v EPA
United States nth ?nn4
EPA/600/P-°99e/002,F
Air Quality Criteria for
Particulate Matter
Volume I of II
-------
EPA/600/P-99/002aF
October 2004
Air Quality Criteria for Particulate Matter
Volume I
National Center for Environmental Assessment-RTF Office
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
PREFACE
National Ambient Air Quality Standards (NAAQS) are promulgated by the United States
Environmental Protection Agency (EPA) to meet requirements set forth in Sections 108 and 109
of the U.S. Clean Air Act (CAA). Sections 108 and 109 require the EPA Administrator (1) to
list widespread air pollutants that reasonably may be expected to endanger public health or
welfare; (2) to issue air quality criteria for them that assess the latest available scientific
information on nature and effects of ambient exposure to them; (3) to set "primary" NAAQS to
protect human health with adequate margin of safety; (4) to set "secondary" NAAQS to protect
against welfare effects (e.g., effects on vegetation, ecosystems, visibility, climate, manmade
materials, etc.); and (5) to periodically (every 5 years) review and revise, as appropriate, the
criteria and NAAQS for a given listed pollutant or class of pollutants.
The original NAAQS for particulate matter (PM), issued in 1971 as "total suspended
particulate" (TSP) standards, were revised in 1987 to focus on protecting against human health
effects associated with exposure to ambient PM less than 10 microns (< 10 jim) that are capable
of being deposited in thoracic (tracheobronchial and alveolar) portions of the lower respiratory
tract. Later periodic reevaluation of newly available scientific information, as presented in the
last previous version of this "Air Quality Criteria for Paniculate Matter" document published in
1996, provided key scientific bases for PM NAAQS decisions published in July 1997. More
specifically, the PM10 NAAQS set in 1987 (150 |ig/m3, 24-h; 50 |ig/m3, annual average) were
retained in modified form and new standards (65 |ig/m3, 24-h; 15 |ig/m3, annual average) for
particles < 2.5 jim (PM25) were promulgated in July 1997.
This final version of revised Air Quality Criteria for P articulate Matter assesses new
scientific information that has become available (published or accepted for publication) mainly
I-ii
-------
between early 1996 through April 2002, although a few important new studies published through
2003 are also considered. Several previous successive drafts of this document were released for
public comment and review by the Clean Air Scientific Advisory Committee (CASAC), to
obtain comments on the organization and structure of the document, the issues addressed, the
approaches employed in assessing and interpreting the newly available information on PM
exposures and effects, and the key findings and conclusions arrived at by this assessment. Public
comments and CASAC review recommendations were taken into account in making revisions to
this document for incorporation into this final draft. Evaluations contained in this document will
be drawn on to provide inputs to associated PM Staff Paper analyses prepared by EPA's Office
of Air Quality Planning and Standards (OAQPS) to pose alternatives for consideration by the
EPA Administrator with regard to proposal and, ultimately, promulgation of decisions on
potential retention or revision of the current PM NAAQS.
The document describes the nature, sources, distribution, measurement, and concentrations
of PM in outdoor (ambient) environments. It also evaluates the latest data on human exposures
to ambient PM and consequent health effects in exposed human populations, to support decision
making regarding primary, health-related PM NAAQS. The document also evaluates ambient
PM environmental effects on vegetation and ecosystems, visibility, and man-made materials, as
well as atmospheric PM effects on climate change processes, to support decision making bearing
on secondary PM NAAQS.
Preparation of this document was coordinated by EPA's National Center for
Environmental Assessment in Research Triangle Park (NCEA-RTP). NCEA-RTP scientific
staff, together with experts from other EPA/ORD laboratories and academia, contributed to
writing of document chapters. The NCEA of EPA acknowledges the contributions provided by
authors, contributors, and reviewers and the diligence of its staff and contractors in the
preparation of this document.
I-iii
-------
Air Quality Criteria for Particulate Matter
VOLUME I
1. INTRODUCTION 1-1
2. PHYSICS, CHEMISTRY, AND MEASUREMENT OF
OF PARTICULATE MATTER 2-1
APPENDIX 2A: Techniques for Measuring of Semivolatile
Organic Compounds 2A-1
APPENDIX 2B: Analytical Techniques 2B-1
3. CONCENTRATIONS, SOURCES, AND EMISSIONS OF
ATMOSPHERIC PARTICULATE MATTER 3-1
APPENDIX 3 A: Composition of Particulate Matter Source Emissions .... 3A-1
APPENDIX 3B: Organic Composition of Particulate Matter 3B-1
APPENDIX 3C: Aerosol Composition Data from the Speciation
Network 3C-1
APPENDIX 3D: Spatial and Temporal Variability of the Nationwide
AIRS PM25 and PM10.25 Data Sets 3D-1
APPENDIX 3E: Characterization of PM25, PM10, and PM10_25
Concentrations at IMPROVE Sites 3E-1
4. ENVIRONMENTAL EFFECTS OF AIRBORNE PARTICULATE
MATTER 4-1
APPENDIX 4A: Common and Latin Names 4A-1
5. HUMAN EXPOSURE TO PARTICULATE MATTER AND
ITS CONSTITUENTS 5-1
I-iv
-------
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
APPENDIX 7A: Rat-to-Human Dose Extrapolation 7A-1
APPENDIX 7B: Ambient Bioaerosols 7B-1
8. EPIDEMIOLOGY OF HUMAN HEALTH EFFECTS ASSOCIATED
WITH AMBIENT PARTICULATE MATTER 8-1
APPENDIX 8A: Short-Term PM Exposure-Mortality Studies:
Summary Tables 8A-1
APPENDIX 8B: Particulate Matter-Morbidity Studies:
Summary Tables 8B-1
9. INTEGRATIVE SYNTHESIS 9-1
I-v
-------
Table of Contents
Page
List of Tables I-xiii
List of Figures I-xix
Authors, Contributors, and Reviewers I-xxxii
U.S. Environmental Protection Agency Project Team for Development of
Air Quality Criteria for Particulate Matter I-xl
U.S. Environmental Protection Agency Science Advisory Board (SAB)
Staff Office Clean Air Scientific Advisory Committee (CASAC)
Particulate Matter Review Panel I-xliii
Abbreviations and Acronyms I-xlvi
1. INTRODUCTION 1-1
1.1 BACKGROUND 1-1
1.1.1 Legislative Requirements 1-1
1.1.2 Criteria and NAAQS Review Process 1-2
1.1.3 History of Earlier PM Criteria and NAAQS Reviews 1-4
1.1.4 The 1997PMNAAQS Revision 1-5
1.1.5 Coordinated PM Research Program 1-7
1.2 CURRENT PM CRITERIA AND NAAQS REVIEW 1-12
1.2.1 Key Milestones and Procedures for Document Preparation 1-12
1.2.2 Assessment Approaches 1-16
1.3 DOCUMENT ORGANIZATION 1-18
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 2-5
2.1.2.3 Ultrafme Particles 2-28
2.1.3 Chemistry of Atmospheric Particulate Matter 2-37
2.1.3.1 Chemical Composition and Its Dependence on
Particle Size 2-37
2.1.3.2 Primary and Secondary Particulate Matter 2-38
2.1.3.3 Particle-Vapor Partitioning 2-39
2.1.3.4 Atmospheric Lifetimes and Removal Processes 2-49
2.1.4 Comparison of Fine and Coarse Particles 2-51
2.2 MEASUREMENT OF PARTICULATE MATTER 2-51
2.2.1 Particle Measurements of Interest 2-51
2.2.2 Issues in Measurement of Particulate Matter 2-54
2.2.2.1 Artifacts Due to Chemical Reactions 2-55
I-vi
-------
Table of Contents
(cont'd)
Page
2.2.2.2 Treatment of Semivolatile Components of
Paniculate Matter 2-56
2.2.2.3 Upper Cut Point 2-57
2.2.2.4 Cut Point for Separation of Fine and Coarse
Paniculate Matter 2-58
2.2.2.5 Treatment of Pressure, Temperature, and
Relative Humidity 2-58
2.2.2.6 Time Resolution 2-59
2.2.2.7 Accuracy and Precision 2-59
2.2.3 Measurement of Semivolatile Paniculate Matter 2-62
2.2.3.1 Particle-Bound Water 2-62
2.2.3.2 Nitrate and Organic Species 2-68
2.2.3.3 Continuous Measurement of Semivolatile and
Nonvolatile Mass 2-75
2.2.4 U.S. Environmental Protection Agency Monitoring Methods 2-79
2.2.4.1 The Federal Reference Methods for Measurement
of Equilibrated Mass for PM10, PM2 5, and PM10.25 2-79
2.2.5 Speciation Monitoring 2-93
2.2.6 Inorganic Elemental Analyses 2-96
2.2.7 Elemental and Organic Carbon in Particulate Matter 2-97
2.2.8 Ionic Species 2-99
2.2.9 Continuous Monitoring 2-99
2.2.10 Measurements of Individual Particles 2-101
2.2.11 Low Flow Filter Samples for Multiday Collection of
Particulate Matter 2-103
2.3 SUMMARY AND KEY POINTS 2-104
2.3.1 Chemistry and Physics of Atmospheric Particles 2-104
2.3.2 Measurement of Atmospheric Particles 2-107
2.3.3 Key Points 2-111
REFERENCES 2-113
APPENDIX 2A: TECHNIQUES FOR MEASURING OF SEMIVOLATILE
ORGANIC COMPOUNDS 2A-1
APPENDIX 2B: ANALYTICAL TECHNIQUES 2B-1
2B.1 INORGANIC ELEMENTS 2B-1
2B.1.1 Energy Dispersive X-Ray Fluorescence 2B-1
2B.1.2 Synchrotron Induced X-ray Fluorescence 2B-2
2B.1.3 Proton (or Particle) Induced X-ray Emission 2B-2
2B.1.4 Proton Elastic Scattering Analysis 2B-3
2B.1.5 Total Reflection X-Ray Fluorescence 2B-3
2B.1.6 Instrumental Neutron Activation Analysis 2B-4
I-vii
-------
Table of Contents
(cont'd)
Page
2B.1.7 Atomic Absorption Spectrophotometry 2B-5
2B. 1.8 Inductively CoupledPlasmawith Atomic
Emission Spectroscopy 2B-6
2B. 1.9 Inductively Coupled Plasma with
Mass Spectroscopy 2B-6
2B.1.10 Scanning Electron Microscopy 2B-7
2B.2 ORGANIC AND INORGANIC CARBON 2B-8
2B.3 CONTINUOUS METHODS 2B-17
2B.3.1 Continuous Measurement of Mass 2B-17
2B.3.2 Continuous Measurement of Organic and/or
Elemental Carbon 2B-21
2B.3.3 Continuous Measurements of Nitrate and Sulfate 2B-24
2B.4 OTHER CONTINUOUS MEASUREMENTS 2B-26
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 inPM Concentrations 3-13
3.2.2 Diurnal (Circadian) Variability in PM Concentrations 3-30
3.2.3 Relationships Among Particulate Matter in Different
Size Fractions 3-32
3.2.4 Relationships Between Mass and Chemical
Component Concentrations 3-34
3.2.5 Spatial Variability in Particulate Matter and its Components 3-40
3.3 SOURCES OF PRIMARY AND SECONDARY PM 3-59
3.3.1 Chemistry of Secondary PM Formation 3-62
3.3.2 Source Contributions to Ambient PM Determined by
Receptor Models 3-72
3.3.3 Background Concentrations of PM in the United States 3-82
3.3.4 Emissions Estimates for Primary PM, and Precursors
to Secondary PM (SO2, NOX, VOCs, and NH3) in the
United States 3-92
3.3.5 Uncertainties of Emissions Inventories 3-98
3.4 SUMMARY AND KEY CONCLUSIONS 3-101
REFERENCES 3-106
APPENDIX 3 A: Spatial and Temporal Variability of the Nationwide AIRS
PM25 and PM10_25 Data Sets 3A-l
REFERENCES 3A-3
I-viii
-------
Table of Contents
(cont'd)
Page
APPENDIX 3B: Aerosol Composition Data from the Speciation Network 3B-1
REFERENCES 3B-3
APPENDIX 3C: Organic Composition of Particulate Matter 3C-1
REFERENCES 3C-14
APPENDIX 3D: Composition of Particulate Matter Source Emissions 3D-1
REFERENCES 3D-25
APPENDIX 3E: Characterization of PM2 5, PM10, and PM10.2 5
Concentrations at IMPROVE Sites 3E-1
REFERENCES 3E-6
4. ENVIRONMENTAL EFFECTS OF AIRBORNE PARTICULATE MATTER 4-1
4.1 INTRODUCTION 4-1
4.2 EFFECTS OF AMBIENT AIRBORNE PARTICULATE MATTER ON
VEGETATION AND NATURAL ECOSYSTEMS 4-1
4.2.1 Ecological Attributes 4-2
4.2.2 Ecosystem Exposures - Particle Deposition 4-4
4.2.2.1 Fine and Coarse Particulate Matter 4-5
4.2.2.2 Diversity of Deposition Modes 4-8
4.2.2.3 Magnitude of Deposition 4-44
4.2.3 Assessment of Atmospheric PM Deposit!on Effects 4-55
4.2.3.1 Effects on Vegetation and Ecosystems 4-61
4.2.3.2 Ecosystem Response to Stresses 4-79
4.2.4 Urban Ecosystems 4-148
4.3 AIRBORNE PARTICLE EFFECTS ON VISIBILITY 4-152
4.3.1 Introduction 4-152
4.3.2 Factors Affecting Atmospheric Visibility 4-154
4.3.2.1 Optical Properties of the Atmosphere and
Atmospheric Particles 4-154
4.3.2.2 Relative Humidity Effects on Particle Size and
Light-Scattering Properties 4-161
4.3.3 Relationships Between Particles and Visibility 4-165
4.3.4 Photographic Modeling of Visibility Impairment 4-173
4.3.5 Visibility Monitoring Methods and Networks 4-174
4.3.6 Visibility Impairment: Trends and Current Conditions 4-177
4.3.6.1 Trends in Visibility Impairment 4-177
4.3.6.2 Current Conditions 4-185
4.3.7 Societal Impacts of Parti culate Matter Visibility Effects 4-186
4.3.7.1 Economic Studies 4-186
4.3.7.2 Public Perception and Attitude Studies 4-188
4.4 PARTICULATE MATTER EFFECTS ON MATERIALS 4-191
4.4.1 Corrosive Effects of Particles and Sulfur Dioxide on
Man-Made Surfaces 4-191
I-ix
-------
Table of Contents
(cont'd)
Page
4.4.1.1 Metals 4-192
4.4.1.2 Painted Finishes 4-193
4.4.1.3 Stone and Concrete 4-196
4.4.2 Soiling and Discoloration of Man-Made Surfaces 4-202
4.4.2.1 Stones and Concrete 4-203
4.4.2.2 Household and Industrial Paints 4-203
4.5 ATMOSPHERIC PARTICULATE MATTER, CLIMATE CHANGE,
AND EFFECTS ON SOLAR UV-B RADIATION TRANSMISSION 4-204
4.5.1 Atmospheric Particle Interactions with Solar and Terrestrial
Radiation Related to Climate Change 4-206
4.5.1.1 The Projected Impacts of Global Climate Change .... 4-206
4.5.1.2 Airborne Particle Relationships to Global
Warming and Climate Change 4-209
4.5.2 Atmospheric Particulate Matter Effects on the Transmission
of Solar Ultraviolet Radiation Transmission: Impacts on
Human Health and the Environment 4-220
4.5.2.1 Potential Effects of Increased Ultraviolet
Radiation Transmission 4-220
4.5.2.2 Airborne Particle Effects on Atmospheric
Transmission of Solar Ultraviolet Radiation 4-223
4.6 SUMMARY AND KEY CONCLUSIONS 4-227
4.6.1 Particulate Matter Effects on Vegetation and Ecosystems 4-227
4.6.2 Particulate Matter-Related Effects on Visibility 4-235
4.6.3 Particulate Matter-Related Effects on Materials 4-238
4.6.4 Effects of Atmospheric Particulate Matter on Global
Warming Processes and Transmission of Solar
Ultraviolet Radiation 4-239
REFERENCES 4-241
APPENDIX 4A: Common and Latin Names 4A-1
5. HUMAN EXPO SURE 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
5.1.4 Chapter Structure 5-4
5.2 BASIC CONCEPTS OF EXPOSURE 5-5
5.2.1 The Concept of Exposure 5-5
5.2.2 Components of Exposure 5-6
5.2.3 Quantification of Exposure 5-9
I-x
-------
Table of Contents
(cont'd)
Page
5.2.4 Methods To Estimate Personal Exposure 5-10
5.2.4.1 Direct Measurement Methods 5-11
5.2.4.2 Indirect or Modeling Methods 5-15
5.2.4.3 Methods for Estimating Personal Exposure to
Ambient Particulate Matter 5-22
5.3 SUMMARY OF PARTICULATE MATTER MASS DATA 5-29
5.3.1 Types of Particulate Matter Exposure Measurement Studies 5-29
5.3.2 Available Data 5-30
5.3.2.1 Personal Exposure Data 5-30
5.3.2.2 Microenvironmental Data 5-35
5.3.2.3 Traffic-Related Microenvironments 5-43
5.3.2.4 Reanalyses of Previously-Reported Particulate
Matter Exposure Data 5-43
5.3.3 Factors Influencing and Key Findings on Particulate
Matter Exposures 5-46
5.3.3.1 Relationship of Personal/Microenvironmental
Particulate Matter with Ambient Particulate Matter .... 5-46
5.3.3.2 Factors That Affect Relationships Between
Personal Exposure and Ambient PM 5-66
5.3.3.3 Effect of Ambient Sources on Exposures to
Particulate Matter 5-87
5.3.3.4 Correlations of Particulate Matter with
Other Pollutants 5-89
5.4 SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA 5-94
5.4.1 Introduction 5-94
5.4.2 Monitoring Studies That Address Particulate Matter
Constituents 5-94
5.4.3 Key Findings 5-94
5.4.3.1 Correlations of Personal and Indoor Concentrations
with Ambient Concentrations of Particulate Matter
Constituents 5-94
5.4.4 Factors Affecting Correlations Between Ambient
Measurements and Personal or Microenvironmental
Measurements of Particulate Matter Constituents 5-105
5.4.5 Limitations of Available Data 5-106
5.5 IMPLICATIONS OF USING AMBIENT PM CONCENTRATIONS
IN TOXICOLOGICAL AND EPIDEMIOLOGICAL STUDIES OF
PM HEALTH EFFECTS 5-107
5.5.1 Toxicology 5-107
5.5.2 Potential Sources of Error Resulting from Using Ambient
Particulate Matter Concentrations in Epidemiological
Analyses 5-107
I-xi
-------
Table of Contents
(cont'd)
Page
5.5.2.1 Associations Between Personal Exposures and
Ambient Particulate Matter Concentrations 5-110
5.5.2.2 Role of Compositional Differences in Exposure
Characterization for Epidemiology 5-114
5.5.2.3 Role of Spatial Variability in Exposure
Characterization for Epidemiology 5-115
5.5.3 Analysis of Exposure Measurement Error Issues in Parti culate
Matter Epidemiology 5-116
5.5.3.1 Time-Series Analyses 5-117
5.5.3.2 Studies of Chronic Effects 5-117
5.5.4 Conclusions from Analysis of Exposure Measurement Errors
on Particulate Matter Epidemiology 5-120
5.6 SUMMARY OF OBSERVATIONS AND LIMITATIONS 5-121
REFERENCES 5-129
I-xii
-------
List of Tables
Number Page
1-1 Key Scientific Uncertainties Related to the Source-To-Response
Framework 1-10
1-2 Research Topics and Questions Recommended by National Research
Council to be Addressed by Expanded M Research Program 1-11
1-3 Key Milestones in the Development of This Document 1-13
2-1 Correlations Between TEOM Measurements in Phoenix 2-26
2-2 Comparison of Ambient Particles, Fine Particles (ultrafme plus
accumulation-mode) and Coarse Particles 2-52
2-3 Particulate Matter Components/Parameters of Interest for Health,
Ecological, or Radiative Effects; for Source Category Apportionment
Studies; or for Air Quality Modeling Evaluation Studies 2-53
2-4 Summary of Sensitivity Studies of WINS Impactor Performance 2-84
2-5 PM2 5 Samplers Currently Designated as FRMs for PM2 5 Mass
Concentrations 2-85
2-6 Measurement Methods for Inorganic Elements 2-97
2-7 Methods for Continuous Measurement of PM Mass, PM
Components, etc 2-100
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 Characteristic (EPA/HEI) Regions
in 1999 3-33
3-2 Concentrations (ng/m3) of PM2 5, PM10_25, and Selected Elements (ng/m3)
in the PM2 5 and PM10_2 5 Size Ranges with Standard Deviations (SD) and
Correlations Between Elements and PM25 Mass in Philadelphia, PA 3-35
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 (SD) and
Correlations (r) Between Elements and PM2 5 and PM10_2 5 Mass in
Phoenix, AZ 3-36
3-4a-d Measures of the Spatial Variability of PM25 Concentrations Within
Selected Metropolitan Statistical Areas 3-41
I-xiii
-------
List of Tables
(Cont'd)
Number Page
3-5 Measures of the Spatial Variability of PM10_25 Concentrations Within
Selected Metropolitan Statistical Areas 3-49
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) 3-57
3-7 Correlation Coefficients for Spatial Variation of PM25 Mass and Different
Components for Pairs of Sampling Sites in Philadelphia (1994) 3-57
3-8 Constituents of Atmospheric Particles and their Major Sources 3-60
3-9 Receptor Model Source Contributions to PM2 5 3-76
3-10 Receptor Model Source Contributions to PM10 3-77
3-11 Emissions of Primary PM25 by Various Sources in 1999 3-94
3-12 Emissions of Precursors to Secondary PM2 5 Formation by Various Sources
in 1999 3-95
3-13 Nationwide Changes in Ambient Concentrations and Emissions of PM10 and
Gaseous Precursors to Secondary Particulate Matter from 1992 to 2001 3-97
3A-1 Performance Metrics forPM25 from Collocated Samplers 3A-4
3B-1 PM2 5 Speciation Samplers by Location: Sites Selected for Summary in
2004 PM AQCD 3B-4
3B-2a Burlington, VT Summary Data (October 2001 to September 2002) 3B-5
3B-2b Philadelphia, PA Summary Data (October 2001 to September 2002) 3B-6
3B-2c Atlanta, GA Summary Data (October 2001 to September 2002) 3B-7
3B-2d Detroit, MI Summary Data (October 2001 to September 2002) 3B-8
3B-2e Chicago, IL Summary Data (October 2001 to September 2002) 3B-9
3B-2f St. Louis, MI Summary Data (October 2001 to September 2002) 3B-10
3B-2g Houston, TX Summary Data (October 2001 to September 2002) 3B-11
I-xiv
-------
List of Tables
(Cont'd)
Number Page
3B-2h Minneapolis, MN Summary Data (October 2001 to September 2002) 3B-12
3B-2i Boulder, CO Summary Data (October 2001 to September 2002) 3B-13
3B-2J Phoenix, AZ Summary Data (October 2001 to September 2002) 3B-14
3B-2k Seattle, WA Summary Data (October 2001 to September 2002) 3B-15
3B-21 Sacramento, CA Summary Data (October 2001 to September 2002) 3B-16
3B-2m Riverside-Rubidoux, CA Summary Data (October 2001 to
September 2002) 3B-17
3B-3 Blank Corrections for Elemental Carbon, Organic Carbon, and Total
Carbon in the Speciation Network 3B-18
3C-1 Paniculate Organic and Elemental Carbon Concentrations (in jig C/m3)
Based on Studies Published After 1995 3C-2
3C-2 Paniculate 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-5
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
Paniculate Matter Exhaust 3D-12
3D-4b Contribution of Organic Carbon to Particulate Matter Carbon Emissions
in Motor Vehicle Exhaust Collected from Vehicles Operated on Chassis
Dynamometers 3D-13
3D-5 Emission Rates (mg/mi) for Constituents of Particulate Matter from
Gasoline and Diesel Vehicles 3D-14
3D-6 Summary of Particle-Phase Organic Compounds Emitted from
Motor Vehicles 3D-17
I-xv
-------
List of Tables
(Cont'd)
Number Page
3D-7 Mass Emissions, Organic Carbon, and Elemental Carbon Emissions from
Residential Combustion of Wood 3D-19
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-22
3E-1 Ranges of Annual Mean PM Concentrations at IMPROVE Monitoring
Sites (jig/m3) 3E-3
3E-2 Ranges of Annual 90th Percentile Values of Daily PM Concentrations at
IMPROVE Monitoring Sites (|ig/m3) 3E-4
4-1 Essential Ecological Attributes and Reporting Categories 4-3
4-2 Types and Determinants of Particulate Deposition and Impact
to Vegetation 4-9
4-3 Relative Importance of Wet, Dry, Parti culate, and Total Deposition to
Three Forest Sites 4-12
4-4 Reported Mean Deposition Velocities for Sulfate, Chlorine, Nitrate,
and Ammonium-Ion-Containing Particles 4-37
4-5 Representative Empirical Measurements of Deposition Velocity for
Paniculate Deposition 4-38
4-6 Reported Mean Deposition Velocities for Potassium, Sodium, Calcium,
and Magnesium Base Cation Containing Particles 4-38
4-7 Mean Annual Nitrogen Deposition (equivalents/ha/year) from Fine and
Coarse Particles Compared to Total Nitrogen Deposition from all Sources
to a Variety of Forest Ecosystems 4-46
4-8 Mean Annual Sulfate Deposition (equivalents/ha/year) from Fine and
Coarse Particles Compared to Total Sulfur Deposition from all Sources
to a Variety of Forest Ecosystems 4-47
I-xvi
-------
List of Tables
(Cont'd)
Number Page
4-9 Mean Annual Base Cation Deposition (equivalents/ha/year) from Fine
and Coarse Particles Compared to Total Base Cation Deposition from
all Sources to a Variety of Forest Ecosystems 4-48
4-10 Mean Particle Size, Deposition Rates, and Derived Deposition Velocities
for Heavy Metal Deposition to the Upper Canopy (inert plates or leaves)
of an Upland Oak Forest 4-52
4-11 Total Heavy Metal Deposition to Temperate Latitude Forests 4-52
4-12 Primary Goods and Services Provided by Ecosystems 4-58
4-13 Ecosystem Functions Impacted by Air Pollution Effects on Temperate
Forest Ecosystems 4-84
4-14 Nitrogen-Saturated Forests in North America, Including Estimated
N Inputs and Outputs 4-102
4-15 Visibility Measurement Techniques 4-171
4-16 Residential Visibility Contingent Valuation Study Results 4-189
4-17 Residential Visibility Valuation Study Results for Los Angeles and
San Francisco 4-190
4-18 Corrosive Effects of Particulate Matter and Sulfur Dioxide on Metals 4-194
4-19 Corrosive Effects of Particulate Matter and Sulfur Dioxide on Stone 4-197
4-20 Examples of Impacts Resulting from Projected Changes in Extreme
Climate Events 4-208
4-21 Effects of Reactive Nitrogen 4-230
5-1 Types of Parti culate Matter Used in Exposure and Concentration Variables . . . 5-7
5-2 Activity Pattern Studies Included in the Consolidated Human
Activity Database 5-18
5-3 Personal Exposure Models for Parti culate Matter 5-19
5-4 Summary of Recent PM Personal Exposure Studies 5-31
I-xvii
-------
List of Tables
(cont'd)
Number Page
5-5 Summary of Recent Microenvironmental PM Measurement Studies 5-36
5-6 Papers Reporting Reanalyses of Particulate Matter Exposure Studies 5-44
5-7 Personal Monitoring Studies for Particulate Matter: Measured
Concentrations and Correlation Coefficients 5-49
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 5-58
5-9 Daily-average Correlation Coefficients Reported for the Baltimore
(Williams et al., 2000a,b,c) and the Fresno (Evans et al., 2000) Studies 5-58
5-10 Regression Analysis Reported for Indoor/Outdoor Relationships for
PM25 in the Baltimore (Williams et al., 2000a,b,c) and Fresno (Evans
et al., 2000) Studies 5-58
5-11 Volume Mean Diameter and Maximum PM2 5 Concentrations of Indoor
Particle Sources 5-84
5-12 Correlations Between Personal PM25 and Ambient Pollutant
Concentrations 5-91
5-13 Correlations Between Hourly Personal PM25 and Gaseous Pollutants 5-93
5-14 Studies That have Measured Particulate Matter Constituents in Personal
Exposure Samples 5-96
5-15 Studies That Have Measured Particulate Matter Constituents in
. 5-97
5-16
5-17
5-18
5-19
5-20
Summary Statistics for Personal, Indoor, and Outdoor Concentrations of
Selected Aerosol Components in Two Pennsylvania Communities
Regression Analysis of Indoor Versus Outdoor Concentrations
Mixed Model Analysis of Personal Versus Outdoor Concentrations
Regression Analysis of Indoor Versus Outdoor Concentrations
Regression Analysis of Indoor Versus Outdoor Concentrations
. . . 5-99
. . 5-101
. . 5-103
. . 5-103
5-104
I-xviii
-------
List of Figures
Number Page
1-1 A general framework for integrating PM research 1-9
2-1 Distribution of coarse (c), accumulation (a), and nuclei (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
2-2 Particle size distributions by number: (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 . . 2-9
2-3 Size distribution by volume for the averaged (a) rural and urban-influenced
rural number distribution shown in Figure 2-2a and a distribution from
south central New Mexico, and (b) for the averaged urban and freeway-
influenced urban number distributions shown in Figure 2-2a 2-10
2-4 Volume size distribution, measured in traffic, showing fine and coarse
particles and the nuclei and accumulation modes of fine particles 2-11
2-5 Sub-micrometer number size distribution observed in a boreal forest in
Finland showing the tri-modal structure of fine particles 2-12
2-6 An idealized size distribution, that might be observed in traffic, showing
fine and coarse particles and the nucleation, Aitken, and accumulation
modes that comprise fine particles 2-13
2-7 Specified particle penetration (size-cut curves) through an ideal (no-particle-
loss) inlet for five different size-selective sampling criteria 2-16
2-8 An idealized distribution of ambient particulate matter showing fine-mode
particles and coarse-mode particles and the fractions collected by size-
selective samplers 2-18
2-9 Comparison of penetration curves for two PM10 beta gauge samplers using
cyclone inlets 2-19
2-10 Three examples of impactor size distributions and distributions resulting
from fitting several log normal distributions to the impactor data 2-24
2-11 Particulate matter concentrations in Spokane, WA, during the August 30,
1996 dust storm 2-27
2-12 Typical engine exhaust size distribution 2-32
I-xix
-------
List of Figures
(cont'd)
Number Page
2-13 Number size distributions showing measurement of a nucleation burst
mode in a boreal forest in Finland 2-33
2-14 Examples of the measured 1-h average particle number size distributions
and the log normal fits to the modes of the data 2-34
2-15a-f Fitted multi-model particle size distribution at different sampling distances
from Freeway 405: (a) 30 m downwind, (b) 60 m downwind, (c) 90 m
downwind, (d) 150 m downwind, (e) 300 m downwind, (f) 300 m upwind . . . 2-35
2-15g Combination of Figures 2-15(a-e), with dN/d logDp scale 2-36
2-16a 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 2-42
2-16b Theoretical predictions and experimental measurements of growth of
NH4HSO4 particles at relative humidities between 95 and 100% 2-43
2-17 Concentration of the dissolved gas in the particle normalized by its initial
concentration as a function of airway generation forH= 104 M atnT1
for particle diameters of 0.1, 0.3, 0.7, and 1 |im 2-46
2-18 Schematic showing major nonvolatile and semivolatile components of PM25 . 2-57
2-19 Aerosol water content expressed as a mass percentage, as a function of
relative humidity 2-65
2-20 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 2-69
2-21 Average concentration (|ig/m3) of nonvolatile and semivolatile PM
components in three cities 2-76
2-22 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) 2-77
I-xx
-------
List of Figures
(cont'd)
Number Page
2-23 Schematic diagram of the sample collection portion of the PM25
FRM sampler 2-81
2-24 Schematic view of the final design of the WINS 2-82
2-25 Evaluation of the final version of the WINS 2-83
2-26 Schematic diagram showing the principle of virtual impaction 2-91
2-27 Size distribution of particles divided by chemical classification into
organic, marine, and soil (or crustal) 2-102
2B-1 Thermogram for a sample containing rock dust (a carbonate source) and
diesel exhaust, shows three traces that correspond to temperature, filter
transmittance, and Flame lonization Detector (FID) detector response 2B-10
2B-2 Examples of thermograms obtained by (a) the IMPROVE protocol,
and by (b) the NIOSH protocol 2B-12
3-la 1999-2001 county-wide average annual mean PM10 concentrations
(|ig/m3) for counties with PM10 monitors 3-7
3-lb 1999-2001 highest county-wide 98th percentile 24-h average PM10
concentrations (|ig/m3) for counties with PM10 monitors 3-7
3-2 Nationwide trend in ambient PM10 concentration from 1992 through 2001 .... 3-8
3-3 Trend in PM10 annual mean concentrations by EPA region, 1992 through
2001 (jig/m3) 3-9
3-4a 1999-2001 county-wide average annual mean PM25 concentrations
(|ig/m3) for counties with PM25 monitors 3-10
3-4b 1999-2001 highest county-wide 98th percentile 24-h average PM25
concentrations (|ig/m3) for counties with PM25 monitors 3-10
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 3-11
I-xxi
-------
List of Figures
(cont'd)
Number Page
3-6a 1999-2000 estimated county-wide average annual mean PM10_25
concentrations (|ig/m3) for counties with collocated PM2 5 and
PM10 monitors 3-14
3-6b 1999-2000 estimated county-wide highest 98th percentile 24-h
PM10_2 5 concentrations (|ig/m3) for counties with collocated PM2 5
and PM10 monitors 3-15
3-7a,b Quarterly distribution of 24-h average PM25 concentrations for selected
monitors in the (a) Philadelphia, PA and (b) Cleveland, OH MS As 3-16
3-7c,d Quarterly distribution of 24-h average PM25 concentrations for selected
monitors in the (c) Dallas, TX and (d) Los Angeles, CA MS As 3-17
3-8 Seasonal concentrations of PM2 5 and PM10 measured in the four
MAACS cities 3-20
3-9a,b Quarterly distribution of 24-h average PM10_25 concentrations for selected
sites in the (a) Cleveland, OH; (b) Dallas, TX MS As 3-22
3-9c Quarterly distribution of 24-h average PM10_25 concentrations for selected
sites in the Los Angeles, CA MSAs 3-23
3-10 Frequency distribution of 24-h average PM25 concentrations measured
at the Presbyterian home (PBY) monitoring site in southwestern
Philadelphia from 1992 to 1995 3-25
3-11 Concentrations of 24-h average PM2 5 measured at the EPA site in
Phoenix, AZ from 1995 to 1997 3-26
3-12 Frequency distribution of 24-h average PM25 concentrations measured
at the EPA site in Phoenix, AZ from 1995 to 1997 3-27
3-13 Frequency distribution of 24-h average PM25 measurements obtained from
all California Air Resources Board dichotomous sampler sites from 1989
to 1998 3-27
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 3-28
I-xxii
-------
List of Figures
(cont'd)
Number Page
3-15 Concentrations of 24-h average PM2 5 measured at the Riverside-Rubidoux
site from 1989 to 1998 3-28
3-16 Frequency distribution of 24-h average PM25 concentrations measured
at the Riverside-Rubidoux site from 1989 to 1994 3-29
3-17 Intraday variability of hourly average PM25 concentrations across the
United States 3-31
3-18 Occurrence of differences between pairs of sites in three MS As 3-47
3-19 Intersite correlation coefficients for PM25, PM10, and PM10_25 3-51
3-20 PM25 chemical components in downtown Los Angeles and Burbank (1986)
have similar characteristics 3-54
3-21 Concentrations of PM25 chemical components in Rubidoux and downtown
Los Angeles (1986) . 3-55
3-22 Monthly average Saharan dust components in all size fractions of the
aerosol sampled in Miami, FL (from 1974 to 1996) 3-85
3-23 PM25 and PM10 concentrations measured at Chilliwack Airport, located in
southwestern British Columbia, just before and during the Asian desert
dust episode of April and May 1998 3-86
3-24 Time series of 24-h average PM10 concentrations observed in the
Rio Grande Valley during May 1998 3-88
3-25 PM10 concentrations observed in St. Louis, MO, during May 1998 3-89
3A-1 Philadelphia, PA-NJ metropolitan statistical area 3A-5
3A-2 Washington, DC metropolitan statistical area 3A-6
3 A-3 Norfolk, VA metropolitan statistical area 3 A-7
3A-4 Columbia, SC metropolitan statistical area 3A-8
3 A-5 Atlanta, GA metropolitan statistical area 3 A-9
I-xxiii
-------
List of Figures
(cont'd)
Number Page
3A-6 Birmingham, AL metropolitan statistical area 3 A-10
3A-7 Tampa, FL metropolitan statistical area 3A-11
3A-8 Cleveland, OH metropolitan statistical area 3A-12
3A-9 Pittsburgh, PA metropolitan statistical area 3 A-13
3A-10 Steubenville, OH-Weirton, WV metropolitan statistical area 3A-14
3A-11 Detroit MI metropolitan statistical area 3A-15
3A-12 Grand Rapids, MI metropolitan statistical area 3A-16
3 A-13 Milwaukee, WI metropolitan statistical area 3 A-17
3A-14 Chicago, IL metropolitan statistical area 3A-18
3A-15 Gary, IN metropolitan statistical area 3A-19
3 A-16 Louisville, KY metropolitan statistical area 3A-20
3 A-17 St. Louis, MO metropolitan statistical area 3A-21
3 A-18 Baton Rouge, LA metropolitan statistical area 3A-22
3 A-19 Kansas City, KS-MO metropolitan statistical area 3A-23
3 A-20 Dallas, TX metropolitan statistical area 3 A-24
3A-21 Boise, ID metropolitan statistical area 3A-25
3 A-22 Salt Lake City, UT metropolitan statistical area 3 A-26
3 A-23 Seattle, WA metropolitan statistical area 3 A-27
3 A-24 Portland, OR metropolitan statistical area 3 A-28
3A-25 Los Angeles-Long Beach, CA metropolitan statistical area 3A-29
3A-26 Riverside-San Bernadino, CA metropolitan statistical area 3A-30
I-xxiv
-------
List of Figures
(cont'd)
Number Page
3A-27 San Diego, CA metropolitan statistical area 3A-31
3A-28 Columbia, SC metropolitan statistical area 3A-32
3A-29 Tampa, FL metropolitan statistical area 3A-33
3A-30 Cleveland, OH metropolitan statistical area 3A-34
3A-31 Steubenville, OH metropolitan statistical area 3A-35
3A-32 Detroit, MI metropolitan statistical area 3A-36
3A-33 Milwaukee, WI metropolitan statistical area 3A-37
3A-34 Chicago, IL metropolitan statistical area 3A-38
3A-35 Gary, IN metropolitan statistical area 3A-39
3A-36 Louisville, KY metropolitan statistical area 3A-40
3A-37 St. Louis, MO metropolitan statistical area 3A-41
3A-38 Baton Rouge, LA metropolitan statistical area 3A-42
3A-39 Dallas, TX metropolitan statistical area 3A-43
3 A-40 Salt Lake City, UT metropolitan statistical area 3 A-44
3A-41 Portland, OR metropolitan statistical area 3A-45
3A-42 Los Angeles, CA metropolitan statistical area 3A-46
3 A-43 Riverside, CA metropolitan statistical area 3 A-47
3 A-44 San Diego, CA metropolitan statistical area 3 A-48
3D-1 Size distribution of particles generated in a laboratory resuspension
chamber 3D-2
3D-2 Size distribution of California source emissions, 1986 3D-4
I-xxv
-------
List of Figures
(cont'd)
Number Page
3D-3 Chemical abundances for PM2 5 emissions from paved road dust in
Denver, CO 3D-5
3D-4 Chemical abundances for PM2 5 emissions from wood burning in
Denver, CO 3D-18
3E-la,b Interannual variability in 24-h average PM25 concentrations observed at
selected IMPROVE sites: (a) Bridger Wilderness, WY; (b) Yellowstone
National Park, WY 3E-7
3E-2a,b Seasonal variability in 24-h average PM2 5 concentrations observed at
selected IMPROVE sites: (a) Bridger Wilderness, WY; (b) Yellowstone
National Park, WY 3E-8
3E-3a,b Seasonal variability in 24-h average PM10_2 5 concentrations observed at
selected IMPROVE sites: (a) Bridger Wilderness, WY; (b) Yellowstone
National Park, WY 3E-9
4-1 The diversity of fine PM from sites in the western and eastern
United States 4-6
4-2 Relative importance of three modes of deposition of nitrate (A) and
sulfate (B) at high elevation sites (Unsworth and Wilshaw, 1989;
Fowler et al., 1989; Mueller, 1988; Aneja and Murthy, 1994) 4-11
4-3 A simplified resistance catena representing the factors controlling
deposition of particles to the surface 4-16
4-4 The relationship between deposition velocity of selected particulate
materials and the distribution of the material between the coarse- and
fine-aerosol fractions 4-17
4-5 The relationship between particle diameter and deposition velocity
for particles 4-18
4-6 Vertical stratification of diverse, chemically speciated particles in a
mixed oak forest 4-33
4-7 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 4-34
I-xxvi
-------
List of Figures
(cont'd)
Number Page
4-8 Range of percent of total deposition delivered in the dry particulate
fraction, across the sites of the Integrated Forest Study 4-40
4-9 Contribution of parti culate deposition to total deposition of nitrogen,
sulfur, and base cations 4-40
4-10 Mean percent of total nitrogen, sulfur, or base cation deposition
contributed by fine plus coarse particles 4-49
4-11 Annual total deposition of heavy metals to Hubbard Brook Experimental
Forest, NH 4-53
4-12 Sample stressors and the essential ecological attributes they affect
(after Science Advisory Board, 2002) 4-56
4-13 Linkages among various ecosystem goods and services (food, water,
biodiversity, forest products) and other driving forces (climate change) 4-60
4-14 Effects of environmental stress on forest trees are presented on a
hierarchial scale for the leaf, branch, tree, and stand levels of organization . . . 4-80
4-15 Illustration of the nitrogen cascade showing the movement of the
human-produced reactive nitrogen (Nr) as it cycles through the various
environmental reservoirs in the atmosphere, terrestrial ecosystems, and
aquatic ecosystems 4-97
4-16 Nitrogen cycle (dotted lines indicate processes altered by
nitrogen saturation) 4-99
4-17 Diagrammatic overview of excess nitrogen in North America 4-103
4-18 Schematic of sources and sinks of hydrogen ions in a forest 4-117
4-19 Key elements of proposed framework for determining critical loads for
nitrogen and sulfur in the United States 4-124
4-20 Calcium deposition in > 2-|im particles, < 2-|im particles, and wet
forms (upper bars) and leaching (lower bars) in the Integrated Forest
Study sites 4-128
I-xxvii
-------
List of Figures
(cont'd)
Number Page
4-21 Magnesium deposition in > 2-|im particles, < 2-|im particles, and wet
forms (upper bars) and leaching (lower bars) in the Integrated Forest
Study sites 4-129
4-22 Potassium deposition in > 2-|im particles, < 2-|im particles, and wet
forms (upper bars) and leaching (lower bars) in the Integrated Forest
Study sites 4-129
4-23 Base cation deposition in > 2-|im particles, < 2-|im particles, and wet
forms (upper bars) and leaching (lower bars) in the Integrated Forest
Study sites 4-130
4-24 Total cation leaching (total height of bar) balanced by sulfate and nitrate
estimated from particulate deposition (assuming no ecosystem retention,
particulate 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 4-131
4-25 Soil exchangeable Ca2+ pools and net annual export of Ca2+ (deposition
minus leaching times 25 years) in the Integrated Forest Study sites 4-133
4-26 Soil exchangeable Mg2+ pools and net annual export of Mg2+ (deposition
minus leaching times 25 years) in the Integrated Forest Study sites 4-133
4-27 Soil exchangeable K+ pools and net annual export of K+ (deposition
minus leaching times 25 years) in the Integrated Forest Study sites 4-134
4-28a Simulated soil solution mineral acid anions in the red spruce site with
no change, 50% N and S deposition, and 50% base cation deposition 4-138
4-28b Simulated soil solution base cations in the red spruce site with no change,
50% N and S deposition, and 50% base cation deposition 4-139
4-29a Simulated soil solution Al in the red spruce site with no change,
50% N and S deposition, and 50% base cation deposition 4-140
4-29b Simulated soil solution soil base saturation in the red spruce site with
no change, 50% N and S deposition, and 50% base cation deposition 4-141
4-3Oa Simulated soil solution mineral acid anions in the Coweeta site with
no change, 50% N and S deposition, and 50% base cation deposition 4-142
I-xxviii
-------
List of Figures
(cont'd)
Number Page
4-3Ob Simulated soil solution base cations in the Coweeta site with no change,
50% N and S deposition, and 50% base cation deposition 4-143
4-31 Relationship of plant nutrients and trace metals with vegetation 4-147
4-32 Light reflected from a target toward an observer 4-156
4-33 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 4-159
4-34 Volume-specific light-scattering efficiency as a function of particle
diameter Dp 4-161
4-35 Particle growth curve as a function of relative humidity (RH) showing
deliquescent growth of ammonium sulfate [(NH4)2 SO4] particles at the
deliquescent point (A, about 80% RH), reversible hygroscopic growth
of ammonium sulfate solution droplets at RH > 80%, and hysteresis
(the droplet remains supersaturated as the RH decreases below 80%)
until the crystallization point (B, about 38% RH) is reached 4-162
4-36 Comparison of extinction (Mm"1) and visual range (km) 4-167
4-37 Proportionality of observed daytime haziness to fine particle mass
concentration in Los Angeles 4-168
4-38 Relative humidity adjustment factor, f(RH), for ammonium sulfate as a
function of relative humidity 4-170
4-39a Aggregate visibility trends (in deciviews) for 10 eastern Class 1 areas 4-180
4-39b Aggregate visibility trends (in deciviews) for 26 western Class 1 areas 4-180
4-40 (a) Eastern class I area aggregate trends in aerosol light extinction on
the 20% haziest days, including trends by major aerosol component;
(b) Western class I area aggregate trends in aerosol light extinction on
the 20% haziest days, including trends by major aerosol component 4-182
4-4la Light extinction trends in Tucson, Arizona from 1993 to 2002 4-184
4-4Ib Light extinction trends in Phoenix, Arizona from 1994 to 2001 4-184
I-xxix
-------
List of Figures
(cont'd)
Number Page
4-42 Estimated global mean radiative forcing exerted by gas and various
particle phase species for the year 2000, relative to 1750 4-213
5-1 Hypothetical exposure time profile: pollutant exposure as a function of
time showing how the averaged exposure, integrated exposure, and peak
exposure relate to the instantaneous exposure 5-9
5-2 Distribution of FINF (a) and a (b) for daytime as estimated from PTEAM
study data 5-25
5-3 Comparison of correlation coefficients for longitudinal analyses of
personal exposure versus ambient concentrations for individual subjects
for PM25 and sulfate 5-61
5-4 Personal exposure versus ambient concentrations for PM2 5 and sulfate for
well, moderately, and poorly ventilated indoor environments 5-64
5-5 Regression analyses of aspects of daytime personal exposure to PM10
estimated using data from the PTEAM study: (a) total personal exposure
to PM, T, regressed on ambient concentration, C; (b) personal exposure
to ambient PM, A, regressed on C; and (c) personal exposure to
nonambient PM, N, regressed on C 5-65
5-6 Air exchange rates measured in homes throughout the United States 5-72
5-7 Box plots of hourly air exchange rates stratified by season in Boston, MA
during 1998 5-73
5-8 Geometric mean infiltration factor (indoor/outdoor ratio) for hourly
nighttime, nonsource data for two seasons 5-74
5-9 Regression of air exchange rate on absolute indoor-outdoor temperature
difference 5-76
5-10 Comparison of deposition rates from Long et al. (2001a) with literature
values (from Abt et al., 2000b) 5-78
5-11 Penetration efficiencies and deposition rates from models of nightly
average data 5-79
5-12 Mean hourly indoor/outdoor particle concentration ratio from an
unoccupied residence in Fresno, CA during spring 1999 5-85
I-xxx
-------
List of Figures
(cont'd)
Number Page
5-13 Personal versus outdoor SO42 in State College, PA 5-100
5-14 Plots of nonambient exposure to PM10 showing (a) daily individual
daytime values from PTEAM data and (b) daily-average values from
THEES data 5-113
I-xxxi
-------
Authors, Contributors, and Reviewers
CHAPTER 1. INTRODUCTION
Principal Authors
Dr. Lester D. Grant—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
CHAPTER 2. PHYSICS, CHEMISTRY, AND MEASUREMENT
OF PARTICULA TE MA TTER
Principal Authors
Dr. William Wilson—National Center for Environmental Assessment (B243-01),
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
Dr. Brooke L. Hemming—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributing Authors
Dr. Tom Cahill, University of California, Davis, One Shields Ave., Davis, CA 95616
Dr. Judith C. Chow, Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512
Max R. Peterson, Research Triangle Institute, P.O. Box 12194, Research Triangle Park, NC
27709
Dr. James J. Schauer, University of Wisconsin, College of Engineering, Department of Civil and
Environmental Engineering, Madison, WI 53706
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. John G. Watson, Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512
I-xxxii
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
Dr. Edward O. Edney—National Exposure Research Laboratory (E205-02)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. R.R. Eldred, Crocker Nuclear Laboratory, University of California, Davis, Davis, CA 95616
Mr. William Ewald—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
James B. Flanagan—Research Triangle Institute, P.O. Box 12194, Research Triangle Park, NC
27709
Dr. Judith Graham—American Chemistry Council, 1300 Wilson Boulevard, Arlington, VA
22207
Dr. Lester D. Grant—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Jim Homolya—Office of Air Quality Planning and Standards (C339-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
R.K.M. Jayanty—Research Triangle Institute, P.O. Box 12194, Research Triangle Park, NC
27709
Mr. Scott Mathias—Office of Air Quality Planning and Standards (C539-01),
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. Joseph Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Joann Rice—Office of Air Quality Planning and Standards (C339-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27709
Ed E. Rickman—Research Triangle Institute, P.O. Box 12194, Research Triangle Park, NC
27709
Dr. Bret Schichtel—Colorado State University, Cooperative Institute for Research in the
Atmosphere, Foothills Campus, Fort Collins, CO 80523-1375
I-xxxiii
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. John J. Vandenberg—National Center for Environmental Assessment (8601D), U.S.
Environmental Protection Agency, 1200 Pennsylvania Avenue, NW, Washington, DC 20460
Dr. Russell Wiener—National Exposure Research Laboratory (D205-03)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
CHAPTER 3. CONCENTRATIONS, SOURCES, AND
EMISSIONS OF A TMOSPHERIC PARTICVLA TE MA TIER
Principal Author
Dr. Joseph P. Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributing Authors
Dr. Allen S. Lefohn—A.S.L. & Associates, 111 North Last Chance Gulch, Helena, MT 59601
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. Robert D. Willis, ManTech Environmental, Inc., Research Triangle Park, NC 27711
Contributors and Reviewers
Dr. JoEllen Bandemeyer—Research Triangle Institute, PO Box 12194, Research Triangle Park,
NC 27709
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
I-xxxiv
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430, Irvine,
CA 92612
Mr. Tom Coulter—National Exposure Research Laboratory (D234-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Edward O. Edney—National Exposure Research Laboratory (E205-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Terence Fitz-Simons—Office of Air Quality Planning and Standards (C3 04-01),
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. Lester D. Grant—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lynn Hildemann—Civil and Environmental Engineering Department, Stanford University,
Stanford, CA 94305
Mr. Phil Lorang—Office of Air Quality Planning and Standards (D205-01),
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 (C539-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Nehzat Mottallebi—California Air Resources Board, 2020 L Street, Sacramento, CA 95814
Mr. Tom Pace—Office of Air Quality Planning and Standards (D205-01), 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. Win Setiawan—California Air Resources Board, 2020 L Street, Sacramento, CA 95814
I-xxxv
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. John J. Vandenberg—National Center for Environmental Assessment (8601D), U.S.
Environmental Protection Agency, 1200 Pennsylvania Avenue, NW, Washington, DC 20460
Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street, Sacramento, CA 95814
CHAPTER 4. ENVIRONMENTAL EFFECTS OF PARTICULA TE MA TTER
Principal Authors
Ms. Beverly Comfort—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. J.H.B. Garner—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lester D. Grant—National Center for Environmental Assessment (B243-01),
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
Dr. Brooke L. Hemming—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dale W. Johnson—Environmental and Resource Science, 1000 Valley Road, University of
Nevada, Reno, NV 89512
Dr. William Malm—National Park Service, Air Resources Division, CIRA, Colorado State
University, Fort Collins, CO
Dr. Joseph P. Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
I-xxxvi
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributing Authors
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. William H. Smith—Professor Emeritus, 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 (D305-01),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Rich Damberg—Office of Air Quality Planning and Standards (C539-02),
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-Cary, NC 27511
Dr. John J. Vandenberg—National Center for Environmental Assessment (8601D), U.S.
Environmental Protection Agency, 1200 Pennsylvania Avenue, NW, Washington, DC 20460
Ms. Debra Meyer Wefering—Duckterather Weg 61, Bergisch Gladbach, Germany 54169
Ms. Kay Whitfield—Office of Air Quality Planning and Standards (C243-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Richard Zepp—National Exposure Research Laboratory (IOD),
U. S. Environmental Protection Agency, Athens, GA
I-xxxvii
-------
Authors, Contributors, and Reviewers
(cont'd)
CHAPTER 5. HUMAN EXPOSURE TO PARTICULA TE MA TIER
AND ITS CONSTITUENTS
Principal Authors
Dr. William E. Wilson—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Linda S. Sheldon—National Exposure Research Laboratory (E205-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Haluk Ozkaynak—National Exposure Research Laboratory (E205-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributing Authors
Dr. Janet Burke—National Exposure Research Laboratory (E205-02),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Roy Fortmann—National Exposure Research Laboratory (E205-02),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David T. Mage—Institute for Survey Research, Temple University,
Philadelphia, PA 19122-6099
Mr. Thomas McCurdy—National Exposure Research Laboratory (E205-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Gary Norris—National Exposure Research Laboratory (E205-03),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Anne Rea—National Exposure Research Laboratory (E205-04),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Alan Vette—National Exposure Research Laboratory (E205-04),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
I-xxxviii
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
Dr. Lester D. Grant—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. John Langstaff—Office of Air Quality Planning and Standards (C539-01)
U.S. Environmental Protection Agency, Research Triangle park, NC 27711
Dr. Joseph Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Harvey Richmond—Office of Air Quality Planning and Standards (C539-01),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. John J. Vandenberg—National Center for Environmental Assessment (8601D), U.S.
Environmental Protection Agency, 1200 Pennsylvania Avenue, NW, Washington, DC 20460
I-xxxix
-------
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 (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Scientific Staff
Dr. Robert W. Elias—PM Team Leader, Health Scientist, National Center for Environmental
Assessment (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Dr. William E. Wilson—Air Quality Coordinator, Physical Scientist, National Center for
Environmental Assessment (B243-01), U.S. Environmental Protection Agency, Research
Triangle Park, NC 27711
Dr. James Brown—Health Scientist, National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly Comfort—Health Scientist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—Health Scientist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. J.H.B. Garner—Ecological Scientist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Brooke Hemming—Health Scientist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis J. Kotchmar—Medical Officer, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph P. Pinto—Physical Scientist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lori White—Health Scientist, National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
I-xl
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
FOR PARTICULATE MATTER
(cont'd)
Technical Support Staff
Ms. Nancy Broom—Information Technology Manager, National Center for Environmental
Assessment (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Mr. Douglas B. Fennell—Technical Information Specialist, National Center for Environmental
Assessment (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Ms. Emily R. Lee—Management Analyst, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Diane H. Ray—Program Specialist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Donna Wicker—Administrative Officer, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Richard Wilson—Clerk, National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Document Production Staff
Ms. Carolyn T. Perry—Manager/Word Processor, Computer Sciences Corporation, 2803 Slater
Road, Suite 220, Morrisville, NC 27560
Dr. Barbara Liljequist—Technical Editor, Computer Sciences Corporation, 2803 Slater Road,
Suite 220, Morrisville, NC 27560
Dr. Carol A. Seagle—Technical Editor, Computer Sciences Corporation, 2803 Slater Road,
Suite 220, Morrisville, NC 27560
Ms. Jessica Long—Graphic Artist, Computer Sciences Corporation, 2803 Slater Road,
Suite 220, Morrisville, NC 27560
Mr. Matthew Kirk—Graphic Artist, Computer Sciences Corporation, 2803 Slater Road,
Suite 220, Morrisville, NC 27560
I-xli
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
FOR PARTICULATE MATTER
(cont'd)
Document Production Staff
(cont'd)
Mr. John A. Bennett—Technical Information Specialist, Library Associates of Maryland,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852
Ms. Sandra L. Hughey—Technical Information Specialist, Library Associates of Maryland,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852
Ms. Rebecca Caffey—Records Management Technician, Reference Retrieval and Database
Entry Clerk, InfoPro, Inc., 8200 Greensboro Drive, Suite 1450, McLean, VA 22102
I-xlii
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
SCIENCE ADVISORY BOARD (SAB) STAFF OFFICE
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE (CASAC)
PARTICULATE MATTER REVIEW PANEL*
CASAC Chair
Dr. Philip Hopke—Bayard D. Clarkson Distinguished Professor, Department of Chemical
Engineering, Clarkson University, Box 5708, Potsdam, NY 13699-5708
CASAC Members
Dr. Ellis Cowling—University Distinguished Professor At-Large, North Carolina State
University, Colleges of Natural Resources and Agriculture and Life Sciences, North Carolina
State University, 1509 Varsity Drive, Raleigh, NC 27695-7632
Dr. James D. Crapo—Chairman, Department of Medicine, National Jewish Medical and
Research Center, 1400 Jackson Street, Denver, CO, 80206, and Chief Executive Officer (CEO)
of Aeolus Pharmaceuticals, Inc.
Dr. Frederick J. Miller—Vice President for Research, CUT Centers for Health Research, 6 Davis
Drive, P.O. Box 12137, Research Triangle Park, NC 27709
Mr. Richard L. Poirot—Environmental Analyst, Air Pollution Control Division, Department of
Environmental Conservation, Vermont Agency of Natural Resources, Bldg. 3 South, 103 South
Main Street, Waterbury, VT 05671-0402
Dr. Frank Speizer—Edward Kass Professor of Medicine, Channing Laboratory, Harvard
Medical School, 181 Longwood Avenue, Boston, MA 02115-5804
Dr. Barbara Zielinska—Research Professor , Division of Atmospheric Science, Desert Research
Institute, 2215 Raggio Parkway, Reno, NV 89512-1095
* Members of this CASAC Panel consist of:
a. CASAC Members: Experts appointed to the statutory Clean Air Scientific Advisory Committee by
the EPA Administrator; and
b. CASAC Consultants: Experts appointed by the SAB Staff Director to serve on one of the
CASAC's National Ambient Air Quality Standards (NAAQS) Panels for a particular criteria air pollutant.
I-xliii
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
SCIENCE ADVISORY BOARD (SAB) STAFF OFFICE
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE (CASAC)
PARTICULATE MATTER REVIEW PANEL*
(cont'd)
CASAC Consultants
Dr. Jane Q. Koenig—Professor, Department of Environmental Health, School of Public Health
and Community Medicine, University of Washington, Box 357234, Seattle, WA 98195-7234
Dr. Petros Koutrakis—Professor of Environmental Science, Environmental Health , School of
Public Health, Harvard University, HSPH, 401 Park Dr., Room 410 West, Boston, MA 02215
Dr. Allan Legge—President, Biosphere Solutions, 1601 11th Avenue NW, Calgary, Alberta,
CANADA T2N 1H1
Dr. Paul J. Lioy—Associate Director and Professor, Environmental and Occupational Health
Sciences Institute, UMDNJ - Robert Wood Johnson Medical School, 170 Frelinghuysen Road,
Rm 301, Piscataway, NJ 08854
Dr. Morton Lippmann—Professor, Nelson Institute of Environmental Medicine, New York
University School of Medicine, 57 Old Forge Road, Tuxedo, NY 10987
Dr. Joe Mauderly—Vice President, Senior Scientist, and Director, National Environmental
Respiratory Center, Lovelace Respiratory Research Institute, 2425 Ridgecrest Drive SE,
Albuquerque, NM, 87108
Dr. Roger O. McClellan—Consultant, 1370 Quaking Aspen Pine, Albuquerque, NM, 87111
Dr. Gunter Oberdorster—Professor of Toxicology, Department of Environmental Medicine,
School of Medicine and Dentistry, University of Rochester, 575 Elmwood Avenue, Box 850,
Rochester, NY 14642
Dr. RobertD. Rowe—President, Stratus Consulting, Inc.,, PO Box 4059, Boulder, CO
80306-4059
Dr. Jonathan M. Samet—Professor and Chair, Department of Epidemiology, Bloomberg School
of Public Health, Johns Hopkins University, 615 N. Wolfe Street, Suite W6041, Baltimore, MD
21205-2179
Dr. Sverre Vedal—Professor of Medicine, School of Public Health and Community Medicine
Department of Environmental and Occupational Health Sciences, University of Washington,
4225 Roosevelt Way NE, Suite 100, Seattle, WA 98105-6099
I-xliv
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
SCIENCE ADVISORY BOARD (SAB) STAFF OFFICE
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE (CASAC)
PARTICULATE MATTER REVIEW PANEL*
(cont'd)
CASAC Consultants
(cont'd)
Mr. Ronald White—Research Scientist, Epidemiology, Bloomberg School of Public Health,
Room W6035, The Johns Hopkins University, 615 N. Wolfe St., Rm W6035, Baltimore, MD
21205
Dr. Warren H. White—Visiting Professor, Crocker Nuclear Laboratory, University of
California-Davis,, Davis, CA 95616-8569
Dr. George T. Wolff—Principal Scientist, General Motors Corporation, 300 Renaissance Center
(MC482-C27-B76), Detroit, MI 48265-3000
EPA Science Advisory Board Staff
Dr. Vanessa Vu—SAB Staff Office Director, EPA Science Advisory Board Staff Office
(Mail Code 1400F), 1200 Pennsylvania Avenue, N.W., Washington DC, 20460
Mr. Fred Butterfield—CASAC Designated Federal Officer, EPA Science Advisory Board Staff
Office (Mail Code 1400F), 1200 Pennsylvania Avenue, N.W., Washington, DC 20460
I-xlv
-------
Abbreviations and Acronyms
AAS
AC
ACGffl
ACS
ADS
AES
AIRS
AM
ANC
APS
AQCD
AQI
ARIES
ASOS
ATOFMS
AWOS
BaP
BASE
BC
BNF
BOSS
BYU
CAA
CAAM
CAMNET
CARB
atomic absorption spectrophotometry
air conditioning
American Conference of Governmental Industrial Hygienists
American Cancer Society
annular denuder system
atomic emission spectroscopy
Aerometric Information Retrieval System
alveolar macrophage
acid neutralizing capacity
aerodynamic particle sizer
Air Quality Criteria Document
Air Quality Index
Aerosol Research and Inhalation Epidemiology Study
Automated Surface Observing System
aerosol time-of-flight mass spectrometry
Automated Weather Observing System
benzo(a)pyrene
Building Assessment and Survey Evaluation
black carbon
bacterial nitrogen fertilization
Brigham Young University Organic Sampling System
Brigham Young University
Clean Air Act
continuous ambient mass monitor
Coordinated Air Monitoring Network
California Air Resources Board
I-xlvi
-------
CASAC
CASTNet
CC
CCPM
CCSEM
CEN
CFCs
CFR
CHAD
CI
GIF
CMAQ
CMB
CMSA
CO AQCD
COD
COH
CONCDEP
COPD
CPC
CRP
CSIRO
CSMCS
CSS
CTM
CV
CVM
D,
Clean Air Scientific Advisory Committee
Clean Air Status and Trends Network
carbonate carbon
continuous coarse particle monitor
computer-controlled scanning electron microscopy
European Standardization Committee
chlorofluorocarbons
Code of Federal Regulations
Consolidated Human Activity Database
confidence interval
charcoal-impregnated cellulose fiber; charcoal-impregnated filter
Community Model for Air Quality
chemical mass balance
Consolidated Metropolitan Statistical Area
Air Quality Criteria Document for Carbon Monoxide
coefficient of divergence
coefficient of haze
average of unitless regional deposition and precipitation-weighted
concentrations
chronic obstructive pulmonary disease
condensation particle counter
Coordinated Research Program
Commonwealth Scientific and Industrial Research Organisation
Carbonaceous Species Methods Comparison Study
coastal sage scrub
chemistry-transport model
coeffi ci ent of van ati on
contigent valuation method
aerodynamic diameter
I-xlvii
-------
DAQ
DMPS
DMS
DP
dv
BAD
EC
EDXRF
EEA
ENSO
EPEC
ESP
ETS
EXPOLIS
FID
FRD
FRM
GAM
GC
GCM
GCVTC
GC/MSD
GHG
GSD
HAPs
H
HBEF
HDS
Department of Air Quality
differential mobility particle sizer
dimethyl sulfide
Stokes particle diameter
deciview
electrical aerosol detector
elemental carbon
energy dispersive X-ray fluorescence
essential ecological attribute
El Nino-Southern Oscillation
Ecological Processes and Effects Committee
electrostatic precipitator
environmental tobacco smoke
Air Pollution Exposure Distributions within Adult Urban Populations in
Europe
flame ionization detection
flame photometric detector
Federal Reference Method
general additive model
gas chromatography
General Circulation Model
Grand Canyon Visibility Transport Commission
gas chromatography/mass-selective detection
greenhouse gas
geometric standard deviation
hazardous air pollutants
Henry's law constant
Hubbard Brook Experimental Forest
honeycomb denuder/filter pack sampler
I-xlviii
-------
HEADS
HEI
HI
hivol
HPEM
HTGC/MS
HVAC
1C
ICP
IPS
IMPROVE
INAA
IOVPS
IP
IPCC
IPM
IPN
ISO
LAI
LOD
LWC
LWCA
MAA
MAACS
MADPro
MAQSIP
MCM
MHVOC
Harvard-EPA Annular Denuder System
Health Effects Institute
Harvard Impactor
high volume sampler
Harvard personal environmental monitor
high temperature gas chromotography/mass spectrometry
heating, ventilation, or air conditioning
ion chromatography
inductively coupled plasma
Integrated Forest Study
Interagency Monitoring of Protected Visual Environments
instrumental neutron activation analysis
integrated organic vapor/particle sampler
inhalable particle
Intergovernmental Panel on Climate Change
inhalable particulate matter
Inhalable Particulate Network
International Standards Organization
leaf area indices
level of detection
liquid water content
liquid water content analyzer
mineral acid anion
Metropolitan Acid Aerosol Characterization Study
Mountain Acid Deposition Program
Multiscale Air Quality Simulation Platform
mass concentration monitor
more highly volatile organic compound
I-xlix
-------
MDL
MOUDI
MS
MSA
MSA
mv
NAAQS
NAMS
NAPAP
NARSTO
NCEA
NDDN
NERL
NESCAUM
NFRAQS
NHAPS
NIOSH
NIR
NIST
NOX
Nr
NRC
NuCM
OAQPS
OC
ORD
PAH
PAN
minimum detection level
microorifice uniform deposit impactor; multiple orifice uniform deposit
impactor
mass spectroscopy
methane sulfonic acid
metropolitan statistical area
motor vehicle
National Ambient Air Quality Standards
National Ambient Monitoring Stations
National Acid Precipitation Assessment Program
North American Research Strategy for Tropospheric Ozone
National Center for Environmental Assessment
National Dry Deposition Network
National Exposure Research Laboratory
Northeast States for Coordinated Air Use Management
North Frontal Range Air Quality Study
National Human Activity Pattern Survey
National Institute for Occupational Safety and Health
near infared radiation
National Institute of Standards and Technology
nitrogen oxides
reactive nitrogen
National Research Council
nutrient cycling model
Office of Air Quality Planning and Standards
organic carbon
Office of Research and Development
polycyclic aromatic hydrocarbon; polynuclear aromatic hydrocarbon
peroxyacetyl nitrate
1-1
-------
PAR
PEL
PBPs
PBW
PBY
PC
PC
PC-BOSS
PCA
PCBs
PCDD
PCDF
PCM
pDR
PEM
PESA
PIXE
PM
PM25
PM10
PM10.25
PM90
PM ACQD
PMF
POP
PRB
PTEAM
PTEP
photosynthetically active radiation
planetary boundary layer
primary biological particles
particle-bound water
Presbyterian Home
particle concentrator
pyrolitic carbon
Particulate Concentrator-Brigham Young University Organic Sampling
System
principal component analysis
polychloronated biphenyls
polychlorinated dibenzo-/>-dioxins
polychlorinated dibenzofurans
particle composition monitor
personal DataRAM
personal exposure monitor
proton (or particle) elastic scattering analysis
proton (or particle) induced X-ray emission
particulate matter
fine particulate matter
combination of coarse and fine particulate matter
coarse particulate matter
90th percentile difference in concentration
Air Quality Criteria Document for Particulate Matter
positive matrix factorization
persistent organic pollutant
policy-relevant background
Particle Total Exposure Assessment Methodology
PM10 Technical Enhancement Program
I-li
-------
PTFE
PUF
PWC
r
RAMS
RAMS
RAPS
RAS
RCS
RFC
RH
RIVM
ROS
RPM
RPM
RRMS
RSP
RSP
RTI
RTF
RUE
SA
SAB
SCAQS
SCENES
SCOS97
sd, SD
sec
polytetrafluoroethylene (Teflon)
polyurethane foam
precipitation-weighted concentrations
Pearson correlation coefficient
Regional Air Monitoring Study
real-time total ambient mass sampler
Regional Air Pollution Study
Roll-Around System
Random Component Superposition
residual fuels oils
relative humidity
Dutch National Institute of Public Health and the Environment
reactive oxygen species
reconstructed particulate mass
respirable particulate matter
relatively remote monitoring sites
respirable particulate matter
respirable suspended particles
Research Triangle Institute
Research Triangle Park
radiation use efficiency
Sierra Anderson
Science Advisory Board
Southern California Air Quality Study
Subregional Cooperative Electric Utility, Department of Defense,
National Park Services, and Environmental Protection Agency Study
1997 Southern California Ozone Study
standard deviation
secondary
I-lii
-------
SEM
SES
SEV
SHEDS
SIP
SLAMS
SMPS
SMSAs
SoCAB
SOPM
SOX
SRM
STN
SUVB
SVOC
SVM
S-XRF
TAR
TC
TDMA
TEO
TEOM
THEES
TMO
TNF
TO
TOFMS
TOR
scanning electron microscopy
sample equilibration system
Sensor Equivalent Visibility
Stochastic Human Exposure and Dose Simulation
State Implementation Plan
State and Local Air Monitoring Stations
scanning mobility particle sizer
Standard Metropolitan Statistical Areas
South Coast Air Basin
secondary organic particulate matter
sulfur oxides
standard reference method
Speciation Trends Network
solar ultraviolet B radiation
semivolatile organic compound
semivolatile material
synchrotron induced X-ray fluoroescence
Third Assessment Report
total carbon
Tandem Differential Mobility Analyzer
trace element oxides
tapered element oscillating microbalance
Total Human Environmental Exposure Study
thermal manganese oxidation
tumor necrosis factor
thermal-optical
time-of-flight mass spectroscopy
thermal-optical reflectance
I-liii
-------
TOT
TPM
TRM
TRXRF
TSP
TVOC
UAM-IV
UCM
UNEP
URG
VAPS
VMD
VOC
WMO
VR
WINS
WRAC
WTA
WTP
XAD
XRF
thermal-optical transmission
thoracic paniculate matter
traffic-related microenvironment
total reflection X-ray fluorescence
total suspended particulate
total volatile organic compounds
Urban Airshed Model Version IV
unresolved complex mixture
United Nations Environment Programme
University Research Glassware
Versatile Air Pollution Sampler
volume mean diameter
volatile organic compound
World Meteorological Organization
visual range
Well Impactor Ninety-Six
Wide Range Aerosol Classifier
willing to accept
willing to pay
polystyrene-divinyl benzene
X-ray fluorescence
I-liv
-------
1. INTRODUCTION
This document is an update of "Air Quality Criteria forF'articulate Matter" published by
the U.S. Environmental Protection Agency (EPA) in 1996, and it will serve as the basis for
Congress!onally-mandated periodic review by EPA of the National Ambient Air Quality
Standards for Particulate Matter (PM NAAQS). The present document critically assesses the
latest scientific information relative to determining the health and welfare effects associated with
the presence of various concentrations of PM in ambient air. It builds upon the assessment in the
previous 1996 EPA Air Quality Criteria Document for Particulate Matter (1996 PM AQCD; U.S.
Environmental Protection Agency, 1996a) by focusing on assessment and integration of
information most relevant to PM NAAQS criteria development, based on pertinent peer-
reviewed literature published or accepted for publication mainly through 2002, as well as a few
more recent important studies. This introductory chapter presents (1) background information
that summarizes legislative requirements, the criteria and NAAQS review process, the history of
PM NAAQS reviews including the chronology of changes in key elements of the standards, and
the coordinated PM research program that has guided much of the more recent research in this
area; (2) an overview of the current PM criteria and NAAQS review (including key milestones),
as well as assessment approaches and procedures used in preparing this document; and (3) an
orientation to the general organizational structure of this document.
1.1 BACKGROUND
1.1.1 Legislative Requirements
As indicated in U.S. Code (1991), Sections 108 and 109 of the U.S. Clean Air Act (CAA)
(42 U.S.C. Sections 7408 and 7409) govern the establishment, review, and revision of NAAQS.
Section 108(a) directs the EPA Administrator to list pollutants, which, in the Administrator's
judgment, cause or contribute to air pollution that may reasonably be anticipated to endanger
public health or welfare and to issue air quality criteria for them. The air quality criteria are to
1-1
-------
reflect the latest scientific information useful in indicating the kind and extent of all identifiable
effects on public health and welfare that may be expected from the presence of the pollutant in
ambient air.
Section 109 directs the Administrator of EPA to propose and promulgate "primary" and
"secondary" NAAQS for criteria air pollutants listed under Section 108. Section 109(b)(l)
defines a primary standard as a level of air quality, the attainment and maintenance of which,
in the judgment of the Administrator, based on the criteria and allowing for an adequate margin
of safety, is requisite to protect the public health. Section 109(b)(2) defines a secondary standard
as one which, in the judgment of the Administrator, based on the criteria, is requisite to protect
public welfare from any known or anticipated adverse effects associated with the presence of the
pollutant in ambient air. Welfare effects, defined in Section 302(h), include (but are not limited
to) effects on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather,
visibility, and climate; damage to and deterioration of property; hazards to transportation; and
effects on economic values, personal comfort, and well-being. Section 109(d)(l) requires the
periodic review and, as appropriate, revision of existing criteria and standards. Also,
Section 109(d)(2) requires an independent committee of non-EPA experts, the Clean Air
Scientific Advisory Committee (CASAC), to provide advice and recommendations to the EPA
Administrator regarding the scientific soundness and appropriateness of criteria and NAAQS
for PM and other criteria air pollutants (i.e., ozone, nitrogen dioxide, sulfur oxides, carbon
monoxide, lead).
1.1.2 Criteria and NAAQS Review Process
The EPA's periodic reviews of criteria and NAAQS for a given criteria air pollutant
progress through a number of steps, beginning with the preparation of an Air Quality Criteria
Document (AQCD) by EPA's National Center for Environmental Assessment Division in
Research Triangle Park, NC (NCEA-RTP), which critically assesses the scientific information
upon which the NAAQS are to be based. Building upon the AQCD, staff within EPA's Office of
Air Quality Planning and Standards (OAQPS) prepare a Staff Paper that evaluates the policy
implications of the key studies and scientific information contained in the AQCD and presents
1-2
-------
staff conclusions and recommendations of options for the Administrator to consider. The Staff
Paper is intended to help "bridge the gap" between the scientific assessment contained in the
AQCD and the judgments required of the Administrator in determining whether it is appropriate
to retain or to revise the NAAQS. Iterative drafts of both of the AQCD and the Staff Paper
(as well as any other analyses supporting the Staff Paper, such as exposure and/or risk
assessments) are made available for public comment and are reviewed by CASAC. Following
CASAC review of these documents, the Administrator proposes decisions on whether to retain
or revise the NAAQS based on the information in these documents, taking into account CASAC
advice and recommendations and public comments. The Administrator's proposed decisions are
published in the Federal Register', with a preamble that presents the rationale for the decisions
and solicits public comment. After taking into consideration comment received on the proposed
decisions, the Administrator's final decisions are promulgated in a final Federal Register notice
that addresses significant comments received on the proposal.
Decisions on the NAAQS involve consideration of the four basic elements of a standard:
indicator, averaging time, form, and level. The indicator defines the pollutant to be measured in
the ambient air for the purpose of determining compliance with the standard. The averaging
time defines the time period over which air quality measurements are to be obtained and
averaged, considering evidence of effects associated with various time periods of exposure.
The form of a standard defines the air quality statistic that is to be compared to the level of the
standard (i.e., an ambient concentration of the indicator pollutant) in determining whether an
area attains the standard. The form of the standard specifies the air quality measurements that
are to be used for compliance purposes (e.g., the 98th percentile of an annual distribution of
daily concentrations; the annual arithmetic average), the monitors from which the measurements
are to be obtained (e.g., one or more population-oriented monitors in an area), and whether the
statistic is to be averaged across multiple years. These basic elements of a standard are the
primary focus of the staff conclusions and recommendations in the Staff Paper and in the
subsequent rulemaking, building upon the policy-relevant scientific information assessed in the
AQCD and on the policy analyses contained in the Staff Paper. These four elements taken
together determine the degree of public health and welfare protection afforded by the NAAQS.
1-2
-------
1.1.3 History of Earlier PM Criteria and NAAQS Reviews
Selection of appropriate indicator(s) for the PM NAAQS has long posed a unique
challenge, in that unlike the other criteria air pollutants, PM is the generic term for a broad class
of physically and chemically diverse substances that exist in ambient air as discrete particles
(liquid droplets or solids) over a wide range of sizes. These airborne particles originate from a
variety of stationary and mobile sources. Primary particles are emitted directly into ambient air;
whereas secondary particles are formed in the atmosphere by transformation of gaseous
emissions such as sulfur oxides (SOX), nitrogen oxides (NOX), and volatile organic compounds
(VOCs). The physical and chemical properties of PM vary greatly with time, region,
meteorology, and source category, thus complicating assessment of ambient PM health and
welfare effects. Particles in ambient air are usually distributed bimodally in two somewhat
overlapping size categories: (1) fine (diameter generally < 2.5 jim) and (2) coarse (diameter
generally > 1.0 jim). Particles in these two size fractions tend to differ in terms of formation
mechanisms, sources of origin, composition, and behavior in the atmosphere and human
respiratory tract.
EPA first promulgated primary and secondary NAAQS for PM on April 30, 1971 (Federal
Register, 1971). These standards measured PM as "total suspended particulate" (TSP), which
refers to ambient PM up to a nominal size of 25 to 45 |im. The primary standards for PM
(measured as TSP) were 260 |ig/m3 (24-h average), not to be exceeded more than once per year,
and 75 |ig/m3 (annual geometric mean). The secondary standard (measured as TSP) was
150 |ig/m3 (24-h average), not to be exceeded more than once per year.
EPA completed the next review of PM air quality criteria and standards in July 1987,
revising the 1971 standards to protect against adverse health effects of inhalable airborne
particles that can be deposited in the lower (thoracic) regions of the human respiratory tract, with
"PM10" (i.e., those particles collected by a sampler with a specified penetration curve yielding an
upper 50% cut-point of 10-|im aerodynamic diameter) as the indicator (Federal Register, 1987).
EPA established identical primary and secondary PM10 standards for two averaging times:
150 |ig/m3 (24-h average, with no more than one expected exceedance per year) and 50 |ig/m3
(expected annual arithmetic mean, averaged over three years).
1-4
-------
1.1.4 The 1997 PM NAAQS Revision
The last previous review of the air quality criteria and standards for PM was initiated in
April 1994 by EPA announcing its intention to develop a revised Air Quality Criteria for
Particulate Matter. Several workshops were held in November 1994 and January 1995 to
discuss important new health effects information useful in preparing initial PM AQCD draft
materials. Also, plans for review of the PM criteria and standards under a highly accelerated,
court-ordered schedule were presented by EPA at a CAS AC public meeting in December 1994.
A court order entered in American Lung Association v. Browner, CIV-93-643-TUC-ACM (U.S.
District Court of Arizona, 1995), as subsequently modified, required publication of EPA's final
decision on the review of the PM NAAQS by July 19, 1997.
Several external review drafts of the revised PM AQCD were prepared by NCEA-RTP and
made available for public comment and CASAC review (at public meetings held in August
1995, December 1995, and February 1996). The CASAC completed its review of the PM
AQCD in early 1996, advising the EPA Administrator in a March 15, 1996 letter (Wolff, 1996)
that "although our understanding of the health effects of PM is far from complete, a revised
Criteria Document which incorporates the Panel's latest comments will provide an adequate
review of the available scientific data and relevant studies of PM." Revisions made in response
to public and CASAC comments were then incorporated by NCEA-RTP, as appropriate, into the
final 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a). The associated PM
Staff Paper, prepared by OAQPS staff, drew upon the 1996 PM AQCD and other assessments to
pose options for the Administrator to consider in making PM NAAQS decisions. Drafts of the
PM Staff Paper also underwent public comment and CASAC review, which informed the final
1996 version (U.S. Environmental Protection Agency, 1996b).
The 1996 PM AQCD and PM Staff Paper provide detailed information on atmospheric
formation, ambient concentrations, and health effects of ambient air PM, as well as quantitative
estimates of human health risks associated with exposure to ambient PM. The main focus of
these documents was on recent epidemiologic evidence reporting associations between ambient
concentrations of PM and a range of serious health effects. Special attention was given to
several size-specific classes of particles, including PM10 and the principal fractions of PM10,
1-5
-------
referred to as the fine (PM2 5) and coarse (PM10_2 5) fractions. PM2 5 refers to those particles
collected by samplers having penetration curves yielding an upper 50% cut point of 2.5 jim
aerodynamic diameter. PM10_2 5 refers to those particles in an aggregate sample having an upper
50% cut point of 10 |im and a lower 50% cut point of 2.5 jim aerodynamic diameter. In other
words, the coarse fraction (PM10_25) refers to inhalable particles that remain if fine (PM25)
particles are removed from a sample of PM10 particles. As discussed in the 1996 PM AQCD,
fine and coarse fraction particles can be differentiated by their sources and formation processes,
by their chemical and physical properties, and by their behavior in the atmosphere.
Taking into account information and assessments presented in the 1996 PM AQCD and
Staff Paper, CAS AC advice and recommendations, and public comments received on proposed
revisions to the PM NAAQS published in December 1996 (Federal Register, 1996), the EPA
Administrator promulgated significant revisions to the PM NAAQS in July 1997 (Federal
Register, 1997). In that decision, the PM NAAQS were revised in several respects. While it was
determined that the PM NAAQS should continue to focus on particles less than or equal to
10 jim in diameter, it was also determined that the fine and coarse fractions of PM10 should be
considered separately. New standards were added, using PM2 5 as the indicator for fine particles;
and PM10 standards were retained for the purpose of regulating coarse-fraction particles. Two
new PM25 standards were set: an annual standard of 15 |ig/m3, based on the 3-year average of
annual arithmetic mean PM2 5 concentrations from single or multiple community-oriented
monitors; and a 24-h standard of 65 |ig/m3, based on the 3-year average of the 98th percentile of
24-h PM2 5 concentrations at each population-oriented monitor within an area. To continue to
address coarse-fraction particles, the annual PM10 standard was retained, and the form, but
not the level, of the 24-h PM10 standard was revised to be based on the 99th percentile of 24-h
PM10 concentrations at each monitor in an area. The secondary standards were revised by
making them identical in all respects to the PM2 5 and PM10 primary standards.
Following promulgation of the revised PM NAAQS, legal challenges were filed by a large
number of parties, addressing a broad range of issues. In May 1998, the U.S. Court of Appeals
for the District of Columbia Circuit issued an initial opinion that upheld EPA's decision to
establish fine particle standards, holding that such standards were amply justified by the growing
1-6
-------
body of empirical evidence demonstrating a relationship between fine particle pollution and
adverse health effects. American Trucking Associations v. Browner. 175 F. 3d 1027, 1055-56
(D.C. Cir. 1999a) (rehearing granted in part and denied in part, 195 F. 3d 4 (D.C. Cir. 1999b),
affirmed in part and reversed in part, Whitman v. American Trucking Associations. 531 U.S. 457
(2001). Further, the court found "ample support" for EPA's decision to regulate coarse fraction
particles, although it vacated the revisions to the 1987 PM10 standards on the basis of PM10 being
a "poorly matched indicator for coarse particulate pollution" because PM10 includes fine
particles (Id. at 1053-55). As a result of this aspect of the court's ruling, which EPA did not
appeal, the 1987 PM10 standards remain in effect.
In addition, the court broadly held that EPA's approach to establishing the level of the
standards in its 1997 decisions on both the PM and Ozone NAAQS (which were promulgated on
the same day and considered together by the court in this aspect of its opinion) effected "an
unconstitutional delegation of legislative authority" (Id. at 1034-40). EPA appealed this aspect
of the court's ruling to the U.S. Supreme Court. In February 2001, the U.S. Supreme Court
unanimously reversed the Court of Appeals' ruling on the constitutional issue and sent the case
back to the Court of Appeals for resolution of any remaining issues that had not been addressed
in that court's earlier rulings. Whitman v. American Trucking Associations. 531 U.S. 457,
475-76 (2001). In March 2002, the Court of Appeals rejected all remaining challenges to the
standards, finding that the 1997 PM25 standards were reasonably supported by the record and
were not "arbitrary or capricious." American Trucking Associations v. EPA. 283 F. 3d 355,
369-72 (D.C. Cir. 2002). Thus, the 1997 PM25 standards also remain in effect.
1.1.5 Coordinated PM Research Program
Shortly after promulgation of the 1997 PM NAAQS decisions, NCEA-RTP published a
PM Health Risk Research Needs Document (U.S. Environmental Protection Agency, 1998a) that
identified research needed to improve scientific information supporting future reviews of the
PM NAAQS. The document provided a foundation for PM research coordination among Federal
agencies and other research organizations, as well as input to National Research Council (NRC)
deliberations on PM research. The Office of Research and Development (ORD) of EPA also
1-7
-------
moved quickly to broaden its ongoing PM research activities by developing, in partnership with
other Federal agencies, a coordinated interagency PM research program. This interagency
program has focused and continues to focus mainly on expanding scientific knowledge of
ambient PM exposure and health effects, as well as on developing improved monitoring methods
and cost-effective mitigation strategies. The interagency effort also promotes substantially
expanded coordination with other research organizations, including the Health Effects Institute
(HEI) and other state-, university-, and industry-sponsored research groups. Beginning in the
fall of 1997, public participation was and continues to be encouraged through workshops and
review of EPA's PM Research Program documentation.
In response to Congressional requirements in EPA's Fiscal Year 1998 Appropriation, the
NRC established its Committee on Research Priorities for Airborne Particulate Matter in January
1998. This NRC PM Research Committee's charge was to identify the most important research
priorities relevant to setting primary (health-based) PM NAAQS, to develop a conceptual plan
for PM research, and to monitor EPA's research progress toward improved understanding of the
relationship between PM and public health. The NRC PM Research Committee issued a series
of reports (National Research Council, 1998, 1999, 2001) which recommended that expanded
PM research efforts be planned and carried out in relation to a general conceptual framework,
as shown in Figure 1-1. That framework essentially calls for research aimed at (a) identifying
sources of airborne particles or gaseous precursor emissions and characterization of processes
involved in atmospheric transformation, transport, and fate of ambient PM; (b) delineating
temporal and spatial patterns of air quality indicators (e.g., PM25, PM10_25, PM10 mass
concentrations) of ambient PM and apportionment of observed variations in such ambient PM
indicators to various emission sources; (c) characterizing human exposures to ambient PM as
one important component of total personal exposure to particles, as modified by time-activity
patterns and varying microenvironmental exposure to particles of indoor or ambient origin;
(d) characterizing resulting respiratory tract deposition, clearance, retention, and disposition of
inhaled particles, as determinants of dose to target tissues (e.g., locally in the lungs or via
systemic translocation to the heart or other organs); and (e) delineating of mechanisms of
damage and repair plausibly leading to (f) human health responses, as extrapolated from or
-------
Sources of Airborne
Particulate Matter or
Gaseous Precursor
Emissions
i
->
k.
Indicator in Ambient
(Outdoor) Air
(e.g. Mass
Concentration)
i
->
k
Personal
Exposure
t
>
k
Dose to
Target
Tissues
i
>
k
Human
Health
Response
Mechanisms determining emissions,
chemical transformation (including
Human time-activity
patterns, indoor (or
Deposition,
Mechanisms of
formation of secondary particles from microenviornmental)
gaseous precursors), and
transport In air
sources and sinks of
particulate matter
clearance, retention damage and repair
and disposition of
particulate matter
presented to
an individual
Figure 1-1. A general framework for integrating PM 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).
quantified by experimental animal or human exposure (toxicology) studies and/or observational
(epidemiology) studies.
Research conducted under a PM Research Program structured in relation to the conceptual
framework shown in Figure 1-1 would be expected to (a) reduce key scientific uncertainties
regarding interrelationships between PM sources, ambient concentrations, exposures, dose to
target tissues, and resulting health effects and thereby (b) improve the scientific underpinnings
for both current and future periodic PM NAAQS reviews. Table 1-1 highlights some types of
key uncertainties identified by the NRC PM Research Committee in relation to elements of the
source-to-response conceptual framework illustrated in Figure 1-1. The NRC Committee went
on to delineate a series of 10 research topics that they recommended be addressed by an
expanded PM research program aimed at answering a set of broadly stated questions, as shown
in Table 1-2.
The EPA PM Research Program, structured to address topics shown in Table 1-2, has
encompassed studies to improve understanding of the formation and composition of fine PM,
to improve the measurement and estimation of population exposures to ambient PM, to delineate
characteristics or components of PM that are responsible for its health effects, to evaluate factors
that increase susceptibility to PM effects in some subpopulations, and to elucidate mechanisms
1-9
-------
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 particulate matter concentrations
• Relative contribution of various sources to the most toxic components of particulate matter
Concentration (indicator) ^ Exposure
• Relationship between ambient (indoor) particulate matter and the composition of particles to
which people are exposed
• Contribution of ambient particulate matter to total personal exposure for:
- Susceptible subpopulations
- General population
• Variation in relationship of ambient particulate matter concentrations to human exposure by place
• Variation in contribution of ambient particulate matter concentrations to total human exposure
over time
• Covariance of particulate matter exposures with exposures to other pollutants
• Relationships between outdoor ambient and personal exposures for particulate matter
and co-pollutants
Exposure ^ Dose
• Relationship between inhaled concentration and dose of particulate 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 particulate matter dose to or via the lungs
- Inflammation
~ Host defenses
~ Neural mechanisms
Source: National Research Council (2001).
1-10
-------
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
co-pollutants 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 epidemiologic
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 epidemiologic studies
influence estimates of health risks from exposures to particulate matter? Can existing methods be
improved? What is the effect of measurement error and misclassiflcation on estimates of the association
between air pollution and health?
Source: National Research Council (2001).
1-11
-------
by which these effects are produced. The results from these efforts, and 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.
1.2 CURRENT PM CRITERIA AND NAAQS REVIEW
1.2.1 Key Milestones and Procedures for Document Preparation
As with other NAAQS reviews, critical assessment of relevant scientific information is
presented in this updated PM AQCD. Key milestones in the development of the present PM
AQCD are shown in Table 1-3 and discussed below. It is important to note that development of
the document involved substantial external peer and public review through (a) public workshops
involving the general aerosol scientific community, (b) iterative reviews of successive drafts by
CASAC, and (c) comments from the public on successive drafts. This final document reflects
extensive external input received through these reviews and serves to ensure that the review of
the PM standards is based on critical assessment of the latest available pertinent science.
The procedures for developing this updated PM AQCD built on the experience derived
from the most recent previous PM, Ozone, and CO AQCD preparation efforts. Briefly, the
respective responsibilities for production of the present PM AQCD are as follows. An NCEA-
RTP PM Team was responsible for developing and implementing a project plan for preparing the
PM AQCD, taking into account inputs from individuals in other EPA program and policy offices
identified as part of the EPA PM Work Group. The resulting project plan (i.e., the PM
Document Development Plan) was discussed with CASAC in May 1998 and was appropriately
revised. A literature search was ongoing throughout the preparation of this document to identify,
to the extent possible, all pertinent PM literature published since early 1996. Additionally, EPA
published in the Federal Register (1) a request for information asking for recently available
research information on PM that may not yet be published and (2) a request for individuals with
the appropriate type and level of expertise to contribute to the writing of PM AQCD materials to
identify themselves (U.S. Environmental Protection Agency, 1998b). The specific authors of
chapters or sections of the proposed document were selected on the basis of their expertise on
1-12
-------
TABLE 1-3. KEY MILESTONES IN THE DEVELOPMENT OF THIS DOCUMENT
Key Milestones
Dates
PM NAAQS Review Plan to CASAC
Prepare AQCD Development Plan
Begin Literature Search
Federal Register Call for Information/Sources Sought
CASAC Meeting on PM AQCD Development Plan
Prepare Workshop Drafts of Chapters
Peer Review Workshop
Prepare First External Review Draft PM AQCD
First External Review Draft PM AQCD
Public Comment Period on First Draft
CASAC Meeting on First Draft
Second External Review Draft PM AQCD
Public Comment Period on Second Draft
CASAC Meeting on Second Draft
Third External Review Draft PM AQCD
Public Comment Period on Third Draft
CASAC Meeting on Third Draft
Fourth External Review Draft PM AQCD
Public Comment Period on Fourth Draft
CASAC Meeting on Fourth Draft
CASAC Consultation on Proposed Revisions to Chapter 9
CASAC Consultation on Revisions to Chapters 7 and 8
Revised Draft Chapters 7 and 8
Public Comment Period on Revised Draft Chapters 7 and 8
CASAC Teleconference Review of Revised Draft Chapters 7 and 8
Revised Draft Chapters 7, 8 and 9
Public Comment Period on Revised Draft Chapters 7, 8 and 9
CASAC Meeting on Revised Draft Chapters 7, 8 and 9
Revised Draft Chapter 9
Public Comment Period on Revised Chapter 9
CASAC Review of Revised Chapter 9
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
June 2003
June to August 2003
August 2003
October 2003
November 2003
December 2003
January 2004
February 2004
June 2004
July 2004
July 2004
August 2004
September 2004
September 2004
October 2004
1-13
-------
the subject areas and their familiarity with the relevant literature; and these included both EPA
and non-EPA scientific experts. The project team defined critical issues and topics to be
addressed by the authors and provided direction in order to emphasize evaluation of those
studies most clearly identified as important for standard setting. It should be noted that materials
contributed by non-EPA authors were, at times, modified by EPA PM Team staff in response to
internal and/or external review comments, e.g., by the public or CASAC, and that EPA is
responsible for the ultimate content of this PM AQCD.
The main focus of this document is the evaluation and interpretation of pertinent
atmospheric science information, air quality data, human exposure information, and health and
welfare effects information newly published since that assessed in the 1996 PM AQCD. Initial
draft versions of AQCD chapters were evaluated via expert peer-review workshop discussions
and/or written peer reviews that focused on the selection of pertinent studies to be included in
the chapters, the potential need for additional information to be added to the chapters, and the
quality of the characterization and interpretation of the literature. The authors of the draft
chapters then revised them on the basis of workshop and/or written expert review comments.
These and other integrative materials were incorporated into the First External Review Draft of
the PM AQCD (October 1999), which was made available for public comment and was the
subject of consultation with CASAC at a December 1999 public meeting.
In order to foster timely presentation and publication of newly emerging PM research
findings, EPA cosponsored and helped to organize an Air and Waste Management Association
International Speciality Conference, entitled "PM 2000: Paniculate Matter and Health," held in
January 2000 in Charleston, SC. The conference was cosponsored in cooperation with several
other government agencies and/or private organizations that also fund PM research. Topics
covered included new research results concerning the latest advances in PM atmospheric
sciences (e.g., PM formation, transport, transformation), PM exposure, PM dosimetry and
extrapolation modeling, PM toxicology (e.g., mechanisms, laboratory animal models, human
clinical responses), and PM epidemiology. The main purpose of the conference was to facilitate
having the latest scientific information available in time for incorporation as quickly as possible
into the Second External Review Draft of the PM AQCD. Hence, arrangements were made for
1-14
-------
scientists to submit written manuscripts on papers or posters presented at the PM 2000
Conference for expedited peer-review by several major journals, so that decisions on acceptance
for publication could be made by mid-2000. The evaluations and findings set forth in the
Second External Review Draft (March 2001) included consideration of such published PM 2000
papers, as well as extensive additional information published elsewhere since that addressed in
the First External Review Draft. The Second External Review Draft was made available for
public comment and was reviewed by CAS AC at a July 2001 public meeting.
Further revisions were then incorporated into the Third External Review Draft (April 2002)
to reflect both public comment and CAS AC review of the Second Draft, as well as assessment of
additional extensive new information published since that addressed in the Second Draft.
Shortly after EPA released the Third External Review Draft in May 2002 for public comment
and CASAC review, the HEI announced that researchers at Johns Hopkins University had
discovered problems with applications of statistical software used in a number of important
epidemiologic studies on links between ambient air PM and mortality and morbidity effects.
The Third External Review Draft was reviewed by CASAC at a July 2002 public meeting,
although it was recognized that discussion of the epidemiology information would need to be
further revised after the newly surfaced statistical issues were appropriately addressed.
In response to the surfacing of such statistical issues, which affected numerous PM time-series
studies that used General Additive Models (GAM) and were published post-1995, EPA took
steps in consultation with CASAC to identify particularly policy-relevant studies and to
encourage researchers to reanalyze affected studies and to submit them expeditiously for peer
review by a special expert panel convened by HEI. The results of reanalyses for more than
30 studies were peer-reviewed and published, along with commentary by the FIEI expert peer-
review panel, in an FIEI Special Report (Health Effects Institute, May 2003).
Discussion of the newly addressed statistical issues and reanalyses results was incorporated
into the Fourth External Review Draft (June 2003), which was made available for public
comment and was reviewed by CASAC at an August 2003 public meeting. The Fourth Draft
also incorporated changes made in response to earlier public comments and CASAC reviews,
including pertinent peer-reviewed literature published or accepted for publication mainly
1-15
-------
through April 2002. CAS AC completed its review of Chapters 1 through 6 of the Fourth Draft
PM AQCD at the August 2003 meeting, but comments on the remaining chapters (7, 8, and 9)
were judged to warrant further revision and review of those chapters. EPA subsequently
consulted with CASAC and made available revised Chapters 7, 8, and 9 for public comment and
CAS AC review, as indicated in Table 1-3. The CASAC completed its review in September,
2004, as indicated in an October 4, 2004 letter (Hopke, 2004) to the EPA Administrator.
Revisions made in response to final CASAC and public comments were then incorporated by the
NCEA-RTP PM team, as appropriate, into this final 2004 PM AQCD (U.S. Environmental
Protection Agency, 2004) released in October, 2004.
1.2.2 Assessment Approaches
The assessment presented in this document is framed by: (1) the selection of pertinent
issues to be addressed; (2) the selection of relevant studies and an approach to the presentation of
information drawn from those studies; and (3) the selection of an approach to interpreting and
integrating the body of evidence evaluated in the document.
As an initial matter, the NCEA-RTP PM team focused on selecting pertinent issues to be
addressed in this assessment. Preliminary issues were drawn from among those highlighted in
the 1996 PM AQCD and Staff Paper and in CASAC and public reviews of those documents,
the 1997 PM NAAQS promulgation process, and the 1998 PM Health Risk Research Needs
Document (U.S. Environmental Protection Agency, 1998a). Further identification and
refinement of issues resulted from the NRC review and reports on PM research priorities,
as discussed in Section 1.1.5 above. The CASAC review of the PM AQCD Development Plan
and public comments on early draft AQCD materials at various stages of their development also
provided important inputs regarding issue identification. The issues selected are reflected
throughout this document and are most concisely identified in the introductory section at the
beginning of each of the ensuing chapters.
The selection of relevant studies to be included in this assessment was based on a detailed
review of new research published in the peer-reviewed literature since early 1996, including
materials accepted for publication mainly through April 2002 (and, thus, appearing mostly
1-16
-------
during 2002). Limited coverage of some more recent studies is also included as deemed
appropriate in light of their special importance. For example, information derived from the HEI
Special Report (Health Effects Institute, May 2003), discussed above in Section 1.2.1, has been
integrated into this assessment. Literature discussed in detail in the 1996 PM AQCD (U.S.
Environmental Protection Agency, 1996a) generally is not discussed in depth in this document;
rather key findings from the 1996 review are concisely summarized as appropriate. Limited
treatment is included, however, for some earlier studies judged to be of particular importance in
this review of the PM NAAQS. Key literature is summarized in tables, whereas text discussions
focus mainly on evaluation and integration of the literature, including discussion of alternative
points of view where scientific controversy exists. This approach to study selection and
presentation reflects recommendations from CASAC aimed at development of a more concise
document than the 1996 PM AQCD.
The scientific assessments presented in this document are primarily driven by the large
body of available epidemiologic evidence evaluating associations between ambient PM, alone
and in combination with other air pollutants, and various health endpoints. In such a case, an
approach to interpreting the evidence most fundamentally focuses on characterizing the causal
significance of observed associations. In so doing, it is appropriate to consider various aspects
of the evidence of associations, in particular drawing from those presented earlier in Hill's
classic monograph (Hill, 1965) and extensively used by the scientific community in conducting
such evidence-based reviews, e.g., in the 2004 Surgeon General's Report on smoking (Centers
for Disease Control and Prevention, 2004). As discussed in Chapters 8 and 9, a number of these
aspects are judged to be particularly salient in evaluating the body of evidence available in this
review, including the aspects described by Hill as strength, consistency, temporality, biologic
gradient, experiment, plausibility, and coherence. These interrelated aspects are considered in
the evaluation of epidemiologic evidence presented in Chapter 8 and are also more broadly
addressed in the Chapter 9 integrative synthesis.
1-17
-------
1.3 DOCUMENT ORGANIZATION
This document is basically organized to assess information related to topics along the same
flow of issues presented in the NRC conceptual framework shown in Figure 1-1, including
information related to effects on both human health and the environment. The document
consists of nine chapters presented in two volumes. Volume I contains this general introduction
(Chapter 1), as well as Chapters 2 through 5. Chapters 2 and 3 provide background information
on the physical and chemical properties of PM and related compounds; sources and emissions;
atmospheric transport, transformation, and fate of PM; methods for the collection and
measurement of PM; and U.S. ambient air PM concentrations. Chapter 4 assesses welfare-
related PM effects on vegetation and ecosystems, visibility, and man-made materials, as well as
climate-related effects (including effects on atmospheric transmission of solar radiation), and it
includes limited information on economic impacts of some welfare effects. Chapter 5 discusses
factors affecting exposure of the general population to ambient PM.
The second volume contains Chapters 6 through 9. Chapter 6 assesses information
concerning dosimetry of inhaled particles in the respiratory tract. Chapter 7 assesses the
toxicology of specific types of PM constituents and potential mechanisms of action, based
primarily on laboratory animal studies and controlled human exposure studies. Chapter 8
assesses the epidemiologic literature. Lastly, Chapter 9 integrates key information on PM-
related health effects, drawing from assessments in prior chapters of the literature on exposure,
dosimetry, toxicology, and epidemiology, as well as highlighting key information regarding
important welfare effects associated with ambient PM.
1-18
-------
REFERENCES
Centers for Disease Control and Prevention. (2004) The health consequences of smoking: a report of the Surgeon
General. Atlanta, GA: U.S. Department of Health and Human Services, National Center for Chronic Disease
Prevention and Health Promotion, Office on Smoking and Health. Available:
http://www.cdc.gov/tobacco/sgr/sgr_2004/chapters.htm (18 August, 2004).
Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30)
36: 8186-8201.
Federal Register. (1987) Revisions to the national ambient air quality standards for paniculate matter. F. R. (July 1)
52: 24,634-24,669.
Federal Register. (1996) National ambient air quality standards for paniculate matter; proposed rule. F. R.
(December 13) 61: 65,638-65,713.
Federal Register. (1997) National ambient air quality standards for paniculate matter; final rule. F. R. (July 18)
62: 38,652-38,752.
Health Effects Institute. (2003) Revised analyses of time-series studies of air pollution and health. Boston, MA:
Health Effects Institute; special report.
Hill, A. B. (1965) The environment and disease: association or causation? Proc. R. Soc. Med. 58: 295-300.
Hopke, P. K. (2004) Clean Air Scientific Advisory Committee (CASAC) Paniculate Matter (PM) Review Panel's
ongoing peer review of the agency's draft Air Quality Criteria for Paniculate Matter [letter to Michael O.
Leavitt, Administrator, U.S. EPA]. Washington, DC: U.S. Environmental Protection Agency, Clean Air
Scientific Advisory Committee; EPA-SAB-CASAC-04-009; October 4.
National Research Council. (1998) Research priorities for airborne paniculate matter. I. Immediate priorities and a
long-range research portfolio. Washington, DC: National Academy Press. Available:
http://www.nap.edu/catalog/6131.html (4 June 2003).
National Research Council. (1999) Research priorities for airborne paniculate matter. II. Evaluating research
progress and updating the portfolio. Washington, DC: National Academy Press. Available:
http://www.nap.edu/books/0309066387/html/ (4 June 2003).
National Research Council. (2001) Research priorities for airborne paniculate matter. III. Early research progress.
Washington, DC: National Academy Press. Available: http://www.nap.edu/books/0309073375/html/
(4 June 2003).
U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air quality
standards. U. S. C. 42: §§7408-7409.
U.S. Court of Appeals for the District of Columbia. (1999a) American Trucking Associations, Inc. vs. U.S.
Environmental Protection Agency. 175 F3d 1027 (D.C. Cir. 1999).
U.S. Court of Appeals for the District of Columbia. (1999b) American Trucking Associations, Inc. vs. U.S.
Environmental Protection Agency. 195 F.3d 4 (D.C. Cir. 1999), affirmed in part, reversed in part, and
remanded..
U.S. Court of Appeals for the District of Columbia. (2002) American Trucking Associations, Inc. vs. U.S.
Environmental Protection Agency. 283 F.3d 355, 378-79 (D.C. Cir. 2002).
U.S. District Court of Arizona. (1995) American Lung Association v. Browner. West's Federal Supplement
884 F.Supp. 345 (No. CIV 93-643 TUC ACM).
U.S. Environmental Protection Agency. (1996a) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
U.S. Environmental Protection Agency. (1996b) Review of the national ambient air quality standards for paniculate
matter: policy assessment of scientific and technical information. OAQPS staff paper. Research Triangle
Park, NC: Office of Air Quality Planning and Standards; report no. EPA/452/R-96-013. Available from:
NTIS, Springfield, VA; PB97-115406REB.
U.S. Environmental Protection Agency. (1998a) Paniculate matter research needs for human health risk assessment
to support future reviews of the national ambient air quality standards for paniculate matter. Research
Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
U.S. Environmental Protection Agency. (1998b) Review of national ambient air quality standards for paniculate
matter. Commer. Bus. Daily: February 19. Available: http://cbdnet.access.gpo.gov/index.html
[1999, November 24].
1-19
-------
U.S. Supreme Court. (2001) Whitman v. American Trucking Association. 531 U.S. 457 (nos. 99-1257 and 99-1426).
Wolff, G. T. (1996) Closure by the Clean Air Scientific Advisory Committee (CASAC) on the draft Air Quality
Criteria for Paniculate Matter [letter to Carol M. Browner, Administrator, U.S. EPA]. Washington, DC:
U.S. Environmental Protection Agency, Clean Air Scientific Advisory Committee.; EPA-SAB-CASAC-LTR-
96-005; March 15.
1-20
-------
2. PHYSICS, CHEMISTRY, AND MEASUREMENT
OF PARTICULATE MATTER
Chapter 3 of the 1996 EPA document Air Quality Criteria for Particulate Matter (1996 PM
AQCD; U.S. Environmental Protection Agency, 1996a) contained an extensive review of the
physics and chemistry of airborne parti culate matter (PM). Chapter 2 of this revised version of
the PM AQCD also provides background information on the physics and chemistry of
atmospheric particles, information useful in understanding the subsequent chapters. The
chapters in this document are organized to generally follow the sequence of key elements that
make up the risk assessment framework described in Chapter 1 (Section 1.2.2) beginning with
sources and continuing to effects as shown in Figure 1-1. Thus, this chapter discusses new
background information useful in evaluating PM effects on human health and welfare and in
preparing related risk assessments used to support PM standard-setting decisions. Information
important for implementation of PM standards, but not essential to the standard setting process,
is not the focus in this chapter. The reader is referred to the NARSTO (North American
Research Strategy for Tropospheric Ozone) Fine Particle Assessment (NARSTO, 2003) for
information relevant to air quality management for PM.
Unlike other criteria pollutants (O3, CO, SO2, NO2, and Pb), PM is not a specific chemical
entity but is a mixture of particles from different sources and of different sizes, compositions,
and properties. Emphasis is placed here on discussion of differences between fine and coarse
particles and differences between ultrafine particles and accumulation-mode particles within fine
particles.
Since PM is defined quantitatively by measurement techniques, it will be useful to discuss
our understanding of the relationships between PM suspended in the atmosphere, PM inhaled by
people, and PM measured by various sampling and analytical techniques. Chapter 4 of the 1996
PM AQCD (U.S. Environmental Protection Agency, 1996a) contained a review of the state of
the art of PM measurement technology. Since that time, considerable progress has been made in
understanding problems in the measurement of PM mass, chemical composition, and physical
parameters. Progress has also been made in developing new and improved measurement
techniques, especially for continuous measurements. Therefore, a more extensive survey on
2-1
-------
measurement problems and on newly developed measurement techniques is included here in
Section 2.2. For more detail and older references, the reader is referred to Chapters 3 and 4 of
the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a).
2.1 PHYSICS AND CHEMISTRY OF PARTICULATE MATTER
2.1.1 Basic Concepts
Atmospheric particles originate from a variety of sources and possess a range of
morphological, chemical, physical, and thermodynamic properties. Examples of atmospheric
particles include combustion-generated particles, such as diesel soot or fly ash; photochemically
produced particles, such as those found in urban haze; salt particles formed from sea spray; and
soil-like particles from resuspended dust. Some particles are liquid; some are solid. Others may
contain a solid core surrounded by liquid. Atmospheric particles contain inorganic ions, metallic
compounds, elemental carbon, organic compounds, and crustal compounds. Some atmospheric
particles are hygroscopic and contain particle-bound water. The organic fraction is especially
complex, containing hundreds (probably thousands) of organic compounds. (See Appendix 3C
for information on the composition of the organic fraction and the concentration of specific
organic compounds.) Primary particles are emitted directly from sources; whereas secondary
particles are formed from gases through chemical reactions in the atmosphere involving
atmospheric oxygen (O2) and water vapor (H2O); reactive species such as ozone (O3); radicals
such as the hydroxyl (»OH) and nitrate (»NO3) radicals; and pollutants such as sulfur
dioxide (SO2), nitrogen oxides (NOX), and organic gases from natural and anthropogenic
sources. The particle formation process includes nucleation of particles from low-vapor pressure
gases emitted from sources or formed in the atmosphere by chemical reactions, condensation of
low-vapor pressure gases on existing particles, and coagulation of particles. Thus, any given
particle may contain PM from many sources. Because a particle from a given source is likely to
be composed of a mixture of chemical components and because particles from different sources
may coagulate to form a new particle, atmospheric particles may be considered a mixture of
mixtures. The composition and behavior of particles are fundamentally linked with those of the
surrounding gas. An aerosol may be defined as a suspension of solid or liquid particles in air.
The term aerosol includes both the particles and all vapor or gas phase components of air.
2-2
-------
However, the term aerosol is sometimes used to refer to the suspended particles only. In this
document, "particulate" is used only as an adjective, as in particulate matter.
A complete description of the atmospheric aerosol would include an accounting of the
chemical composition, morphology, and size of each particle, as well as the relative abundance
of each particle type as a function of particle size (Friedlander, 1970). However, the physical
and chemical characteristics of particles are usually measured separately. Size distributions by
particle number used to calculate surface area and volume distributions often are determined by
physical means, such as electrical mobility, aerodynamic behavior, or light scattering. Chemical
composition usually is determined by analysis of collected samples, although some species can
be measured in situ. The mass and average chemical composition of particles segregated
according to aerodynamic diameter by cyclones or impactors can also be determined. However,
recent developments in single particle analysis techniques by electron microscopy with X-ray
analysis of single particles (but not agglomerates) collected on a substrate or by mass
spectroscopy of individual suspended particles provide elemental composition of individual
particles by particle size and, thus, are bringing the description envisioned by Friedlander closer
to reality.
2.1.2 Physical Properties and Processes
2.1.2.1 Definitions of Particle Diameter
The diameter of a spherical particle may be determined by optical or electron microscopy,
by light scattering and Mie theory, by its electrical mobility, or by its aerodynamic behavior.
However, atmospheric particles often are not spherical. Therefore, their diameters are described
by an "equivalent" diameter (i.e., the diameter of a sphere that would have the same physical
behavior). An optical diameter is the diameter of a spherical particle, with the same refractive
index as the particle used to calibrate the optical particle sizer, that scatters the same amount of
light into the solid angle measured. Diffusion and gravitational settling are important physical
behaviors for particle transport, collection, and removal processes, including deposition in the
respiratory tract. Different equivalent diameters are used depending on which process is more
important. For smaller particles, diffusion is more important and the Stokes diameter is often
used. For larger particles, gravitational setting is more important and the aerodynamic diameter
is often used.
-------
The Stokes diameter, Dp, describes particle size based on the aerodynamic drag force
imparted on a particle when its velocity differs from that of the surrounding fluid. For a smooth,
spherically shaped particle, Dp exactly equals the physical diameter of the particle. For
irregularly shaped particles, Dp is the diameter of an equivalent sphere that would have the same
aerodynamic resistance. Electrical mobility analyzers classify particles according to their
electrical mobility. Particles of equal Stokes diameters that carry the same electric charge will
have the same electrical mobility. Hence, for spherical particles, the electrical mobility diameter
would equal the Stokes diameter. The mobility diameter can be considered the diameter of a
spherical particle that would have the same electrical mobility. The particle mobility can be
related to the particle diffusion coefficient and Brownian diffusion velocity through the Stokes-
Einstein equation. Thus, the Stokes diameter is the appropriate parameter for particle behavior
governed by diffusion. The Stokes diameter, Dp, is used in size distributions based on light
scattering and mobility analysis. The Stokes diameter is independent of density.
The aerodynamic diameter, Da, however, depends on particle density. It is defined as the
diameter of a spherical particle with an equal gravitational settling velocity but a material density
of 1 g/cm3. Cascade impactors separate particles based on their aerodynamic diameter, and
aerodynamic particle sizers measure the aerodynamic diameter. Respirable, thoracic, and
inhalable sampling and PM2 5 and PM10 sampling are based on particle aerodynamic diameter.
For particles greater than about 0.5 jim, the aerodynamic diameter is generally the quantity of
interest. For smaller particles, the Stokes diameter may be more useful. Particles with the same
physical size and shape but different densities will have the same Stokes diameter but different
aerodynamic diameters.
The aerodynamic diameter, Da, is related to the Stokes diameter, Dp, by:
(2-D
where p is the particle density, and Cp and Ca are the Cunningham slip factors evaluated for the
particle diameters Dp and Da respectively. The slip factor is a function of the ratio between
particle diameter and mean free path of the suspending gas (0.066 jim for air at one atmosphere
pressure and 20 °C). C is an empirical factor that accounts for the reduction in the drag force on
2-4
-------
particles due to the "slip" of the gas molecules at the particle surface. C is an important factor
for particles less than 1 |_im in diameter, for which the surrounding air cannot be modeled by a
continuous fluid. For large particles (Dp > 5 jim) C = 1; while for smaller particles C > 1.
For particles with diameters greater than the mean free path, A, the aerodynamic diameter
given by equation (2-1) is approximated by:
Da= Dp(p)l/2 (forDp»X) (2-2)
This expression, which shows that aerodynamic diameter is directly proportional to the square
root of the particle density, is often used for particles as small as 0.5 jim. For particles with
diameters much smaller than the mean free path, the slip factor must be taken into account.
In this case, the aerodynamic diameter is directly proportional to the particle density,
(forDp«X) (2-3)
Detailed definitions of the various sizes and their relationships are given in standard aerosol
textbooks (e.g., Friedlander [2000], Reist [1984, 1993], Seinfeld and Pandis [1998], Hinds
[1999], Vincent [1989, 1995], Willeke and Baron [1993], Baron and Willeke [2002], and Fuchs
[1964, 1989]).
2.1.2.2 Aerosol Size Distributions
Particle size, as indexed by one of the "equivalent" diameters, is an important parameter in
determining the properties, effects, and fate of atmospheric particles. The atmospheric
deposition rates of particles and, therefore, their residence times in the atmosphere are a strong
function of their Stokes and aerodynamic diameters. Particle diameters also influence the
deposition patterns of particles within the lung. Because light scattering is strongly dependent
on the optical particle size, the amount of light scattering per unit PM mass will be dependent on
the size distribution of atmospheric particles. Therefore, the effects of atmospheric particles on
visibility, radiative balance, and climate will be influenced by the size distribution of the
2-5
-------
particles. Studies using cascade impactors or cyclones measure the particle-size distribution
directly in aerodynamic diameter. The diameters of atmospheric particles range from 1 nm to
100 jam, spanning 5 orders of magnitude. A variety of different instruments, measuring a variety
of equivalent diameters, are required to cover this range.
Older particle counting studies used optical particle counters to cover the range of 0.3 to
30 |im diameter. Diameters of particles below 0.5 jim were measured as mobility diameters.
The particle diameters used in size distribution graphs from these studies usually are given as
physical or Stokes diameters rather than aerodynamic diameters. In recent years, aerodynamic
particle sizers have been developed that give a direct measurement of the aerodynamic diameter
in the range of approximately 0.7 to 10 |im diameter. These instruments have been used with
electrical mobility analyzers that measure the mobility diameter of particles from 3 nm to
approximately 0.5 jim (McMurry, 2000). Unfortunately, there is no agreed-upon technique for
combining the various equivalent diameters. Some workers use various assumptions to combine
the various measurements into one presentation; others report each instrument separately.
Therefore, the user of size distribution data should be careful to determine exactly which
equivalent diameter is reported.
Particle Size Distribution Functions
The distribution of particles with respect to size is an important physical parameter
governing particle behavior. Because atmospheric particles cover several orders of magnitude in
particle size, size distributions often are expressed in terms of the logarithm of the particle
diameter on the X-axis and the measured differential concentration on the Y-axis:
AN/A(logDp) = the number of particles per cm3 of air having diameters in the size range from
log Dp to log(Dp + ADp). Because logarithms do not have dimensions, it is necessary to think
of the distribution as a function of log(Dp/Dp0), where the reference diameter Dp0 = 1 jim is
not explicitly stated. If AN/A(logDp) is plotted on a linear scale, the number of particles
between Dp and Dp + ADp is proportional to the area under the curve of AN/A(logDp) versus
logDp. Similar considerations apply to distributions of surface, volume, and mass. When
approximated by a function, the distributions are usually given as dN/d(log Dp) rather than
AN/A(logDp).
2-6
-------
Atmospheric Aerosol Size Distributions
Whitby (1978) published an analysis of over 1,000 particle size distributions measured at
various locations in the United States. Figure 2-1 shows the number, surface area, and volume
distributions for the grand average continental size distribution. Volume, surface area, and
number distributions are plotted on an arithmetic scale such that the volume, surface area, or
number of particles in any specified size range is proportional to the corresponding area under
the curve. These distributions show that most of the particles are quite small, below 0.1 jam;
whereas most of the particle volume (and therefore most of the mass) is found in particles
> 0.1 jim. Other averaged atmospheric size distributions are shown in Figures 2-2 and 2-3
(Whitby, 1978; Whitby and Sverdrup, 1980). Figures 2-2a and 2-2b describe the number of
particles as a function of particle diameter for rural, urban-influenced rural, urban, and freeway-
influenced urban aerosols. For some of the same data, the particle volume distributions are
shown in Figures 2-3a and 2-3b. Whitby (1978) observed that the size distributions typically
had three peaks which he called "modes." The entire size distribution could be characterized
well by a trimodal model consisting of three additive log-normal distributions. The mode with a
peak between 5 and 30 jim diameter formed by mechanical processes was called the coarse
particle mode; the mode with a peak between 0.15 and 0.5 jim formed by condensation and
coagulation was called the accumulation mode; and the mode with a peak between 0.015 and
0.04 |im whose size was influenced by nucleation as well as by condensation and coagulation
was called the transient nuclei or Aitken nuclei range, subsequently shortened to the nuclei
mode. The nuclei mode could be seen in the number and surface distribution but only in special
situations was it noticeable in the mass or volume distributions. The accumulation and nuclei
modes taken together were called fine particles. An experimental size distribution showing
modes and formation mechanisms is given in Figure 2-4. This size distribution was measured in
traffic. Therefore, the nuclei mode is clearly separated from the accumulation mode and larger
than it would be in size-distributions measured farther from sources of nuclei-mode particles.
Whitby (1978) concluded
"The distinction between 'fine particles' and 'coarse particles' is a fundamental
one. There is now an overwhelming amount of evidence that not only are two
modes in the mass or volume distribution usually observed, but that these fine and
coarse modes are usually chemically quite different. The physical separation of
2-7
-------
Tt
O
15-
1_ -
o E
J2 O
Ez
3 •*"
D)
O
03
o
n
E
o
II
D)
O
10-
5-
600-
< E
-------
o
Q.
Q
O)
o
1,000,000 -
10,000 -
100-
1-
0.01-
0.0001 -
0.000001
Clean Rural
Urban Influenced Rural
Average Urban
— • — • Urban + Freeway
200,000
0.01
0.1
10
100
Particle Diameter, D (pm)
0.01 0.1 1
Particle Diameter, Dp (|jm)
10
Figure 2-2. Particle size distributions by number: (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 the linear scale, the area under any part of the curve is proportional to
particle number in that size range.
Source: Whitby (1978); Whitby and Sverdrup (1980).
the fine and coarse modes originates because condensation produces fine particles
while mechanical processes produce mostly coarse particles . . . the dynamics of
fine particle growth ordinarily operate to prevent the fine particles from growing
larger than about 1 (im. Thus, the fine and coarse modes originate separately, are
transformed separately, are removed separately, and are usually chemically
different. . . practically all of the sulfur found in atmospheric aerosol is found in
the fine particle fraction. Thus, the distinction between fine and coarse fractions is
of fundamental importance to any discussion of aerosol physics, chemistry,
measurement, or aerosol air quality standards."
Whitby's (1978) conclusions were based on extensive studies of size distributions in a
number of western and midwestern locations during the 1970s (Whitby et al., 1974; Willeke
and Whitby, 1975; Whitby, 1978; Wilson et al., 1977; Whitby and Sverdrup, 1980). No size
2-9
-------
1
o
CO
|
Q.
Q
O)
O
T3
5*
TJ
m
65-
60-
55-
50-
45-
40-
35-
30-
25-
20-
15-
10-
5-
a I
Cl I ,' • Plaan Rural
; \ Mrh I fl H
I \ Rural
; OUUu i-wt2[!Utal
: • New Mexico
/ " /^\
/ / \\
/ "\\\
/ -.--^=^^ •^' "'~-:'.-_\^
CO
.O
CO
Q.
Q
O)
.
•o
70-
65-
60-
55-
50-
45-
40-
35-
30-
25-
20-
15-
10-
5-
Average Urban
Urban + Freeway
0.01 0.1 1 10 100
Particle Diameter, Dp (pm)
0,01 0.1 1 10 100
Particle Diameter, Dp (jjm)
Figure 2-3. Size distribution by volume for the averaged (a) rural and urban-influenced
rural number distribution shown in Figure 2-2a and a distribution from
south-central New Mexico, and (b) urban and freeway-influenced urban
number distributions shown in Figure 2-2a.
Source: Whitby and Sverdrup (1980); Kim et al. (1993) and south-central New Mexico.
distribution studies of similar scope have been published since then. Newer results from particle
counting and impactor techniques, including data from Europe (U.S. Environmental Protection
Agency, 1996a) and Australia (Keywood et al., 1999, 2000), show similar results for the
accumulation and coarse modes. Extensive measurements of particle size distributions, as part
of the EPA's Supersites program, are providing considerable new data for analysis.
Whitby's (1978) conclusions have held up remarkably well. However, ideas about the
sub-0.1 |im diameter range have changed somewhat as newer instruments provided
measurements extending to smaller sizes and with greater resolution in size and time (McMurry
et al., 2000). Depending on the source, temperature, saturated vapor pressure of the components,
2-10
-------
Mechanically
Generated
0
0.001
0.01
Nuclei Mode
0.1 1
Particle Diameter, Dp (him)
10
Coarse Mode
Accumulation
Mode
Fine Particles
Coarse Particles
Ultrafine Particles
Figure 2-4. Volume size distribution, measured in traffic, showing fine and coarse
particles and the nuclei and accumulation modes of fine particles. 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 et al. (1977) and Wilson and Suh (1997).
and the age of the aerosol, size distributions have been observed with peaks (including multiple
peaks) throughout the sub-0.1 jim diameter size range. Sub-0.1 jim diameter peaks have been
observed in rural areas (O'Dowd, 2002) as well as for brief periods (nucleation bursts) in urban
areas (Woo et al., 2001a). Based on these and other observations, discussed in detail in
Section 2.1.2.3, aerosol scientists now classify particles in the sub-0.1 jim size range as ultrafme
particles and divide this size range into a nucleation region (< 10 nm) and an Aitken (nuclei)
region (10 to 100 nm), as shown in Figure 2-5. Other studies, discussed in detail in the 1996 PM
2-11
-------
_o
«
1,500 - -
1,000 - -
a.
Q
at
o
T3 500+-
Nucleation
Mode
Aitken Mode
Accumulation
Mode
10 20 50 100 200
Particle Diameter Dp (nm)
500 1,000
Figure 2-5. Sub-micrometer number size distribution observed in a boreal forest in
Finland showing the trimodal structure of fine particles. The total particle
number concentration was 1,011 particles/cm3 (10-min average).
Source: Makelaetal. (1997).
AQCD (U.S. Environmental Protection Agency, 1996a), have shown that in fog or clouds or at
very high relative humidities the accumulation mode may split into a larger size (more
hygroscopic or droplet) submode and a smaller size (less hygroscopic or condensation) submode.
Definitions of Particle Size Fractions
In the preceding discussion several subdivisions of the aerosol size distribution were
identified. Aerosol scientists use several different approaches or conventions in the
classification of particles by size. These include: (1) modes, based on the observed size
distributions and formation mechanisms; (2) dosimetry or occupational health sizes, based on the
entrance into various compartments of the respiratory system; and (3) cut point, usually based on
the 50% cut point of the specific sampling device, including legally specified, regulatory cut
points for air quality standards.
2-12
-------
Modal. The modal classification as first proposed by Whitby (1978) is shown in
Figures 2-1 and 2-4. The newer modes introduced since 1978 are shown in Figure 2-5.
An idealized distribution showing all four modes is shown in Figure 2-6. The nucleation and
Aitken modes are best observed in the number distribution. However, the Aitken mode can be
seen in the volume distribution in traffic or near traffic or other sources of ultrafine particles
(Figures 2-3b and 2-4). The observed modal structure is frequently approximated by several
lognormal distributions. Definitions of terms used to describe size distributions in modal terms
are given below.
0.001
0.1 1
Particle Diameter, Dp (pm)
Nucleation Mode Accumulation Mode
Aitken Mode
r
10
Coarse Mode
100
Fine Particles
Ultrafine Particles
Coarse Particles
Figure 2-6. An idealized size distribution, that might be observed in traffic, showing fine
and coarse particles and the nucleation, Aitken, and accumulation modes that
comprise fine particles. Also shown are the major formation and growth
mechanisms of the four modes of atmospheric particles.
2-13
-------
NucleationMode: Freshly formed particles with diameters below about 10 nm, observed
during active nucleation events. The lower limit, where particles and molecular clusters or
large molecules overlap, is uncertain. Current techniques limit measurements to particles
3 nm or greater.
AitkenMode: Larger particles with diameters between about 10 and 100 nm. The Aitken
mode may result from growth of smaller particles or nucleation from higher concentrations
of precursors.
Accumulation Mode: Particles with diameters from about 0.1 |im to just above the
minimum in the mass or volume distributions which usually occurs between 1 and 3 jim.
Fine Particles: Fine particles include the nucleation, Aitken, and accumulation modes,
i.e., particles from the lowest measurable size, currently about 3 nm, to just above the
minimum in the mass or volume distribution which generally occurs between 1 and 3 |im.
Coarse Mode or Coarse Particles: Particles with diameters mostly greater than the
minimum in the particle mass or volume distributions, which generally occurs between
1 and 3 |im.
Ultrafine Particles: Ultrafine particles are not a mode. In the air pollution literature, they
are generally defined by size alone, i.e., particles with diameters of 0.1 jim (100 nm) or
less. They include the nucleation mode and much of the Aitken mode. They may also be
defined as particles whose properties differ from those of the bulk material because of their
small size.
Modes are defined primarily in terms of their formation mechanisms but also differ in
sources, composition, transport and fate, as well as in size. Nucleation mode applies to newly
formed particles which have had little chance to grow by condensation or coagulation. Aitken
mode particles are also recently formed particles that are still actively undergoing coagulation.
However, because of higher concentrations of precursors or more time for condensation and
2-14
-------
coagulation, Aitken particles have grown to larger sizes. Fine particles grow by coagulation
(two particles combining to form one) or by condensation (low-equilibrium vapor pressure gas
molecules condensing on a particle). As the particle size increases, the rate of growth by
coagulation and condensation decreases and particles "accumulate" in the accumulation mode
size range. Thus, accumulation-mode particles normally do not grow into the coarse particle
size range. However, during conditions of high relative humidity, hygroscopic accumulation
mode particles grow in size, increasing the overlap of fine and coarse particles. The
accumulation mode may split into a (hygroscopic) droplet mode and a (non-hygroscopic)
condensation mode. In addition, gas-phase pollutants may dissolve and react in the particle-
bound water of hygroscopic particles, leading to an increase in the dry size. The combination of
nucleation, Aitken, and accumulation modes are called fine particles (or sometimes fine-mode
particles). Fine particles are formed primarily by combustion or chemical reactions of gases
yielding products with low saturated vapor pressures. Fine particles are composed of metals
(and metal oxides), black or elemental carbon, primary and secondary organic compounds, and
sulfate, nitrate, ammonium and hydrogen ions.
The coarse mode refers to particles formed by the mechanical breakdown of minerals,
crustal material, and organic debris. In addition to primary minerals and organic material, the
coarse mode may include sea salt, nitrate formed from the reaction of nitric acid with sodium
chloride, and sulfate formed from the reaction of sulfur dioxide with basic particles. The
accumulation mode and the coarse mode overlap in the region between 1 and 3 jim (and
occasionally over an even larger range). In this region, the chemical composition of individual
particles can usually, but not always, allow identification of a source or formation mechanism,
permitting identification of a particle as belonging to the accumulation or coarse mode.
Occupational Health or Dosimetric Size Cuts. The occupational health community has
defined size fractions in terms of their entrance into various compartments of the respiratory
system. This convention classifies particles into inhalable, thoracic, and respirable particles
according to their upper size cuts. Inhalable particles enter the respiratory tract, beginning with
the head airways. Thoracic particles travel past the larynx and reach the lung airways and the
gas-exchange regions of the lung. Respirable particles are a subset of thoracic particles that are
more likely to reach the gas-exchange region of the lung. In the past, exact definitions of these
2-15
-------
terms have varied among organizations. As of 1993, a unified set of definitions was adopted by
the American Conference of Governmental Industrial Hygienists (ACGIH, 1994), the
International Standards Organization (ISO), and the European Standardization Committee
(CEN). The curves which define inhalable (IPM), thoracic (TPM), and respirable (RPM)
particulate matter are shown in Figure 2-7. These curves should not be taken to indicate that
particles > 4 jim Da do not reach the gas exchange regions or that particles < 4 jim Da do not
deposit in the bronchi. See Figure 6-13 in Chapter 6 for a graphical characterization of particle
deposition in regions of the respiratory system as a function of particle size.
100
D)
O
.C
C
_o
I
"5
0)
0.
_0)
o
7i 20
re
O.
2.5 4 10 20 50
Aerodynamic Diameter, Da (|jm)
100
Figure 2-7. 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 [200 Ic], PM10
[2001a]). PM25 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).
2-16
-------
Size-Selective Sampling. Another set of definitions of particle size fractions arises from
considerations of size-selective sampling. Size-selective sampling refers to the collection of
particles below or within a specified aerodynamic size range. Size fractions are usually specified
by the 50% cut point size; e.g., PM2 5 refers to particles collected by a sampling device that
collects 50% of 2.5 jim particles and rejects 50% of 2.5 jim particles. However, size fractions
are defined not merely by the 50% cut point, but by the entire penetration curve. Examples of
penetration curves are given in Figure 2-7. Thus, as shown by Figure 2-7, a PM2 5 sampler, as
defined by the Federal Reference Method, rejects 94% of 3 jim particles, 50% of 2.5 jim
particles, and 16% of 2 jim particles. Samplers with the same 50% cut point, but differently
shaped penetration curves, would collect different fractions of PM. Size-selective sampling has
arisen in an effort to measure particle size fractions with some special significance (e.g., health,
visibility, source apportionment, etc.), to measure mass size distributions, or to collect size-
segregated particles for chemical analysis. Dichotomous samplers split the particles into smaller
and larger fractions that may be collected on separate filters. However, some fine particles
(~ 10%) are collected with the coarse particle fraction. Cascade impactors use multiple size cuts
to obtain a distribution of size cuts for mass or chemical composition measurements. One-filter
samplers with a variety of upper size cuts are also used, e.g., PM25, PM10.
An idealized particle size distribution with the normally observed division of ambient
aerosols into fine and coarse particles and the size fractions collected by the WRAC, TSP, PM10,
PM2 5 and PM10_2 5 samplers is shown in Figure 2-8. PM10 samplers, as defined in Appendix J to
Title 40 Code of Federal Regulations (40 CFR), Part 50 (Code of Federal Regulations, 200 la;
Federal Register, 1987), collect all of the fine-mode particles and part of the coarse-mode
particles. The upper cut point is defined as having a 50% collection efficiency at 10 ± 0.5 jim
aerodynamic diameter. The slope of the collection efficiency curve is defined in amendments to
40 CFR, Part 53 (Code of Federal Regulations, 200 Ib).
An example of a PM25 size-cut curve is also shown in Figure 2-7. The PM25 size-cut
curve, however, is defined by the design of the Federal Reference Method (FRM) sampler. The
basic design of the FRM sampler is given in the Federal Register (1997, 1998) and in 40 CFR,
Part 50, Appendix L (Code of Federal Regulations, 200 Ic). Additional performance
specifications are given in 40 CFR, Parts 53 and 58 (Code of Federal Regulations, 2001b,d).
In order to be used for measurement of PM2 5 to determine compliance with the PM2 5 NAAQS,
2-17
-------
Accumulation Mode
-------
countries use PM10 to refer not to samplers with a 50% cut at 10 jim Da, but to samplers with
100% rejection of all particles greater than 10 jim Da. Such samplers miss a fraction of coarse
thoracic PM. An example is shown in Figure 2-9.
100
90-
80-
T 70~
^ 60-
.0
li 50 -
•*-•
oj
40-
0)
30-
20-
10-
Kimoto cyclonic Inlet
^^^ Manufacturer
• Tsai and Cheng (1996)
0
1
Wedding Cyclonic Inlet
O U = 2 km/h
D U = 8 km/h
A U = 24 km/h
Wedding and
Weigand (1993)
2 4 6 8 10
Aerodynamic Diameter, Da (pm)
20
-a—i
30
Figure 2-9. Comparison of penetration curves for two PM10 beta gauge samplers using
cyclone inlets. The Wedding PM10 sampler uses the EPA definition of PMX as
x = 50% cut point. The Kimoto PM10 defines PMX as x = the 100% cut point
(or zero penetration).
Source: Tsai and Cheng (1996).
PM10, as defined by EPA, refers to particles collected by a sampler with an upper 50% cut
point of 10 |im Da and a specific, fairly sharp, penetration curve. PM25 is analogously defined.
Although there is not yet an FRM, PM10_2 5 refers either to particles collected by a sampler with
an upper 50% cut point of 10 |im Da and a lower 50% cut point of 2.5 |im Da or to the difference
between the particle concentration measured by a PM10 monitor and a PM2 5 monitor. In all
2-19
-------
cases, the fraction of PM collected depends on the entire penetration curve (or curves); i.e., for
PM25 some particles > 2.5 jim Da are collected and not all particles < 2.5 jim Da are collected.
A PM10_2 5 size fraction may be obtained from a dichotomous sampler or by subtracting
the mass collected by a PM2 5 sampler from the mass collected by a PM10 sampler. The
resulting PM10_2 5 mass, or PM10_2 5, is sometimes called "coarse" particles or "thoracic coarse"
particles. However, it would be more correct to call PM10_2 5 an indicator of the thoracic
component of coarse particles (because it excludes some coarse particles below 2.5 jim Da and
above 10 jim Da). Also, PM25 should be considered an indicator of fine particles (because it
contains some coarse particles). It would also be appropriate to call PM10 an indicator of
thoracic particles. PM10 and thoracic PM, as shown in Figure 2-7, have the same 50% cut point.
However, the thoracic cut is not as sharp as the PM10 cut; therefore, thoracic PM contains some
particles between 10 and 30 jam diameter that are excluded from the PM10 fraction.
Over the years, the terms fine and coarse, as applied to particles, have lost the precise
meaning given in Whitby's (1978) definition. In any given article, therefore, the meaning of fine
and coarse, unless defined, must be inferred from the author's usage. In this document, "fine
particles" means all particles in the nucleation, Aitken, and accumulation modes; and "coarse
particles" means all particles in the coarse mode. Fine particles and PM2 5 are not equivalent
terms.
Selection of Cut Points for Regulatory Size Cuts
TSP. Regulatory size cuts are a specific example of size-selective sampling. Prior to
1987, the indicator for the NAAQS for PM was TSP. TSP is defined by the design of the High
Volume sampler (hivol) that collects all of the fine particles but only part of the coarse particles
(Figure 2-8). The upper cut-off size of the hivol depends on the wind speed and direction and
may vary from 25 to 40 |im. Newer PM samplers are usually designed to have an upper cut
point and its standard deviation that are independent of wind direction and relatively independent
of wind speed.
PM10. In 1987, the NAAQS for PM were revised to use PM10, rather than TSP, as the
indicator for the NAAQS for PM (Federal Register, 1987). The use of PM10 as an indicator is an
example of size-selective sampling based on a regulatory size cut (Federal Register, 1987). The
2-20
-------
selection of cut point characteristics depends upon the application for the sampling device.
A separation that simulates the removal of particles by the human upper respiratory system
would appear to be a good choice for both health risk and regulatory monitoring (i.e., it would
measure what gets into the lungs). The ACGIH-ISO-CEN penetration curve for thoracic
particles (particles able to pass the larynx and penetrate into the bronchial and alveolar regions of
the lung) has a 50% cut point at 10 jim aerodynamic diameter (Da). The selection of PM10 as an
indicator was based on health considerations and was intended to focus regulatory concern on
those particles small enough to enter the thoracic region of the human respiratory tract. The
PM10 is an indicator of thoracic particles and is a compromise between the desire to collect all
particles that might enter the thoracic regions and the need to design a sampler with a collection
efficiency independent of particle size or wind speed. As shown in Figure 2-7, the PM10
penetration efficiency curve matches the definition for thoracic PM given by the American
Conference of Government and Industrial Hygienists (1994) very well except for a slight under-
collection of particles between 10 and 30 jim in diameter. While the U.S. PM10 separation curve
is sharper than the thoracic penetration curve, it has the advantage of reducing the problem of
maintaining the finite collection efficiency specified by the thoracic penetration curve for
particles larger than 10 |im Da. (See Section 2.1.2.2 and Figure 2-7.)
Current PM10 samplers have upper cut points that are stable under normal operating
conditions. However, problems may occur under unusual or adverse conditions. Ono et al.
(2000) reported the results of a study in which several PM10 samplers were collocated and
operated at various sites at Owens Lake, C A, a location with high concentrations of coarse PM.
Samplers included the Partisol, the dichotomous, the Wedding high-volume, and the Graseby
high-volume samplers in addition to the tapered element oscillating microbalance (TEOM)
monitor. They found that the TEOM and Partisol samplers agreed to within 6%, on average.
The dichotomous sampler and the Graseby and Wedding high-volume samplers, however,
measured significantly lower PM10 concentrations than the TEOM (on average 10, 25, and 35%
lower, respectively). These lower concentrations were attributed to a decrease in the cut point at
higher wind speeds and to a dirty inlet. Since the 10 jim cut point is on a part of the size
distribution curve where the concentration is changing rapidly, the amount of PM collected is
sensitive to small changes in the cut point. Therefore, the cut point needs to be specified very
2-21
-------
precisely and the balance between a design standard and a performance standard will need to be
considered to deal with this problem.
PM25: Cut Point for Separation of Fine and Coarse Particles. The PM2 5 standard set in
1997 is also an example of size-selective sampling based on a regulatory size cut (Federal
Register, 1997). The PM25 standard was based primarily on epidemiologic studies using
concentrations measured with PM2 5 samplers as an exposure index. However, the PM2 5 sampler
was not designed to collect all respirable particles; it was designed to separate fine and coarse
particles and to collect fine particles because of their different sources and properties (Whitby,
1978; Miller et al., 1979). Thus, the need to attain a PM25 standard tends to focus regulatory
concern on control of the sources of fine particles.
Fine and coarse particles differ not only in size but also in formation mechanisms, sources,
and chemical, physical, and biological properties. They also differ in concentration-exposure
relationships, dosimetry (deposition and retention in the respiratory system), toxicity, and health
effects as observed in epidemiologic studies. Thus, it is desirable to measure fine and coarse PM
separately in order to properly allocate health effects to either fine or coarse PM and to correctly
determine sources by receptor modeling approaches. For example, sulfates in fine particles are
associated with hydrogen or ammonium ions, while sulfates in coarse particles are associated
with basic metal ions. Transition metals in coarse particles are likely to be associated with soil
and tend to be less soluble (and presumably less bioavailable) than transition metals in fresh
combustion particles found in fine particles.
In the early 1970s, aerosol scientists were beginning to recognize the existence of a
minimum between 1 and 3 jim in the distribution of particle size by volume (Whitby et al.,
1974). However, the limited size distribution information available at that time (two
distributions with a minimum near 1 jim and two with a minimum near 2.5 jim) did not permit a
unambiguous definition of the appropriate cut point size for the separation of the two modes.
A cut point of 2.5 jim was chosen for a new dichotomous sampler (Loo et al., 1976; Jaklevic
et al., 1977) designed for use in the Regional Air Pollution Study in St. Louis, MO. The 2.5 jim
cut point was subsequently used as an indicator of fine particles in a number of studies including
the Harvard Six-City Studies of the relationships between mortality and PM concentrations
(Dockery et al., 1993; Schwartz et al., 1996). In an analysis reported in 1979, EPA scientists
2-22
-------
endorsed the need to measure fine and coarse particles separately (Miller et al., 1979). Based on
the availability of a dichotomous sampler with a separation size of 2.5 jim Da, they
recommended 2.5 jim Da as the cut point between fine and coarse particles. Because of the wide
use of this cut point, the PM2 5 fraction is frequently referred to as "fine" particles. However,
although the PM25 sample will contain most of the fine particles, except to a lesser extent during
conditions of high relative humidity, it may also collect a small fraction of the coarse particles,
especially in dry areas or during dry conditions. A 2.5 jim cut point was also used in the
Inhalable Particle Network (Suggs and Burton, 1983), which provided data for another major
epidemiologic study of PM-mortality relationships using an American Cancer Society cohort
(Pope et al., 1995). Therefore, at the time of the last previous review of the NAAQS for PM
(U.S. Environmental Protection Agency, 1996a,b), a number of epidemiologic studies
demonstrating a statistical relationship between PM2 5 concentrations and mortality were
available.
During the previous review of the PM standards, EPA conducted an extensive review of
the cut point to be used for a fine particle standard, including consideration of PMl as an
alternative to PM25. The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a)
contains a review of the available information on size distributions. As shown in Figure 2-10
(adapted from the 1966 PM AQCD; U.S. Environmental Protection Agency, 1996a), published
size distributions exhibit considerable variability in the intermodal region (1 to 3.0 jim
diameter). In Figure 2-10a, Philadelphia, very little mass is found in the intermodal region and it
is not clear whether it should be considered fine or coarse. In Figure 2-1 Ob, Phoenix, a tail of the
coarse mode extends to 1 |im or below. These two size distributions were fit with three
lognormal distributions. However, the intermodal mode probably does not have physical
significance. In Figure 2.10c, Claremont, South Coast Basin, a size distribution is shown with
the accumulation mode split into a condensation mode and a droplet mode, which extends to
3.0 |im or above (John et al., 1990). The droplet mode occurs under high relative humidity (RH)
conditions, usually with very high fine PM concentrations, and is believed to result from
reactions involving gases dissolving and reacting in the particle-bound water of the particle
droplets (Hering and Friedlander, 1982; McMurry and Wilson, 1983). Although Figure 2-10c
shows nitrate concentrations, similar distributions have been observed for sulfate (John et al.,
1990).
2-23
-------
Philadelphia - WRAC
1.0 10.0
Aerodynamic Diameter, Da (|jm)
100.0
Phoenix-WRAC
90.0
re
Q
Mode MM AD
Ac 0.188
IM 1.70
C 16.4
ag %Mass
1.54 22.4
1.90 13.8
2.79 63.9
Mass
1.0 10.0
Aerodynamic Diameter, Da
100.0
South Coast Air Basin - Berner Impactor
0)
Q
O)
O
-a
300
250 .
200
150 -|
100
50
^
/?
/
/
V
\
\
\
>/
Mode MMAD ag %Mass
Accumulation
Condensation 0.20 1.44 9.0
Droplet 0.73 1.68 41.0
Coarse 4.40 1.82 50.0
_^*-~ ^
/ '
\
V
Nitrate
0.1
1.0 10.0
Aerodynamic Diameter, Da
100.0
Figure 2-10. Three examples of impactor size distributions and distributions resulting
from fitting several lognormal distributions to the impactor data. These
distributions show the variation in the particle size range of the minimum in
mass between the accumulation mode and the coarse mode.
Source: a and b, adapted from Lundgren and Hausknecht (1982); c, John et al. (1990).
2-24
-------
It is now understood that the size range between 1.0 and 3.0 jam, sometimes called the
intermodal region, may contain either accumulation-mode or coarse-mode material or both, i.e.,
the two modes may overlap in this region (Kegler et al., 2001). The experimental information on
the composition and source of the intermodal mass was discussed extensively in the 1996 PM
AQCD (U.S. Environmental Protection Agency, 1996a). Depending largely on the RH, a
significant amount of either accumulation- or coarse-mode material may be found in the
intermodal region between 1.0 and 3.0 jim. The analysis demonstrated the important role of
relative humidity in influencing the size of particles in both the accumulation and coarse modes.
As the RH increases, hygroscopic accumulation-mode particles will increase in size due to
accumulation of particle-bound water. At high RH, some originally submicrometer-sized
accumulation-mode PM may be found with a Da above 1 jim. At an RH of 100%, such as is
found in fog and clouds, accumulation-mode PM may extend above 2.5 jim Da. What is not well
understood is whether such particles will shrink to diameters below 1 jim as the RH decreases or
whether reactions occurring in the wet particles will result in an increase in nonaqueous mass, so
that, even at low RH, the diameters would exceed 1 jim. On the other hand, at very low RH,
coarse particles may be fragmented into smaller sizes, and small amounts of coarse PM may be
found with an Da < 2.5 jim (Lundgren et al., 1984; Lundgren and Burton, 1995). Thus, a PM2 5
sample will contain all of the accumulation-mode PM except during periods of high RH.
However, under low-RH conditions, it may also contain a small fraction of the coarse PM.
Considerations that led the EPA to choose to retain 2.5 jim as the cut point for the
separation of fine and coarse particles included the following three points. First, epidemiologic
data showing statistical relationships between fine PM and health outcomes were based largely
on PM25. Second, while PMl would exclude a tail of the coarse mode in some locations, in other
locations it would miss a portion of the fine PM. Since the growth of the droplet mode is
associated with very high fine PM concentrations, this would result in falsely low fine PM
measurements on days with the highest fine PM concentrations. Third, only limited data on the
concentration and composition of the intermodal PM mass was available. The selection of a cut
point of 2.5 |im as a basis for the EPA's 1997 NAAQS for fine particles (Federal Register, 1997)
and its continued use in many health effects studies reflect the importance placed on a more
complete inclusion of accumulation-mode particles, while recognizing that the intrusion of
coarse particles can occur under some conditions with this cut point.
2-25
-------
Since the 1966 PM AQCD (U.S. Environmental Protection Agency, 1996a), several papers
have addressed the issue of intermodal mass in terms of PM2>5-PM1. Kegler et al. (2001)
analyzed data from Phoenix including TEOM measurements of PMl3 PM25, and PM10 as well as
filter measurements of PM25 and concluded that while PM2 ^ was dominated by soil
components, it also contained some nonsoil components. Their analysis suggested that there
were two sources of the coarse mass found in PM2 5^: (1) resuspension of soil dust by natural
wind (windblown dust), which would be prominent at high wind speeds, and (2) resuspension
due to roadway turbulence generated by motor vehicles (road dust), which would occur at all
natural wind speeds. Correlations among the various PM size ranges measured in Phoenix are
given in Table 2-1. The high correlation found between PM25 and PMX (r = 0.97) suggests that
the use of PMX instead of PM25 would not significantly change epidemiologic relationships
between PM mass and health outcomes in Phoenix.
TABLE 2-1. CORRELATIONS BETWEEN TEOM MEASUREMENTS IN PHOENIX
PMj
PM2.5.!
PM25
PM10.2.5
PM10
PM,
I
0.69
0.97
0.65
0.81
PM2,_,
0.69
1
0.84
0.84
0.89
PM25
0.97
0.84
1
0.75
0.89
PMW_2,
0.65
0.84
0.75
1
0.97
PM10
0.81
0.89
0.89
0.97
1
In areas where winds cause high concentrations of windblown soil, there is evidence that a
significant amount of coarse-mode PM may be found below 2.5 |im. An example, taken from
data collected during the August 1996 dust storm in Spokane, WA, is shown in Figure 2-11.
Note that the PM10 scale is 10 times that of the other size fractions. PMl3 although high in the
morning, goes down as the wind increases and PM10, PM2 5, and PM2 ^ go up. During the peak
of the dust storm, around 9:00 p.m., PM2 ^ was 88% of PM2 5. For the 24-h period, PM2 ^ was
54% of PM25. However, PMX was not affected by the intrusion of coarse particles. Similar
2-26
-------
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).
considerations probably apply to short-term intrusions of dust transported from distant sources
such as the Sahara and Gobi deserts (Husar et al., 2001). In Spokane, the correlations
between PMj and PM^s.! (Haller et al., 1999) were lower, r = -0.37 (spring) to 0.26 (winter),
than those observed in Phoenix (Kegler et al., 2001), r = 0.69.
Pope et al. (1999) found that removing days with high PM10 values due to windblown dust
resulted in the observation of a statistically significant relationship between PM10 and mortality
in Salt Lake City. This relationship was not observed when days with windblown dust events
were included in the PM10 concentration time series. The elevated values of PM25, but not PMl3
observed during the windblown dust events in Spokane (Claiborn et al., 2000) suggest that
a PM2 5 time series would also be impacted by windblown dust while a PMj time series would
not. Thus, the contribution of soil to PM2 5 could possibly contribute to erroneous conclusions
from some epidemiologic analyses.
2-27
-------
A cut point of 1.0 jim could reduce the misclassification of coarse-mode material as fine,
especially in areas with high levels of windblown soil, but under high RH conditions could result
in some accumulation-mode material being misclassified as coarse. A reduction in RH, either
intentionally or inadvertently, will reduce the mass mean diameter of the accumulation- mode
particles. Studies of the changes in particle size with changes in relative humidity suggest that
only a small fraction of accumulation-mode particles will be above 1.0 jim in diameter at RH
below 60%, but a substantial fraction will grow to above 1.0 jim at RH above 80% (Hitzenberger
et al., 1997; McMurry and Stolzenburg, 1989; U.S. Environmental Protection Agency, 1996a).
Studies in Europe (Berner, 1989; see Figure 3-31 in the 1966 PM AQCD) show that
dehumidification of ambient PM by heating will reduce the size of accumulation mode particles,
leaving little accumulation mode PM above 1 |im in aerodynamic diameter. However, heating
will also remove ammonium nitrate and semivolatile organic compounds. Currently, the data are
insufficient to determine how much dehumidification by diffusion drying would reduce the size
of accumulation mode particles.
Under high RH circumstances, a monitor using a 1.0 jim Da cut point can achieve better
modal separation if the air stream is dehumidified to some fixed humidity that would remove all
or most particle-bound water without evaporating semivolatile components. New techniques
requiring the reduction of RH by diffusion drying prior to collection have been developed for
measuring fine PM minus particle-bound water but including semivolatile nitrate and organic
compounds. With such techniques, measurements with a 1 |im (or slightly higher) cut point, in
conjunction with concurrent PM2 5 measurements, would be useful for exposure, epidemiologic,
and source-apportionment studies, especially in areas where intrusion of coarse-mode particles
into the intermodal range is likely.
2.1.2.3 Ultrafine Particles
As discussed in Chapter 7 (Toxicology of Particulate Matter in Humans and Laboratory
Animals) and in Chapter 8 (Epidemiology of Human Health Effects Associated with Ambient
Particulate Matter), some scientists argue that ultrafine particles may pose potential health
problems and that some health effects may be associated with particle number or particle surface
area as well as, or more closely than, with particle mass. Some additional attention will be given
2-28
-------
here to ultrafine particles, because they contribute the major portion of particle number and a
significant portion of particle surface area.
Formation and Growth of Fine Particles
Several processes influence the formation and growth of particles. New particles may be
formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
material condenses on existing particles; particles may also grow by coagulation as two particles
combine to form one. Gas phase material condenses preferentially on smaller particles, and the
rate constant for coagulation of two particles decreases as the particle size increases. Therefore,
ultrafine particles grow into the accumulation mode; but accumulation-mode particles do not
normally grow into the coarse mode (see Figure 2-6). More information and references on the
formation and growth of fine particles can be found in the 1996 AQCD PM (U.S. Environmental
Protection Agency, 1996a).
Equilibrium Vapor Pressures
An important parameter in particle nucleation and in particle growth by condensation is the
saturation ratio, S, defined as the ratio of the partial pressure of a species, p, to its equilibrium
vapor pressure above a flat surface at a specified temperature, p0: S = p/p0. For either
condensation or nucleation to occur, the species vapor pressure must exceed its equilibrium
vapor pressure. For particles, the equilibrium vapor pressure is not the same as p0. Two effects
are important: (1) the Kelvin effect, which is an increase in the equilibrium vapor pressure
above the surface due to its curvature (very small particles have higher vapor pressures and will
not be stable to evaporation until they attain a critical size) and (2) the solute effect, which is a
decrease in the equilibrium vapor pressure of the liquid due to the presence of other compounds
in solution. Organic compounds may also be adsorbed onto ultrafine carbonaceous particles.
For an aqueous solution of a nonvolatile salt, the presence of the salt decreases the
equilibrium vapor pressure of the water vapor around the droplet. This effect is in the opposite
direction of the Kelvin effect, which increases the equilibrium water vapor pressure around a
droplet because of its curvature. The existence of an aqueous solution will also influence the
vapor pressure of water-soluble species. The vapor pressure behavior of mixtures of several
liquids or of liquids containing several solutes is complex.
2-29
-------
New Particle Formation
When the vapor concentration of a species exceeds its equilibrium concentration
(expressed as its equilibrium vapor pressure), it is considered condensable. Condensable species
can either condense on the surface of existing particles or can nucleate to form new particles.
The relative importance of nucleation versus condensation depends on the rate of formation of
the condensable species and on the surface or cross-sectional area of existing particles (McMurry
and Friedlander, 1979). In ambient urban environments, the available particle surface area is
usually sufficient to rapidly scavenge the newly formed condensable species. Formation of new,
ultrafine particles is usually not observable in mass or volume distributions except near sources
of condensable species. Wilson et al. (1977) reported observations of the Aitken nuclei mode in
traffic. However, bursts of new particle formation have been observed in urban areas in the
number distribution (Woo et al., 2001a; McMurray et al., 2000). New particle formation also
can be observed in cleaner, remote regions. Bursts of new particle formation in the atmosphere
under clean conditions usually occur when aerosol surface area concentrations are low (Covert
et al., 1992). High concentrations of nuclei mode particles have been observed in regions with
low particle mass concentrations, indicating that new particle formation is inversely related to
the available aerosol surface area (Clarke, 1992).
Sources of Ultrafine Particles
Ultrafine particles are the result of nucleation of gas phase species to form condensed
phase species with a very low equilibrium vapor pressure. In the atmosphere, four major classes
of substances yield PM with equilibrium vapor pressures low enough to form nuclei mode
particles:
(1) Particles containing heavy metals. Nuclei mode particles of metal oxides or other metal
compounds are generated when metallic impurities in coal or oil are vaporized during
combustion and the vapor undergoes nucleation. Metallic ultrafine particles also may be
formed from metals in lubricating oil or fuel additives that are vaporized during
combustion of gasoline or diesel fuels. Ultrafine metallic particles were discussed in
Section 6.9 of the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a).
(2) Elemental carbon (EC) or soot. Elemental carbon particles are formed primarily by the
condensation of C2 molecules generated during the combustion process. Because EC
has a very low equilibrium vapor pressure, ultrafine EC particles can nucleate even at
high temperatures (Kittelson, 1998; Morawska et al., 1998).
2-30
-------
(3) Organic carbon (OC). Recent smog chamber studies and indoor experiments show that
atmospheric oxidation of certain organic compounds often found in the atmosphere can
produce highly oxidized organic compounds with an equilibrium vapor pressure
sufficiently low to result in nucleation (Kamens et al., 1999; Weschler and Shields,
1999).
(4) Sulfates. Sulfuric acid (H2SO4) molecules are generated in the atmosphere by
conversion of sulfur dioxide (SO2) to H2SO4. As H2SO4 is formed, it can either nucleate
to form new ultrafme particles, or it can condense on existing ultrafine or accumulation
mode particles (Clark and Whitby, 1975; Whitby, 1978). Nucleation theory allows the
calculation of nucleation rates for both binary nucleation, involving water and sulfuric
acid (Easter and Peters, 1994) or ternary nucleation, which requires sulfuric acid,
ammonia (NH3), and water (Korhonen et al., 1999). Kulmala et al. (2000) compared
binary and ternary nucleation rates and concluded that the ternary rate is 1,000 times the
binary rate. Results from an aerosol dynamics model with a ternary nucleation scheme
indicate that nucleation in the troposphere should be ubiquitous and yield a reservoir of
thermodynamically stable sulfate clusters 1 to 3 nm in diameter. The growth of these
clusters to a detectable size (> 3 nm diameter) is limited by the availability of
condensable vapor. Observations of atmospheric particle formation and growth from a
continental and a coastal site suggest that a growth process including ternary nucleation
is responsible for the formation of cloud condensation nuclei. Nucleation processes in
the atmosphere may also involve organic compounds as well as sulfuric acid, ammonia,
and water. However, current formulations of nucleation theory only include the three
inorganic components. (The possible formation of ultrafine NH4NO3 by reaction of NH3
and nitric acid [HNO3] vapor has not been investigated.)
Vehicle engine exhaust may include all the substances described above. Ultrafine particles
are observed in the emissions from spark, diesel, and jet engines (Kittelson, 1998). In these
cases, it seems likely that EC, organic compounds, ammonia and sulfuric acid from sulfur in the
fuel, as well as metal additives in the fuel or fuel oil, may contribute to the formation of ultrafine
particles (Tobias et al., 2001). An example of particles from a diesel engine showing number
and volume distributions is shown in Figure 2-12.
Recent Measurements of Ultrafine Particles
Instruments developed during the past decade permit measurement of size distributions of
particles down to 3 nm in diameter. The use of these techniques has led to new information on
the formation of new particles by nucleation. Such measurements have been carried out during
intensive field measurement campaigns, during continuous measurements in urban areas in
several European cities, and in the United States as a part of the Supersite program (McMurry
2-31
-------
Q
Ol
O
O
•O
O
'^
ra
O
O
•c
0)
N
ra
Number Weighting
Mass Weighting
Ultrafine Particles
Dp < 100 nm
H
Accumulation
Mode
Coarse
Mode
0.001
0.01 0.1
Particle Diameter, D (|Jm)
10
Figure 2-12. Typical engine exhaust size distribution.
Source: Kittelson (1998).
et al., 2000; Woo et al., 2001a). Nucleation has been observed in the free troposphere (Weber
et al., 1999; Clarke, 1992; Schroder and Strom, 1997; Raes et al., 1997), in outflows of
evaporating convective clouds (Clarke et al., 1998; Hegg et al., 1990, 1991; Radke and Hobbs,
1991; Perry and Hobbs, 1994), in the marine boundary layer (Covert et al., 1992; Hoppel et al.,
1994; Van Dingenen et al., 1995; Weber et al., 1998; Clarke et al., 1998), downwind of coastal
regions during low tide (McGovern et al., 1996; McGovern, 1999), on mountains (Weber et al.,
1995, 1997; Raes et al., 1997; Wiedensohler et al., 1997), over forests (Makela et al., 1997;
Kulmala et al., 1998; O'Dowd et al., 2002), downwind of certain biogenic emissions (Weber
et al., 1998), in urban areas (Birmili and Wiedensohler, 1998; McMurry et al., 2000; Woo et al.,
2001a), near freeways (Zhu et al., 2002a,b), in engine exhaust (Kittelson, 1998; Tobias et al.,
2001), and in homes (Wallace and Howard-Reed, 2002). Nucleation events in outdoor air
almost always occur during daylight, indicating that photochemistry plays a role in producing
the gas phase precursors of new particles.
The number size distributions observed over a boreal forest in Finland before and during
the initial stages of a nucleation event are shown in Figure 2-13. The Aitken and accumulation
modes can be seen clearly before the nucleation event. The nucleation mode, with a peak
2-32
-------
<*J
I
Q
0)
O
T3
102
Before Nucleation
During Nucleation
10 100
Particle Diameter, Dp (nm)
1,000
Figure 2-13. Number size distributions showing measurement of a nucleation burst mode
in a boreal forest in Finland.
Source: O'Dowd et al. (2002).
between 3 and 7 nm, appears during the event. Figure 2-14 shows the variety of size
distributions that may be observed as nuclei are formed and grow, based on size distributions
measured in the Arctic marine boundary layer (Covert et al., 1996). These distributions all show
a trimodal distribution within the fine-particle size range. The changes in size distribution due to
coagulation (and dilution) immediately downwind of a freeway (Zhu et al., 2002b) are shown in
Figures 2-15(a-f) and 2-15(g). At 30 m downwind, the nucleation mode number concentration is
larger than that of the Aitken mode, but by 60 m downwind that has reversed, because
coagulation removes particles from the nucleation mode and adds particles to the Aitken mode.
Strong evidence suggests that sulfuric acid vapor often participates in nucleation.
However, condensation of sulfuric acid and its associated water and ammonium ions typically
accounts for only 10 to 20% of the observed growth rates for freshly nucleated particles.
Therefore, organic compounds may account for much of the formation and growth of freshly
nucleated particles. Evidence of nucleation of organic particles comes from smog chamber
studies (Kamens et al., 1999) and from field studies over forests (Makela et al., 1997; Kulmala
et al., 1998; O'Dowd et al., 2002). Nucleation of organic particles may also occur indoors due to
the reaction of infiltrated ozone with indoor terpenes from air fresheners or cleaning solutions
(Weschler and Shields, 1999). The observation of bursts of nuclei-mode particles in Atlanta
2-33
-------
CO
E
•^
£
Q
o
_0
T!
1400 -
1200 -
1000 -
800 -
600 -
400 -
200 -
n -
_§J i August 30, 1991, 1000 UTC
A
•A
eo
U
Q
O)
5
z
•O
600
600
500 -
400 -
300 -
200 -
100 -
0
August 31, 1991, 0200 UTC
10
100
1000
1
10
100
1000
700
600 -
500 -
400 -
300 -
200 -
100 -
0
September 22, 1991, 1200 UTC
10 100 1000
Particle Diameter, Dp (nm)
1 10 100 1000
Particle Diameter, Dp (nm)
100
,« 80 -
6
Z_ 60 -
| 4°-
5 20 -
z
•° o
September 24,1991, 0800 UTC
1 10 100 1000
Particle Diameter, Dp (nm)
Figure 2-14. Examples of the measured 1-h average particle number size distributions
and the log normal fits to the modes of the data. Squares are measured data,
solid lines are the fitted lognormal modes determined by DistFit™. These
modes (nucleation between 3 and 20 nm, Aitken between 20 and 100 nm, and
accumulation above 100 nm) can be observed in most examples.
Source: Covert etal. (1996).
(Woo et al., 200 la), perhaps due to unusually high rates of production of condensable species,
suggests that high concentrations of ultrafine particles may be a normal occurrence in polluted
urban areas.
2-34
-------
2.06+5
E
u
Q
D)
_O
5
•D
1.56+5-
1.06+5 -
5.06+4 -
1.26+5
1.06+4
8.06+3
2.06+3
00
= 40.1 nm
|jg = 15.7 nm
crn = 1.37
60m
Down
10 100
Particle Diameter Dp (nm)
6.06+4
5.06+4
150m
Down
10 100
Particle Diameter Dp (nm)
= 20.3nm
<-x ..
"^ * •
|jg = 75.4 nm
0%, = 1.56
4V
300m
Down
1.06+4
S.Oe+3
•o"~"
JJ
"x 6.0e+3
§> 4.0e+3
5
Z
2.06+3
o.o
10 100
Particle Diameter Dp (nm)
|jg = 41.3 nm
10 100
Particle Diameter Dp (nm)
|jg = 64.7 nm
|jg = 14.6 nm
V a =2.23
300m
10 100
Particle Diameter Dp (nm)
10 100
Particle Diameter Dp (nm)
Figure 2-15a-f. Fitted multi-model particle size distribution at different sampling
distances from Freeway 405: (a) 30 m downwind, (b) 60 m downwind,
(c) 90 m downwind, (d) 150 m downwind, (e) 300 m downwind,
(f) 300 m upwind. Size distributions were normalized to the control
CPC's reading. Note different scales for dN/d log Dp axis. Modal
parameters given are: geometric mean diameter, jig; and geometric
standard deviation, og.
Source: Zhu et al. (2002b).
2-35
-------
1.66+5
ST 1-2e+5
.O
d
Q
O)
8.0e+4
4.0e+4
0.0
30 m
90 m
60 m
150 m
300 m upwind
10 100
Particle Diameter, Dp (nm)
Figure 2-15g. Combination of Figures 2-15(a-e), with dN/d logDp scale. Ultrafine
particle size distribution at different sampling locations near Freeway 405
in Los Angeles, CA.
Source: Zhu et al. (2002b).
Concentration of Ultrafine Particles: A Balance Between Formation and Removal
Nuclei-mode particles may be removed by dry deposition or by their growth into the
accumulation mode. Such growth takes place as other low vapor pressure material condenses
on the particles or as nuclei-mode particles coagulate with themselves or with accumulation
mode particles. Because the rate of coagulation would vary with the concentration of
accumulation-mode particles, it might be expected that the concentration of nuclei-mode
particles would increase with a decrease in accumulation-mode mass. On the other hand, the
concentration of particles would be expected to decrease with a decrease in the rate of generation
of particles by reduction in emissions of metal and carbon particles or a decrease in the rate of
generation of H2SO4 or condensable organic vapor. The rate of generation of H2SO4 depends on
the concentrations of SO2 and the hydroxyl radical (»OH), which is generated primarily by
reactions involving ozone (O3). Thus, reductions in SO2 and O3 would lead to a decrease in the
2-36
-------
rate of generation of H2SO4 and condensable organic vapor and to a decrease in the
concentration of nuclei-mode particles. The balance between formation and removal is
uncertain. However, these processes can be modeled using a general dynamic equation for
particle size distribution (Friedlander, 2000) or by aerosol dynamics modules in newer air
quality models (Binkowski and Shanker, 1995; Binkowski and Ching, 1995).
2.1.3 Chemistry of Atmospheric Particulate Matter
The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
ions; particle-bound water; elemental carbon; a great variety of organic compounds; crustal
material; and (at coastal locations) sea salt. Atmospheric PM also contains a large number of
elements in various compounds and concentrations. More information and references on the
composition of PM measured in a large number of studies in the United States, may be found in
the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a). Also, in this document,
the concentrations and composition of ambient PM are discussed in Chapter 3 (Section 3.1), and
ambient data for concentrations and composition of PM25 are given in Appendices 3 A, 3B,
and 3C.
2.1.3.1 Chemical Composition and Its Dependence on Particle Size
Studies conducted in most parts of the United States indicate that sulfate, ammonium, and
hydrogen ions; elemental carbon, secondary organic compounds, and primary organic species
from cooking and combustion; and certain transition metals are found predominantly in the fine
particle mode. Crustal materials such as calcium, aluminum, silicon, magnesium, and iron are
found predominately in the coarse particles. Some primary organic materials such as pollen,
spores, and plant and animal debris are also found predominantly in the coarse mode. Certain
components such as potassium and nitrate may be found in both the fine and coarse particle
modes, but they originate from different sources or mechanisms. Potassium in coarse particles
comes from soil. Potassium in fine particles originates in emissions from burning wood or
cooking meat. Nitrate in fine particles comes primarily from the reaction of gas-phase nitric acid
with gas-phase ammonia forming particulate ammonium nitrate. Nitrate in coarse particles
comes primarily from the reaction of gas-phase nitric acid with preexisting coarse particles.
2-37
-------
2.1.3.2 Primary and Secondary Particulate Matter
Particulate matter can be primary or secondary. PM is called "primary" if it is in the same
chemical form in which it was emitted into the atmosphere, but it is called "secondary" if it is
formed by chemical reactions in the atmosphere. Primary coarse particles are usually formed by
mechanical processes and include windblown dust, sea salt, road dust, and combustion-generated
particles such as fly ash and soot. Primary fine particles are emitted from sources either directly
as particles or as vapors that rapidly condense to form ultrafine or nuclei-mode particles. This
includes soot from diesel engines, a great variety of organic compounds condensed from
incomplete combustion or cooking, and compounds of As, Se, Zn, etc. that condense from vapor
formed during combustion or smelting. The concentration of primary particles depends on their
emission rate, transport and dispersion, and removal rate from the atmosphere.
Secondary PM is formed by chemical reactions of free, adsorbed, or dissolved gases. Most
secondary fine PM is formed from condensable vapors generated by the chemical reactions of
gas-phase precursors. Secondary formation processes can result in either the formation of new
particles or the addition of PM to preexisting particles. Most of the sulfate and nitrate and a
portion of the organic compounds in atmospheric particles are formed by chemical reactions that
occur in the atmosphere. Secondary aerosol formation depends on numerous factors, including
the concentrations of precursors; the concentrations of other gaseous reactive species such as
ozone, hydroxyl radical, peroxy radicals, or hydrogen peroxide; atmospheric conditions
including solar radiation and RH; and the interactions of precursors and preexisting particles
within cloud or fog droplets or in the liquid film on solid particles. As a result, it is considerably
more difficult to relate ambient concentrations of secondary species to sources of precursor
emissions than it is to identify the sources of primary particles. A significant effort is currently
being directed toward the identification and modeling of organic products of photochemical
smog, including the conversion of gases to PM. More information of the transformation of
precursor gases into secondary PM is given in Chapter 3 (Section 3.3.1).
Particle strong acidity is due almost entirely to the presence of H2SO4 or NH4HSO4. Thus,
the acidity of atmospheric particles depends on both the amount of SO2 that is oxidized to SO3
and subsequently forms H2SO4 as well as the amount of ammonia available to react with the
sulfuric acid. Nitric acid is more volatile than sulfuric acid. Thus, if gas phase SO3 or sulfuric
acid or particles containing H2SO4 or NH4HSO4 contact particles containing NH4NO3, nitric acid
2-38
-------
gas will be released with the remaining ammonia contributing to the further neutralization of the
acid. Little NH4NO3 is found in atmospheres containing significant particle strong acidity.
However, as SO2 emissions are reduced to the point that there is more than enough ammonia to
neutralize the sulfuric acid, NH4NO3 particles will begin to form. Thus, ammonia emissions and
concentrations relative to those of SO2 and H2SO4 are important in determining the strong acidity
in the atmosphere and the concentration of particulate NH4NO3. Therefore, once SO2 emissions
have been reduced to the point that ammonia and sulfate are in balance to form (NH4)2SO4,
further reductions in SO2 will not result in an equivalent reduction in airborne PM, because
one (NH4)2SO4 unit will be replaced by two NH4NO3 units.
2.1.3.3 Particle-Vapor Partitioning
Several atmospheric aerosol species, such as water, ammonium nitrate and certain organic
compounds, exist in an equilibrium between gaseous and condensed phases and are called
semivolatile. The equilibrium between water vapor and particle-bound water is well understood
and can be modeled accurately. A variety of thermodynamic models have also been developed
to predict the temperature and relative humidity dependence of the ammonium nitrate equilibria
with gaseous nitric acid and ammonia. However, under some atmospheric conditions (such as
cool, cold, or very clean air), the relative concentrations of the gas and solid phases are not
accurately predicted by equilibrium considerations alone, and transport kinetics can be
important. The gas-particle distribution of semivolatile organic compounds depends on the
equilibrium vapor pressure of the compound, total particle surface area, particle composition,
atmospheric temperature, and relative humidity. Although it generally is assumed that the gas-
particle partitioning of semivolatile organics is in equilibrium in the atmosphere, neither the
equilibria nor the kinetics of redistribution are well understood. Diurnal temperature fluctuations
cause gas-particle partitioning to be dynamic on a time scale of a few hours and can cause
semivolatile compounds to evaporate during the sampling process. The pressure drop across the
filter can also contribute to the loss of semivolatile compounds. The dynamic changes in
gas-particle partitioning caused by changes in temperature, pressure, and gas-phase
concentration, both in the atmosphere and after collection, cause serious sampling problems that
are discussed in Section 2.2.3, Measurement of Semivolatile Particulate Matter.
2-39
-------
Equilibria with Water Vapor
As a result of the equilibrium of water vapor with liquid water in hygroscopic particles,
many ambient particles contain liquid water (particle-bound water). Unless removed, this
particle-bound water will be measured as a component of the particle mass. Particle-bound
water plays an important role in atmospheric chemistry and physics. It influences the size of
particles, and in turn, the light scattering and aerodynamic properties of particles. Particle
aerodynamic properties are important in dry deposition to surfaces, deposition to airway surfaces
during breathing, and deposition in sampling instrumentation. The aqueous solution provides a
medium for reactions of dissolved gases, including reactions that do not take place in the gas
phase. The aqueous solutions also may act as a carrier to convey soluble toxic species to the
gas-exchange regions of the respiratory system, including species that would be removed by
deposition in the upper airways if no particles were present. An extensive review of the
equilibrium of water vapor with particle-bound water as it pertains to ambient aerosols was
given in Chapter 3 of the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a).
The interaction of particles with water vapor may be described briefly as follows.
As relative humidity increases, particles of crystalline soluble salts, such as (NH4)2SO4,
NH4HSO4, or NH4NO3, undergo a phase transition to become aqueous solution particles.
According to the phase rule, for particles consisting of a single component, this phase transition
is abrupt, taking place at a relative humidity that corresponds to the vapor pressure of water
above the saturated solution (the deliquescence point). With a further increase in relative
humidity, the solution particle adds water (and the concentration of the solute decreases) so that
the vapor pressure of the solution is maintained equal to that of the surrounding relative
humidity; thus, the solution particle tends to follow the equilibrium growth curve. As relative
humidity decreases, the solution particle follows the equilibrium curve to the deliquescence
point. However, rather than crystallizing at the deliquescence relative humidity, the solute
remains dissolved in a supersaturated solution to considerably lower relative humidities.
Ultimately, the solution particle abruptly loses its water vapor (efflorescence) and typically
returns to the initial crystalline form.
For particles consisting of more than one component, the solid to liquid transition will take
place over a range of RHs, with an abrupt onset at the lowest deliquescence point of the several
components. Subsequent growth will occur as crystalline material in the particle dissolves
2-40
-------
according to the phase diagram for the particular multicomponent system. Under such
circumstances, a single particle may undergo several more or less abrupt phase transitions until
the soluble material is fully dissolved. At decreasing relative humidity, such particles tend to
remain in solution to relative humidities well below the several deliquescence points. In the case
of the sulfuric acid-ammonium sulfate-water system, the phase diagram is fairly well
understood. For particles of composition intermediate between NH4HSO4 and (NH4)2SO4, this
transition occurs in the range from 40% to below 10% relative humidity, indicating that for
certain compositions the solution cannot be fully dried in the atmosphere. At low relative
humidities, particles of this composition would likely be present in the atmosphere as
supersaturated solution droplets (liquid particles) rather than as solid particles. Thus, they would
exhibit hygroscopic rather than deliquescent behavior during relative humidity cycles.
Other pure compounds, such as sulfuric acid, are hygroscopic (i.e., they form aqueous
solutions at any relative humidity and maintain a solution vapor pressure over the entire range of
relative humidity). Soluble organic compounds may also contribute to the hygroscopicity of the
atmospheric aerosol (Saxena et al., 1995; Saxena and Hildeman, 1996), but the equilibria
involving organic compounds and water vapor, and, especially for mixtures of salts, organic
compounds, and water, are not so well understood. These equilibrium processes may cause an
ambient particle to significantly increase its diameter at relative humidities above about 40%
(Figure 2-16a). A particle can grow to five times its dry diameter as the RH approaches 100%
(Figure 2-16b). The Federal Reference Methods, for filter measurements of PM25 and PM10
mass, require equilibration at a specified, low relative humidity after collection (for PM2 5,
between 30 and 40% RH with control of ± 5% RH [Code of Federal Regulations, 2001a]).
This equilibration removes much of the particle-bound water and provides a relatively stable PM
mass for gravimetric measurements (see Section 2.2 for details and references). Otherwise,
particle mass would be a function of relative humidity, and the particle mass would be largely
particle-bound water at higher relative humidities. However, some particle-bound water may be
retained even after equilibration. Recent studies have shown that significant amounts of particle-
bound water are retained in particles collected on impaction surfaces even after equilibration and
that the amount of retained particle-bound water increases with relative humidity during
collection (Hitzenberger et al., 1997).
2-41
-------
2.0-
o
Q
re
a:
2
o
i_
0)
•5
re
Q
1.5-
1.0-
Hygroscopic Growth
Curve for H2SO4
/
Hysteresis Loop
for(NH4)2SO4
H
30 50 70
Relative Humidity RH (%)
- 7
- 3
- 2
- 1
90
Q.
-5 .2
- 4
0)
E
Figure 2-16a. 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.
Source: National Research Council (1993) adapted from Tang (1980).
The retention of particle-bound water is a greater problem for continuous monitors, which
measure changes in mass collected on a filter over long sampling times. If particle-bound water
is not removed, changes in relative humidity would cause changes in the mass of PM collected
over previous hours or days. These changes could be much greater than amount of PM mass
added in one hour. Therefore, continuous monitoring techniques generally attempt to remove
particle-bound water before measurement either by heating or dehumidification. However, other
semivolatile materials (e.g., ammonium nitrate and organic compounds) that may be partially
2-42
-------
216
I I
Theoretical Prediction at 22 °C
O O O O O Experimental Measurements
0 50 100 150 200
NH4HSO4 Dry Particle Diameter, Dp (nm)
Figure 2-16b. Theoretical predictions and experimental measurements of growth of
NH4HSO4 particles at relative humidities between 95 and 100%.
Source: Lietal. (1992).
lost during sampling or equilibration of an unheated filter are certainly lost when the collected
sample is heated above ambient temperature. These changes in particle size with relative
humidity also mean that particle measurements such as surface area or volume, or composition
as a function of size, should be made at the same RH in order for the results to be comparable.
These problems are addressed in more detail in Section 2.2, Measurement of Paniculate Matter.
Particle-Bound Water as a Carrier
Water vapor is not a pollutant, and particle-bound water (PBW) is not included in the mass
of PM subject to regulation and control. However, while water is not itself a pollutant, PBW
may act as a carrier for pollutants, as stated by Wilson (1995):
2-43
-------
"It is possible that water-soluble gases, which would be removed by deposition to
wet surfaces in the upper respiratory system during inhalation, could dissolve in
particle-bound water and be carried with the particles into the deep lung. Water-
soluble gases in polluted air include oxidants such as O3, H2O2, and organic
peroxides; acid gases such as SO2, HC1, HNO3, HONO, and formic acid; and polar
organic species such as formaldehyde. Some of these species may have biological
effects, but current techniques do not measure particle-bound water or the species
dissolved in it."
Kao and Friedlander (1995) also suggest that aerosols could carry adsorbed free radicals or
dissolved H2O2 into the lung. Friedlander and Yeh (1998) point out that the aqueous component
of the atmospheric aerosol carries short-lived, reactive chemical species, including hydrogen
peroxide and organic peroxides. Equilibrium calculations and limited experimental data suggest
that aerosol-phase concentrations of hydrogen peroxide will fall within a range (around 1 mM)
in which significant biochemical effects have been observed when respiratory tract epithelial
cells were treated with hydrogen peroxide solutions.
Wexler and Sarangapani (1998) used a physical model of "gas-particle-mucus heat and
mass transport in the human airways," developed in order to investigate the hypothesis that
endogenous ammonia neutralizes aerosol acidity (Sarangapani and Wexler, 1996), to investigate
the transport by particles of soluble vapors to the tracheobronchial and air exchange (also called
alveolar or pulmonary) regions of the lung. Wexler and Sarangapani (1998) provides a concise
description of this process:
"Air pollutants are deposited in the human airway via two pathways — particle
and vapor deposition. In the absence of particles, vapors deposit at different
locations in the lung depending on their solubility in mucus, which is over 99%
water. High-solubility compounds, such as nitric acid or hydrogen peroxide, are
rapidly removed in the upper airways while low-solubility compounds, such as
oxygen or ozone, are less well removed and so penetrate to the alveoli. Pollutant
deposition in the upper airways is less harmful than in the lower airways because
upper airways clearances is more rapid and the epithelium is protected by a mucus
layer. As a result, low-solubility pollutants, such as ozone, may harm the alveoli
while high-solubility pollutants, such as nitric acid, do not reach these tissues.
"In the presence of aerosol particles, this scenario changes. Under most ambient
conditions, aerosol particles contain some liquid water so that soluble compounds
are partitioned between the gas phase and the aerosol liquid-water phase. The
degree of deposition via the gas compared to that via the particles is a function of a
number of factors including the solubility of the compound and the liquid-water
content of the aerosol. Since highly soluble compounds deposit in the upper
2-44
-------
airways, particles may provide a vector for deposition of these compounds in the
lower airways. Lower-solubility compounds may persist in the vapor phase and so
may deposit in lower airway segments.
"When both a water-soluble vapor and an aerosol liquid phase are present together,
some of the vapor will dissolve in the liquid-water phase until an equilibrium is
established. The equilibrium condition is given by the Henry's law coefficient,
H (M atnf:), which gives the relationship between the partial pressure of the
vapor, typically in units of atmospheres, and the concentration of the dissolved
compound in the aqueous phase, typically in units of molal.
"Direct deposition of the vapor is primarily a function of its solubility. Since the
airways are lined with mucus, deposition is higher for compounds with a higher
solubility than for those with lower solubility. Compounds with Henry's law
constant (H) greater than 100 M atnT1 deposit relatively uniformly over the first
10 generations (e.g., formaldehyde and hydrogen peroxide [and sulfur (IV) {SO2
(aq) plus H2SO3 (aq) plus HSO3 (aq)}], so they affect health via interactions with
the upper airway epithelia. All the vapor is deposited iftf is greater than about
0.1. Low-solubility compounds do not deposit effectively in the mucus until past
the 20th generation so are still present in significant concentrations in the
pulmonary region. These compounds (e.g., ozone, nitrogen dioxide) affect human
health via interactions in the air-exchange regions.
"As aerosols are inhaled, soluble vapors deposit on the mucus, disrupting the
gas-particle equilibrium, and the compound begins to evaporate from the aerosol
particles. If the evaporation is rapid, the pattern of deposition of the compound
will not be influenced by the presence of the particles, i.e., the deposition pattern is
essentially that of the vapor alone. If the evaporation is very slow compared to a
breathing cycle, a significant amount of the compound will remain in the particle
phase and the pattern of deposition may be shifted toward the pulmonary region by
the particle."
Hygroscopic particles, present in the inhaled aerosol disrupt the gas-particle equilibria.
The soluble gas will deposit on the mucus, reducing the gas phase vapor pressure, and reducing
the equilibrium concentration of the gas in solution. However, in the higher relative humidity of
the lung, hygroscopic particles will add PBW and grow in volume. Therefore, even though the
concentration of the soluble gas may be lower, there will be a greater amount of the soluble gas
in solution. The amount of the soluble gas in the particle will depend on the Henry's Law
coefficient of the gas, the size of the particle, and the position of the particle in the respiratory
tract. Figure 2-17 shows how the concentration of the soluble gas (H= 104 M atirT1) in the PBW
of the aerosol relative to that in the gas phase changes as the particle moves deeper into the lung.
2-45
-------
1.5
Q.
O
O
0.5
Airway Vapor Deposition via Particles
—•— D = 0.1 |jm
— *— D = 0.3 |jm
- B - D = 0.7 |jm
--X-- D = 1.0 |jm
H
\
10 15
Generation
20
25
Figure 2-17. Concentration of the dissolved gas in the particle normalized by its initial
concentration as a function of airway generation for H= 104 M atnT1 for
particle diameters of 0.1, 0.3, 0.7, and 1 um.
Source: Wexler and Sarangpani (1998).
In similar figures for other values of//, Wexler and Sarangapani (1998) show that,
regardless of particle size, the greatest relative increase in the amount of soluble gas dissolved in
PBW (a factor of 10 to 100) occurs for lower solubility gases (H< 0 M atnT1). However, the
deposition of such gases by particles will be small since the concentration of dissolved gas will
be low. For higher solubility gases (H= 102 M atnT1), the enhancement may be 2- to 20-fold,
depending on particle size. There may be a sufficient amount of the soluble gas in the PBW to
influence the deposition pattern, with more soluble gas being transported to higher generations
(deeper into the lung). For the highest solubility gases, the enhancement in less than 2, but a
larger amount of gas is dissolved in the PBW. The highest solubility gases (H> 104 M atnT1)
will evaporate from particles < 0.1 jim before the particles reach the air exchange region of the
lung. However, as shown in Figure 2-17, particles > 0.3 jim can efficiently transport high
solubility gases into the air exchange region.
2-46
-------
Wexler and Sarangapani (1998) observed that even for a high liquid water content
(PBW = 100 |ig/m3) and for a highly soluble gas such as H2O2 (H= 10s M atnT1), only about 1%
of the soluble gas will dissolve in the PBW. Therefore, any influence of the particles on the gas
concentration will cause little change in the quantity of soluble gas deposited on the upper
airways. The Wexler and Sarangapani (1998) analysis was conducted under the assumption that
equilibrium with the gas phase is obtained quickly and considered only the "physical" solubility
of the soluble gas. For example, their discussion of SO2 assumes an H of 1 M atnT1 which is
correct for the equilibrium of SO2 (aq) with SO2 gas. However, SO2 reacts chemically with
water to yield SO2 (aq), H2SO3 (aq), and HSO 3 (aq). The H for sulfur (IV) in solution in
equilibrium with SO2 gas is dependent on the pH of the solution but may reach 104 M atnT1 at a
pH of 6 (Schwartz, 1984). Thus, if SO2 (aq), H2SO3 (aq), and HSO3 (aq) are in equilibrium in
the liquid phase, and any one of the three species is toxic, significantly more toxic sulfur (IV)
will be carried to the air exchange region than predicted by Wexler and Sarangapani (1998).
Friedlander and Yeh (1998) point out that H2O2 forms adducts called peroxohydrates with many
substances in which it acts like water of hydration (Elvers et al., 1991). If such adducts form in
the PBW of atmospheric aerosols, a much higher percentage of the atmospheric peroxide would
dissolve in the PBW and be carried into the air exchange region of the lungs.
Thus, soluble gases, like H2O2, SO2, and formaldehyde, which would be completely
removed in the upper airways in the absence of particles, can dissolve in the PBW of
hygroscopic particles and be transported by the particles into the air exchange (alveolar or
pulmonary) region of the lungs.
Friedlander and Yeh (1998) also suggested that the epidemiologic associations found
between adverse health effects and sulfate or PM mass may represent a response to atmospheric
peroxides, or other toxic substances dissolved in PBW. Friedlander and Yeh (1998) stated,
"This hypothesis supports reduction of the total submicron aerosol mass as away
to reduce adverse health effects, because the total submicron mass is closely linked
to the aqueous component that carries the reactive species. To test this hypothesis,
studies are needed of the effects of exposures of cellular layers and/or animals to
submicron hydrogen peroxide-containing aerosols that also contain salts such as
ammonium sulfate."
2-47
-------
Mono et al. (2001) conducted an inhalation study in which rats were exposed to (NH4)2SO4
(429 or 215 |ig/m3; 0.3 to 0.4 jim mass median diameter) or H2O2 (10, 20, or 100 ppb) alone or in
combination for 2 h. The authors summarized the study as follows.
"In summary, the present studies find that exposure of rats to (NH4)2SO4 + H2O2
results in tissue damage and altered alveolar macrophage activity. Our findings
suggest that macrophage-derived TNF-oc and peroxynitrite are potential factors
contributing to the toxicity of (NH4)SO4 + H2O2. It is possible that alterations in
macrophage functioning following inhalation of (NH4)2SO4 + H2O2 underlie
increases in individual susceptibility to infection following the fine PM exposure."
The potential for PBW to carry toxic gases into the air exchange region of the lung
suggests that it would be useful to measure both the amount of PBW in the atmospheric aerosol
and the composition of the material dissolved in it. Although measurements have been made of
the amount of PBW (discussed in 2.2.3.1), there is little information on the composition or
concentration of soluble species that would evaporate as the PBW evaporates. Hung and Wang
(2001) measured the concentrations of reactive oxygen species (ROS) in various size fractions of
atmospheric aerosols and vehicle emissions in Taipei using a technique which would respond to
ROS adsorbed on particle surfaces or dissolved in PBW. The ROS were not identified, but the
ROS concentration was correlated with the atmospheric ozone concentration. The ROS
concentrations were highest in the 0.18 to 1.0 jim particle size range and decreased by over 50%
between analyses at 1 h and 115 h after collection. While this study does not demonstrate that
PBW carries H2O2, it does demonstrate the ability of particles to act as carriers of reactive
species and offers a technique for the measurement of ROS absorbed on particles or dissolved in
PBW.
PBW Effects on Visibility
Light scattering, and the resultant effects on visibility, depend on the size and refractive
index of the particle. PBW contributes to light scattering just like any other component of fine
PM. As discussed in detail in Chapter 4, the light scattering due to a given mass of PM can be
estimated from the composition and the relative humidity using mass scattering coefficients for
the chemical components that are higher for hygroscopic components.
2-48
-------
PEW Effects on Dosimetry
As particles are inhaled, they will be exposed to the high relative humidities of the
respiratory tract. Although supersaturation may occur in the early generations of the airways, by
the time the aerosol reaches the air exchange region, the relative humidity will have reached
99.5% and the particles will have attained their equilibrium size (Sarangapani and Wexler,
1996). As discussed in Chapter 6, these size changes can have significant effects on the
fractions of inhaled particles deposited in the various portions of the respiratory tract.
PEW as a Reaction Medium
If the air becomes saturated with water vapor, hygroscopic particles will grow into cloud
droplets. The liquid water content of cloud droplets is much higher than that of aerosols at
relative humidities below 100%. This liquid water provides a reaction medium where reactions
can take place that do not take place in the air. For example, dissolved SO2 can react with
dissolved O3 or dissolved H2O2 to form H2SO4 in cloud droplets (Lazrus et al., 1983; Schwartz,
1984, 1986; Martin, 1994; Hegg and Hobbs, 1982, 1986).
The chemistry of fog and cloud droplets is important in the formation of secondary
pollutants. An extensive discussion was given in Section 3.3.1.4 of the 1996 PM AQCD (U.S.
Environmental Protection Agency, 1996a). The possibility of chemical reactions in PBW is
controversial because of the much lower quantity of liquid water at lower relative humidities
(Brock and Durham, 1984; McMurray and Wilson, 1982, 1983; McMurray et al., 1981).
2.1.3.4 Atmospheric Lifetimes and Removal Processes
The lifetimes of particles vary with size. Nuclei-mode particles rapidly grow into the
accumulation mode. However, the accumulation mode does not grow into the coarse mode.
Accumulation-mode fine particles are kept suspended by normal air motions and have very low
deposition rates to surfaces. They can be transported thousands of km and remain in the
atmosphere for a number of days. Coarse particles can settle rapidly from the atmosphere
(within hours) and normally travel only short distances. However, when mixed high into the
atmosphere, as in dust storms, the smaller-sized coarse-mode particles have longer lives and
travel greater distances. Dry deposition rates are expressed in terms of a deposition velocity that
varies with particle size, reaching a minimum between 0.1 and 1.0 jim aerodynamic diameter
2-49
-------
(e.g., Lin et al., 1994). Accumulation-mode particles are removed from the atmosphere
primarily by cloud processes. Fine particles, especially particles with a hygroscopic component,
grow as the relative humidity increases, serve as cloud condensation nuclei, and grow into cloud
droplets. If the cloud droplets grow large enough to form rain, the particles are removed in the
rain. Falling rain drops impact coarse particles and remove them. Ultrafme or nuclei-mode
particles are small enough to diffuse to the falling drop, be captured, and be removed in rain.
Falling rain drops, however, are not nearly as effective in removing accumulation-mode particles
as the cloud processes mentioned above. A more detailed discussion of particle deposition,
including acid deposition, especially as it applies to deposition to vegetation, soil, and water
surfaces, is given in Chapter 4 (Environmental Effects of Airborne Paniculate Matter). Acid
deposition and PM are intimately related, first, because particles contribute to the acidification of
rain and, secondly, because the gas-phase species that lead to dry deposition of acidity are also
precursors of particles. Therefore, reductions in SO2 and NOX emissions will decrease both
acidic deposition and PM concentrations.
Sulfate, nitrate, and some partially oxidized organic compounds are hygroscopic and act as
nuclei for the formation of cloud droplets. These droplets serve as chemical reactors in which
(even slightly) soluble gases can dissolve and react. Thus, SO2 can dissolve in cloud droplets
and be oxidized to sulfuric acid by dissolved ozone or hydrogen peroxide. These reactions take
place only in aqueous solution, not in the gas phase. Sulfur dioxide may also be oxidized by
dissolved oxygen. This process will be faster if metal catalysts such as iron or manganese are
present in solution. If the droplets evaporate, larger particles are left behind. If the droplets
grow large enough, they will fall as rain; and the particles will be removed from the atmosphere
with potential effects on the materials, plants, or soil on which the rain falls. (Similar
considerations apply to dew.) Atmospheric particles that nucleate cloud droplets also may
contain other soluble or nonsoluble materials such as metal salts and organic compounds that
may add to the toxicity of the rain. Sulfuric acid, ammonium nitrate, ammonium sulfates, and
organic particles also are deposited on surfaces by dry deposition. The utilization of ammonium
by plants leads to the production of acidity. Therefore, dry deposition of particles can also
contribute to the ecological impacts of acid deposition. These effects are discussed in Chapter 4
(Section 4.1).
2-50
-------
2.1.4 Comparison of Fine and Coarse Particles
The physical and chemical properties of fine particles (including ultrafine particles and
accumulation-mode particles) and coarse particles are summarized for comparison purposes in
Table 2-2. These include important differences in sources, formation mechanisms, composition,
atmospheric residence time, removal processes, and travel distances. Ensuing chapters in this
document will also show that fine and coarse particles differ in aspects of concentrations,
exposure, dosimetry, toxicology, and epidemiology. Collectively, these differences continue to
warrant consideration of fine particles as a separate air pollutant class from coarse particles and
the setting of separate standards for fine and coarse particles.
2.2 MEASUREMENT OF PARTICULATE MATTER
The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a) summarized the
sampling and analytical techniques for PM and acid deposition that had appeared in the literature
since the earlier 1982 PM AQCD (U.S. Environmental Protection Agency, 1982). Excellent
reviews have also been published by Chow (1995) and McMurry (2000). This section discusses
problems in measuring PM; new techniques that attempt to alleviate these problems or measure
problem species; Federal Reference Methods, speciation monitors, analytical methods for
inorganic elements, organic and elemental carbon, and ionic species; and continuous and
multiday monitors.
2.2.1 Particle Measurements of Interest
There are many PM components and parameters that are of interest across the various
types of uses to which PM measurement data are applied. These uses include analyses of
compliance with air quality standards and trends; source category apportionment studies related
to the development of pollution reduction strategies and the validation of air quality models;
studies related to health, ecological, and radiative effects; and characterization of current air
quality for presentation to the public in the context of EPA's Air Quality Index. Particulate
matter measurement components and parameters of specific interest for these various purposes
are noted below and summarized in Table 2-3.
2-51
-------
TABLE 2-2. COMPARISON OF AMBIENT PARTICLES,
FINE PARTICLES (Ultrafine plus Accumulation-Mode) AND COARSE PARTICLES
Fine
Ultrafine
Accumulation
Coarse
Formation
Processes:
Formed by:
Composed
of:
Solubility:
Sources:
Combustion, high-temperature
processes, and atmospheric reactions
Nucleation
Condensation
Coagulation
Sulfate
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
Diffuses to raindrops
-------
TABLE 2-3. 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 [PM25] and coarse thoracic PM mass [PM10_25]) 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
Particle measurements are needed to determine if a location is in compliance with air
quality standards, to determine long-term trends in air quality patterns, and for use in
epidemiologic studies. For these purposes, the precision of the measurements made by a variety
of measurement instruments in use is a critical consideration. Therefore, the intercomparisons of
various samplers under a variety of atmospheric and air quality conditions are important.
2-53
-------
In order to reduce pollution to attain a standard, pollution control agencies and national
research organizations need measurements to identify source categories and to develop and
validate air quality models. For these purposes, PM parameters other than mass, such as
chemical composition and size distribution, must also be measured. Moreover, measurements
are needed with shorter time resolutions in order to match changes in pollution with the diurnal
changes in the boundary layer.
A number of PM measurements are needed for use in epidemiologic and exposure studies
and to determine components of PM to guide the planning and interpretation of toxicologic
studies. Thus, size and chemical composition measurements are important, as are measurement
across different time intervals. For epidemiologic studies of acute (i.e., short-term) PM
exposures, 1-h or continuous measurements can provide important information beyond that
provided by 24-h measurements. However, for epidemiologic studies of chronic PM exposures,
measurements that permit integration over longer intervals (e.g., a week to a month) are more
relevant. For dosimetric studies and modeling, information will be needed on the particle size
distribution and on the behavior of particles as the relative humidity and temperature changes
found in the atmosphere are increased to those found in the respiratory system.
For studies of ecological effects and materials damage, measurements of particles and of
the chemical components of PM in rain, fog, and dew are needed to understand the contributions
of PM to the soiling of surfaces and damage to materials and to understand the wet and dry
deposition of acidity and toxic substances to surface water, soil, and plants. Some differentiation
into particle size is needed to determine dry deposition.
For studies of visibility impairment and radiative effects, information is needed regarding
how particles scatter and absorb light, including data on refractive index, ratio of scattering to
absorption, size distribution, and change in particle size with change in relative humidity.
2.2.2 Issues in Measurement of Particulate Matter
The EPA decision to revise the PM standards by adding daily and yearly standards
for PM2 5 has led to a renewed interest in the measurement of atmospheric particles and to a
better understanding of the problems in developing precise and accurate measurements of
particles. It is very difficult to measure and characterize particles suspended in the atmosphere;
however, numerous improvements in PM monitoring are in use and others are in development.
2-54
-------
EPA's PM standards are based, in part, on epidemiologic relationships between health effects
and PM concentrations as measured with existing monitoring methods. As understanding of
suspended PM has advanced and new monitoring information has become available, EPA has
changed the indicator for the PM NAAQS from TSP to PM10, and added PM2 5. During the
current PM NAAQS review, consideration will be given to a standard for coarse thoracic PM.
The U.S. Federal Reference Methods (FRM) for PM25 and PM10 provide relatively precise
(± 10%) methods for determining the mass of material remaining on a Teflon filter after
equilibration. However, numerous uncertainties remain as to the relationship between the mass
and composition of material remaining on the filter as determined by the FRM measurement
procedure and the mass and composition of material that existed in the atmosphere as suspended
PM. As a result, the EPA defines accuracy for PM measurements in terms of agreement of a
candidate sampler with a reference sampler. Therefore, intercomparison of samplers is very
important in determining how well various samplers agree and how various design choices
influence what is actually measured.
There are eight general areas where choices are made in the design and use of an aerosol
sampler. These include (1) consideration of positive artifacts due to chemical reaction or
adsorption, (2) treatment of semivolatile components, (3) selection of particle size cut
characteristics for the upper cut point, (4) separation of fine and coarse PM, (5) treatment of
pressure, temperature, and relative humidity, (6) time resolution, (7) assessment of the reliability
of the measurement technique, and (8) operation and maintenance procedures needed to maintain
consistent measurements over time. In many cases, choices have been made without adequate
recognition of their consequences. As a result, measurement methods developed by different
organizations may give different results when sampling the same atmosphere even though the
techniques appear to be similar.
2.2.2.1 Artifacts Due to Chemical Reactions
When TSP was collected on glass-fiber filters, the reaction of SO2 (and other acid gases)
with basic sites on the glass fiber or with basic coarse particles on the filter led to the formation
of sulfate (or other nonvolatile salts, e.g., nitrate, chloride). These positive artifacts led to the
overestimation of mass, sulfate, and probably also of nitrate. The metal impurities in the glass
fiber resulted in high background levels that led to low precision in the measurement of trace
2-55
-------
metals. These problems were largely overcome by changing to quartz fiber or Teflon filters and
by the separate collection of PM2 5. However, the possible reaction of acidic gases with basic
coarse particles remains a possibility, especially with PM10 and PM10_2 5 measurements. The
reaction of NH3 with acidic particles, either during sampling or during transportation, storage,
and equilibration remains a problem in areas such as the eastern United States where PM is
frequently acidic. One technique that has been developed to overcome this problem makes use
of a denuder to remove NH3 during sampling and to protect the collected PM from NH3 (Suh
et al., 1992, 1994; Brauer et al., 1991; Koutrakis et al., 1988a,b). However, this technique has
been applied primarily for measurement of particle strong acidity, not for the measurement of
artifact-free ammonium or mass concentrations. In the measurement of particle strong acidity,
basic coarse particles must be separated from acidic fine particles (Koutrakis et al., 1992).
2.2.2.2 Treatment of Semivolatile Components of Particulate Matter
Current filtration-based mass measurements can result in significant evaporative losses,
during and possibly after collection, of a variety of semivolatile components (i.e., species that
exist in the atmosphere in dynamic equilibrium between the condensed phase and gas phase).
Important examples include ammonium nitrate, semivolatile organic compounds, and particle-
bound water. This problem is illustrated in Figure 2-18.
The following approaches, that have been used to address the problem of potentially lost
semivolatile components, will be discussed in more detail in subsequent sections.
(1) Collect/measure all components present in the atmosphere in the condensed phase except
particle-bound water. Examples: Brigham Young absorptive sampler and Harvard
pressure drop monitor. Both require preconcentration of the accumulation mode and
reduction of ambient humidity by diffusion denuder techniques.
(2) Stabilize PM at a specified temperature high enough to remove all, or almost all, particle-
bound water. This results in loss of much of the semivolatile PM. Examples: tapered
element oscillating microbalance (TEOM) operated at 50 °C beta gauge with heated inlet.
(3) Equilibrate collected material at fixed, near-room temperature and moderate relative
humidity to reduce particle-bound water. Accept the loss of an unknown but possibly
significant fraction of semivolatile PM. Examples: U.S. Federal Reference Method and
most filter-weighing techniques. Equilibration originally was designed to remove
adsorbed water vapor from glass fiber filters in order to maintain a stable filter weight.
2-56
-------
Q.
Q
w
TO
E
Should be
Retained
Particle-Bound Water
Should be Removed
(NH4)XS0
X = 0 to 2/ii-
(NH4)XS04
X = 0 to 2
Elemental Carbon
Mineral/Metal
0.1 1.0 2.5
Aerodynamic Diameter, Da (pm)
Semivolatile components subject to evaporation during or after sampling
Figure 2-18. 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.
The designated RH (35 ± 5%) was a compromise. If the RH is too low, electrostatic
charging becomes a problem. The equilibration process does help provide a stable and
reproducible mass. The equilibrium process reduces the amount of particle-bound water
but may not remove all particle-bound water. Moreover, the equilibration process may
lead to the loss of other semivolatile PM components.
2.2.2.3 Upper Cut Point
The upper cut point of the high volume sampler varies with wind speed and direction.
However, it is usually desirable to have an upper cut point that is independent of these factors.
Considerations in selecting an upper cut point are discussed in Section 2.1.2.2.
2-57
-------
2.2.2.4 Cut Point for Separation of Fine and Coarse Particulate Matter
As shown in Table 2-2, fine and coarse particles differ not only in size but also in
formation mechanisms, sources, and chemical, physical, and biological properties. They also
differ in concentration-exposure relationships, dosimetry (deposition and retention in the
respiratory system), toxicity, and health effects as observed in epidemiologic studies. Thus, it is
desirable to measure fine and coarse PM separately in order to properly allocate health effects to
either fine or coarse PM and to correctly determine sources by receptor modeling approaches.
Considerations in selecting a cut point to separate fine and coarse particles are discussed in
Section 2.1.2.2.
2.2.2.5 Treatment of Pressure, Temperature, and Relative Humidity
There are a variety of techniques for defining (or ignoring) the pressure, temperature, and
relative humidity during and after sampling. For example, the sample volume may be based on
the mass or volumetric flow corrected to standard temperature and pressure (298 K and 1 atm)
(as in the current FRM for PM10), or it may be based on the volumetric flow at ambient
conditions of temperature and pressure (current FRM for PM2 5).
There are also a variety of options for the control of temperature during collection. The
particles may be heated enough to remove much of the particle-bound water (i.e., TEOM at
50 °C); the particles may be heated several degrees, just enough to prevent condensation of
water in the sampling system; the particles and the sampler may be maintained near ambient
temperature (± 5 °C of ambient temperature is required for EPA FRM samplers for PM25);
or the particles and sampler may be maintained at a constant temperature inside a heated or air
conditioned shelter. There are also options for controlling of temperature after collection:
(a) no control (room temperature) or (b) ship and store at cool temperature (4 °C is the current
EPA FRM requirement).
Consideration must also be given to relative humidity (RH). Changes in RH cause changes
in particle size of hygroscopic or deliquescent particles. Changing RH by adding or removing
water vapor affects particle number, particle surface area, and particle size distribution
measurements and the amount of overlap of accumulation-mode and coarse-mode particles.
Changing RH by intentional or inadvertent changes in temperature also affects the relative loss
of ammonium nitrate and semivolatile organic compounds. Monitoring personnel should be
2-58
-------
aware of the various options for the treatment of pressure, temperature, and RH factors; make
appropriate selections; and document which options are used.
Studies of relationships between personal/indoor/outdoor measurements present special
problems. Indoor environments are typically dryer than outdoors and may be warmer or, if
air-conditioned, cooler. These differences may change particle size and the amount of
volatilization of semivolatile components. Such changes between indoors and outdoors will
complicate the comparison of indoor to outdoor PM concentrations, the modeling of personal
exposure to all particles, and apportionment of personal exposure into particles of ambient
origin, particles of indoor origin, and particles originating from personal activity.
2.2.2.6 Time Resolution
The classic 24-h filter collection technique is being supplemented by a variety of
continuous monitors for various PM constituents. This process is being accelerated by the lower
operational cost of continuous monitors and the availability of new continuous monitors for
mass, number, and certain chemical components, as well as refinements of older methods based
on beta attenuation or light scattering. Most epidemiologic studies have used 24-hour
concentrations as exposure indicators. However, one epidemiologic study of chronic effects
uses a filter sampler with a 2-week collection period (Gauderman et al., 2000). Another recent
study used 1 to 2 h concentrations (see Peters et al., 2000). Continuous methods are discussed in
Section 2.2.5.
2.2.2.7 Accuracy and Precision
Precision is typically determined by comparing measurements obtained with collocated
samplers or through replicate analyses; whereas accuracy is determined through the use of
traceable calibration standards. Unfortunately, no standard reference calibration material or
procedure has been developed for suspended, atmospheric PM. It is possible to determine the
accuracy of certain components of the PM measurement system (e.g., flow control, inlet
aspiration, PM2 5 cut, weighing, etc.). The absolute accuracy for collecting a test aerosol can also
be determined by isokinetic sampling in a wind tunnel. However, it is not currently feasible to
provide a simulated atmospheric aerosol with naturally occurring semivolatile components. It is
particularly challenging to develop an atmospheric aerosol calibration standard suitable for
2-59
-------
testing samplers in the field. Therefore, it is not possible at the present time to establish the
absolute accuracy of a PM monitoring technique. Intercomparison studies may be used to
establish the precision of identical monitors and the extent of agreement between different types
of monitors. Such studies are important for establishing the reliability of PM measurements.
Intercomparison studies have contributed greatly to our understanding of the problems in PM
measurement. Such studies will be discussed as they apply to specific measurement problems,
monitoring instruments, or analytical techniques.
Some measurement errors of concern in PM10 sampling, including those that arise due to
uncertainty tolerances in cut point, particle bounce and reentrainment; impactor surface
overloading; and losses to sampler internal surfaces, were discussed in detail in the 1996 PM
AQCD (U.S. Environmental Protection Agency, 1996a). Other measurement errors of concern
in PM25 sampling arise because of our inability to assess accuracy in an absolute sense due to a
lack of an atmospheric aerosol calibration standard, the inclusion in PM2 5 of a small amount of
coarse particles as discussed in Section 2.2.1.3, and problems associated with the definition
of PM2 5 as what remains on a filter after collection and equilibration, rather than as the mass of
particles as they exist in the air. Still, it is possible to measure PM indicators with high
precision. Detailed information on precision and quality assurance may be found on the EPA's
Technology Transfer Network website (U.S. Environmental Protection Agency, 2002a). See
discussion in Section 2.2.4.
Because of the difficulties associated with determining the accuracy of PM measurements,
EPA has sought to make FRM measurements equivalent by specifying operating conditions and,
in the case of PM2 5 samplers, by specifying details of the sampler design. Thus, both the PM10
and the PM25 standards are defined with consistency of measurement technique rather than with
the accuracy of the true mass concentration measurement in mind (McMurry, 2000). It is
acknowledged in the Federal Register (1997) that, "because the size and volatility of the particles
making up ambient particulate matter vary over a wide range and the mass concentration of
particles varies with particle size, it is difficult to define the accuracy of PM25 measurements in
an absolute sense. . . ." Thus, accuracy is defined as the degree of agreement between a
field PM2 5 sampler and a collocated PM2 5 reference method audit sampler (McMurry, 2000).
The Federal Reference Method for PM25 is discussed in Section 2.2.3.3. As mentioned earlier,
volatilization of organic compounds and ammonium nitrate during sampling or post-sampling
2-60
-------
handling can lead to significant underestimation of the fine particulate mass concentration in
some locations. Sources of error in the measurement of mass of PM25 suspended in the
atmosphere also arise because of adsorption or desorption of semivolatile vapors onto or from
collected PM, filter media, or other sampler surfaces; neutralization of acid or basic vapors on
either filter media or collected PM; and artifacts associated with particle-bound water.
During the past 25 years, there have been advancements in the generation and
classification of monodisperse aerosols as well as in the development of electron microscopy and
imaging analyses that have contributed to the advancement in aerosol calibration (Chen, 1993).
Still, one of the limitations in PM sampling and analysis remains the lack of primary calibration
standards for evaluating analytical methods and for intercomparing laboratories. Klouda et al.
(1996) examined the possibility of resuspending the National Institute of Standards and
Technology (NIST) Standard Reference Material 1649 (Urban Dust) in air for collection on up to
320 filters simultaneously using SRI's dust generation and collection system. However, the fine
component does not resuspend and the semivolatile component evaporates. Consequently, this
material is not a suitable standard for suspended PM. NIST is continuing work in this area with
EPA support.
Method validation was discussed in the 1996 PM AQCD (U.S. Environmental Protection
Agency, 1996a), and the usefulness of intercomparisons and "internal redundancy" was
emphasized. For example, a number of internal consistency checks are applied to the IMPROVE
network (Malm et al., 1994). These include mass balances, sulfur measurements by both
proton-induced X-ray emission (PIXE) and ion chromatography (1C), and comparison of organic
matter by combustion and by proton elastic scattering (PESA) of hydrogen. Mass balances
compare the gravimetrically determined mass with the mass calculated from the sum of the
major chemical components (i.e., crustal elements plus associated oxygen, organic carbon,
elemental carbon, sulfate, nitrate, ammonium, and hydrogen ions). Mass balances are useful
validation techniques; however, they do not check for, or account for, artifacts associated with
the absorption of gases during sampling, the loss of semivolatile material during sampling, or
errors in assumptions regarding unmeasured "associated species." The mass balance check may
appear reasonable even if such artifacts are present, because only the material collected on the
filter is included in the balance. Mass balance checks may also suffer from errors due to some
particle-bound water remaining in the PM even after equilibration and from the use of an
2-61
-------
arbitrary factor to account for the amount of oxygen and hydrogen atoms per carbon atom in the
organic carbon fractions (Andrews et al., 2000).
2.2.3 Measurement of Semivolatile Particulate Matter
Some atmospheric species, such as ammonium nitrate, water, and certain organic
compounds have vapor pressures such that, for typical ambient concentrations and temperatures,
they may exist in an equilibrium state with some material in the gas phase and some in the
condensed (particulate) phase. Such species are known as semivolatile materials (SVM) and
may partition to varying degrees between the gas and particle phases (Pankow, 1994a). Organic
compounds with vapor pressure < 10~6 Torr are nonvolatile, i.e., primarily in the particle phase.
Organic compounds with vapor pressure > 10~5 Torr will be primarily in the gas phase. Organic
compounds with vapor pressures on the order of 10~5 to 10~6 are semivolatile. Semivolatile
material, originally in the atmosphere in the particulate phase and collected on a filter, may
subsequently be lost from the filter. For example, SVM may evaporate during sampling due to a
reduction in its concentration in the atmosphere being sampled or due to the pressure drop across
the filter. Or, SVM may evaporate after sampling; during intentional equilibration at a low
relative humidity; or during transport, handling, and storage if exposed to an atmosphere in
which the vapor pressure of one or more semivolatile components is lower than in the
atmosphere sampled. Since water is not a pollutant, it is necessary to remove most of the
particle-bound water before weighing (Chow, 1995). However, the collection and measurement
of ammonium nitrate and semivolatile organic compounds in suspended atmospheric PM
represents a major analytical challenge (McMurry, 2000).
2.2.3.1 Particle-Bound Water
It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
compounds. However, for many measurements of suspended particle mass, it is desirable to
remove the particle-bound water before determining the mass. The mass of particle-bound water
is strongly dependent on the relative humidity and the particle composition. However, the
dependence on relative humidity is not linear, because there is significant hysteresis in the water
adsorption-desorption pathways (Seinfeld and Pandis, 1998). Water vapor cannot be controlled,
and particle-bound water is not included in the mass of PM subject to regulation and control.
2-62
-------
components, a measurement of PM mass including particle-bound water would depend more on
relative humidity than pollution. For all these reasons, it is usually desirable to remove most,
if not all, particle-bound water before weighing collected PM. On the other hand, in some
situations it may be important to know how much of the suspended particle mass or volume
results from particle-bound water. Figures 2-16a and 2-16b show the change in diameter of
sulfate particles as a function of relative humidity. Figure 2-16a also shows hysteresis resulting
from the difference between deliquescence and crystallization points.
Pilinis et al. (1989) calculated the water content of atmospheric PM above and below the
deliquescent point. They predicted that aerosol water content is strongly dependent on
composition and concluded from their calculations that liquid water could represent a significant
mass fraction of aerosol concentration at relative humidities above 60%. Since then, a few
researchers have attempted to measure the water content of atmospheric aerosols. Most
techniques have focused on tracking changes in the particle mass as the relative humidity is
changes. Only a few demonstrations have been carried out using atomospheric particles.
Of particular interest is the development of the Tandem Differential Mobility Analyzer (TDMA)
and its applications in investigations of the effects of relative humidity on particle growth.
Lee et al. (1997) examined the influence of relative humidity on the size of atmospheric
aerosols using a TDMA coupled with a scanning mobility particle sizer (SMPS). They reported
that the use of the TDMA/SMPS system allowed for the abrupt size changes of aerosols at the
deliquescence point to be observed precisely. They also reported that at relative humidities
between 81 and 89% the water content of ammonium sulfate aerosols (by mass) ranged from
47 to 66%.
Andrews and Larson (1993) investigated the interactions of single aerosol particles coated
with an organic film within a humid environment. Using an electrodynamic balance, they
conducted laboratory experiments in which sodium chloride and carbon black particles were
coated with individual organic surfactants (intended to simulate the surface-active, organic films
that many atmospheric aerosol particles may exhibit) and their water sorption curves were
examined. Their results showed that when ordinarily hydrophobic carbon black particles were
coated with an organic surfactant, they sorbed significant amounts of water (20 to 40% of the dry
mass of the particle).
2-63
-------
Liang and Chan (1997) developed a fast technique using the electrodynamic balance to
measure the water activity of atmospheric aerosols. In their technique, the mass of a levitated
particle is determined as the particle either evaporates or grows in response to a step change in
the relative humidity. Their technique was demonstrated using laboratory experiments with
NaCl, (NH4)2SO4, NaNO3, and (NH4)2SO4/NH4NO3 solutions. They concluded that one of the
advantages of their fast method is the ability to measure the water activity of aerosols containing
volatile solutes such as ammonium chloride and some organics.
Mclnnes et al. (1996) measured aerosol mass concentration, ionic composition, and the
associated water mass of marine aerosol over the remote Pacific Ocean. The mass of PBW was
determined by taking the difference between the mass obtained at 48% RH and at 19% RH,
assuming the aerosol particles were dry at 19% RH. Based on a comparison of the remote
Pacific aerosol to an aerosol collected at a site at the marine/continental interface of the
Washington coast, the amount of water associated with the aerosol was observed to be a function
of the ammonium to sulfate ratio. They found that the amount of water associated with the
submicrometer aerosol comprised 29% of the total aerosol mass collected at 47% RH and 9% of
the total mass at 35% RH.
Ohta et al. (1998) characterized the chemical composition of atmospheric fine particles
(50% cut point of 2 jim) in Sapporo, Japan and, as part of their measurements, determined the
water content using the Karl Fischer method (Meyer and Boyd, 1959). After exposing a Teflon
filter, a portion of the filter was equilibrated at 30% RH for 24 h. Then the filter piece was
placed in a water evaporator and heated at 150 °C, vaporizing the particle-bound water. The
evolved vapor was analyzed for water in an aqua-counter where it was titrated coulometrically in
Karl Fischer reagent solution (containing iodine, sulfur, and methanol). The accuracy of the
aqua-counter is ± 1 mg. Using this technique, they determined that the water content of the
particles ranged from 0.4 to 3.2% of the total particulate mass (at RH < 30%). This represents a
smaller portion of water compared to their previous reported values (Ohta and Okita, 1990),
which were determined by calculation at 50% RH.
Speer et al. (1997) developed an aerosol liquid water content analyzer (LWCA) in which
aerosol samples are collected on Teflon filters and then placed in a closed chamber in which
the relative humidity is closely controlled. The aerosol mass is monitored using a beta-gauge,
first as the relative humidity is increased from low RH to high RH, and then as the RH is
2-64
-------
decreased again. They demonstrated the LWCA on a laboratory-generated aerosol and on an
ambient PM2 5 sample collected in Research Triangle Park, NC. The ambient aerosol sample
was also analyzed for chemical constituents. It is interesting to note that, although their
laboratory-generated (NH4)2SO4 aerosol demonstrated a sharp deliquescent point, their
atmospheric aerosol, which was essentially (NH4)2SO4, did not show a sharp deliquescent point.
Hygroscopic properties of aerosols have been studied from the viewpoint of their ability to
act as condensation nuclei. The hygroscopic properties of fresh and aged carbon and diesel soot
particles were examined by Weingartner et al. (1997) who found that fresh, submicron-size
particles tended to shrink with increasing relative humidity because of a structural change.
Lammel and Novakov (1995) found, in laboratory studies, that the hygroscopicity of soot
particles could be increased by chemical modification and that the cloud condensation nucleation
characteristics of diesel soot were similar to those of wood smoke aerosol.
The results of several of the above studies in which aerosol water content was determined
as a function of relative humidity are summarized in Figure 2-19, which includes the results of
Lee et al. (1997), Mclnnes et al. (1996), and Ohta et al. (1998). Relative humidity ranged from
9%, at which the aerosol water content was assumed to be zero (Mclnnes et al., 1996), to 89%, at
which the aerosol water content was determined to be 66% by mass (Lee et al., 1997). Koutrakis
et al. (1989) and Koutrakis and Kelly (1993) also have reported field measurements of the
equilibrium size of atmospheric sulfate particles as a function of relative humidity and acidity.
100
S 90 -
•D
| 80-
m
i 70
u
t
CO
Q.
2 50 -
01
Q 40 -
60 -
30-
I 2°H
1
Mclnnes et al. (1996)
Leeetal. (1997)
Ohtaetal. (1998)
10 20 30 40 50 60 70
Relative Humidity, RH (%)
80
90
100
Figure 2-19. Aerosol water content expressed as a mass percentage, as a function of
relative humidity.
2-65
-------
The effects of relative humidity on particle growth were also examined in several studies.
Fang et al. (1991) investigated the effects of flow-induced relative humidity changes on particle
cut sizes for aqueous sulfuric acid particles in a multi-nozzle microorifice uniform deposit
impactor (MOUDI). Laboratory experiments were conducted in which polydisperse sulfuric
acid aerosols were generated and the RH was adjusted. The aerosols were analyzed by a
differential mobility analyzer. Fang et al. (1991) observed that for inlet RH less than 80%, the
cut sizes for the sulfuric acid aerosols were within 5% of that for nonhygroscopic particles
except at the stage for which the cut size was 0.047 |im, (10.7% larger than the nonhygroscopic
particle cut size). They concluded that flow-induced RH changes would have only a modest
effect on MOUDI cut sizes at RH < 80%.
Hitzenberger et al. (1997) collected atmospheric aerosol in the size range of 0.06 to 15 jim
in Vienna, Austria using a nine-stage cascade impactor and measured the humidity-dependent
water uptake when the individual impaction foils were exposed to high RH. They observed
particle growth with varying growth patterns. Calculated extinction coefficients and single
scattering albedo increased with humidity.
The hygroscopic properties, along with the mixing characteristics, of the submicrometer
particles sampled in Los Angeles, CA during the summer of 1987 SCAQS study and at the
Grand Canyon, AZ during the 1990 Navajo Generating Station Visibility Study were reported by
Zhang et al. (1993). They used a tandem differential mobility analyzer (TDMA; McMurry and
Stolzenburg, 1989) to measure the hygroscopic properties for particles in the 0.05 to 0.5 jim
range. In their experimental technique, monodisperse particles of a known size are selected from
the atmospheric aerosol with the first analyzer. Then, the relative humidity of the monodisperse
aerosol is adjusted, and the new particle size distribution is measured with the second analyzer.
At both sites, they observed that monodisperse particles could be classified according to "more"
hygroscopic and "less" hygroscopic. Aerosol behavior observed at the two sites differed
markedly. Within the experimental uncertainty (± 2%) the "less" hygroscopic particles sampled
in Los Angeles did not grow when the RH was increased to 90%; whereas at the Grand Canyon,
the growth of the "less" hygroscopic particles varied from day to day, but ranged from near 0 to
40% when the RH was increased to 90%. The growth of the "more" hygroscopic particles in
Los Angeles was dependent on particle size (15% at 0.05 jim to 60% at 0.5 jim); whereas at the
Grand Canyon, the "more" hygroscopic particles grew by about 50% with the growth not
2-66
-------
varying significantly with particle size. After comparing the TDMA data to impactor data,
Zhang et al. (1993) surmised that the more hygroscopic particles contained more sulfates and
nitrates, while the less hygroscopic particles contained more carbon and crustal components.
Although most of the work to date on the hygroscopic properties of atmospheric aerosols
has focused on the inorganic fraction, the determination of the contribution of particle-bound
water to atmospheric particulate mass is greatly complicated by the presence of organics. The
effect of RH on the adsorption of semivolatile organic compounds is discussed elsewhere in this
chapter. Saxena et al. (1995) observed that particulate organic compounds can also affect the
hygroscopic behavior of atmospheric particles. They idealized the organic component of aerosol
as containing a hydrophobic fraction (e.g., high-molecular weight alkanes, alkanoic acids,
alkenoic acids, aldehydes, and ketones) and a hydrophilic fraction (e.g., lower molecular weight
carboxylic acids, dicarboxylic acids, alcohols, aldehydes, etc.) that would be likely to absorb
water. They then analyzed data from a TDMA in conjunction with particle composition
observations from an urban site (Claremont, CA) and from a nonurban site (Grand Canyon) to
test the hypothesis that, by adding particulate organics to an inorganic aerosol, the amount of
water absorbed would be affected, and the effect could be positive or negative, depending on the
nature of the organics added. They further presumed that the particulate organic matter in
nonurban areas would be predominantly secondary, and thus hydrophilic, compared to the urban
aerosol that was presumed to be derived from primary emissions and thus hydrophobic in nature.
Their observations were consistent with their hypothesis, in that at the Grand Canyon, the
presence of organics tended to increase the water uptake by aerosols; whereas at the Los Angeles
site, the presence of organics tended to decrease water uptake.
Peng and Chan (2001) also recently studied the hygroscopic properties of nine water
soluble organic salts of atmospheric interest using an electrodynamic balance operated at 25 °C.
Salts studied included sodium formate, sodium acetate, sodium succinate, sodium pyruvate,
sodium methanesulfonate, sodium oxalate, ammonium oxalate, sodium malonate, and sodium
maleate. They observed that hygroscopic organic salts have a growth factor of 1.76 to 2.18 from
RH = 10 to 90%, which is similar to that of typical hygroscopic inorganic salts such as NaCl
and (NH4)2SO4.
Nonequilibrium issues may be important for the TDMA, as well as for other methods of
measuring water content. Although the approach to equilibrium as the RH is increased is
2-67
-------
expected to be rapid for pure salts, it may be much slower for aerosols containing a complex mix
of components (Saxena et al., 1995). For example, if an aerosol contains an organic film or
coating, that film may impede the transport of water across the particle surface, thus increasing
the time required for equilibrium (Saxena et al., 1995). Insufficient time to achieve equilibrium
in the TDMA could result in underestimation of the water content.
2.2.3.2 Nitrate and Organic Species
Particulate Nitrates
It is now well known that volatilization losses of particulate nitrates occur during sampling
on Teflon filters (e.g., Zhang and McMurry [1992]; see also Hering and Cass [1999] and Babich
et al. [2000]). The effect on the accuracy of atmospheric particulate measurements from these
volatilization losses is more significant for PM2 5 than for PM10. The FRM for PM2 5 will likely
suffer a loss of nitrates similar to that experienced with other simple filter collection systems.
Sampling artifacts resulting from the loss of particulate nitrates represents a significant problem
in areas such as southern California that experience high amounts of nitrates. Hering and Cass
(1999) reported on errors in PM25 mass measurements due to the volatilization of particulate
nitrate (Figure 2-20). They examined data from two field measurement campaigns that were
conducted in southern California: (1) the Southern California Air Quality Study (SCAQS)
(Lawson, 1990) and (2) the 1986 CalTech study (Solomon et al., 1992). In both these studies,
side-by-side sampling of PM25 was conducted. One sampler collected particles directly onto a
Teflon filter. The second sampler consisted of a denuder to remove gaseous nitric acid followed
by a nylon filter that absorbed the HNO3 as it evaporated from ammonium nitrate. In both
studies, the denuder consisted of MgO-coated glass tubes (Appel et al., 1981). Fine particulate
nitrate collected on the Teflon filter was compared to fine particulate nitrate collected on the
denuded nylon filter. In both studies, the PM2 5 mass lost because of ammonium nitrate
volatilization represented a significant fraction of the total PM25 mass. The fraction of mass lost
was higher during summer than during fall (17% versus 9% during the SCAQS study, and 21%
versus 13% during the CalTech study; Figure 2-20). In regard to percentage loss of nitrate, as
opposed to percentage loss of mass discussed above, Hering and Cass (1999) found that the
amount of nitrate remaining on the Teflon filter samples was on average 28% lower than that on
the denuded nylon filters.
2-68
-------
(U
*j
re
SCAQ3 Data Set
Summer Measurements
Fall Measurements
50 100 150 200 250
PM2 5 Gravimetric Mass (jjg/m3)
o April - September
» October - March
PM2 5 Gravimetric Mass (|jg/m3)
Figure 2-20. 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.
Source: Hering and Cass (1999).
Hering and Cass (1999) also analyzed these data by extending the evaporative model
developed by Zhang and McMurry (1987). The extended model used by Hering and Cass (1999)
takes into account the dissociation of collected particulate ammonium nitrate on Teflon filters
into nitric acid and ammonia via three mechanisms: (1) the scrubbing of nitric acid and
ammonia in the sampler inlet (John et al. [1988] showed that clean PM10 inlet surfaces serve as
an effective denuder for nitric acid); (2) the heating of the filter substrate above ambient
temperature by sampling; and (3) the pressure drop across the Teflon filter. For the sampling
systems modeled, the flow-induced pressure drop was measured to be less than 0.02 atm, and the
corresponding change in vapor pressure was 2%, so losses driven by pressure drop were not
considered to be significant in this work. Losses from Teflon filters were found to be higher
during the summer compared to the winter, higher during the day compared to night, and
reasonably consistent with modeled predictions.
Finally, during the SCAQS (Lawson, 1990) study, particulate samples also were collected
using a Berner impactor and greased Tedlar substrates in size ranges from 0.05 to 10 jim in
aerodynamic diameter. The Berner impactor PM2 5 nitrate values were much closer to those
2-69
-------
from the denuded nylon filter than those from the Teflon filter, the impactor nitrate values being
-2% lower than the nylon filter nitrate for the fall measurements and -7% lower for the summer
measurements. When the impactor collection was compared to the Teflon filter collection for a
nonvolatile species (sulfate), the results were in agreement. Chang et al. (2000) discuss reasons
for reduced loss of nitrate from impactors.
It should be noted that filters or collection surfaces were removed immediately after
sampling and placed into vials containing a basic extraction solution during these
intercomparison studies to preclude losses that might occur during their handling, storage, and
equilibration. The loss of nitrate observed from Teflon filters and impaction surfaces in this
study, therefore, is a lower limit compared to losses that might occur during the normal
processes involved in equilibration and weighing of filters and impaction surfaces. Brook and
Dann (1999) observed much higher nitrate losses during a study in which they measured
particulate nitrate in Windsor and Hamilton, Ontario, Canada, by three techniques: (1) a single
Teflon filter in a dichotomous sampler, (2) the Teflon filter in an annular denuder system (ADS),
and (3) total nitrate including both the Teflon filter and the nylon back-up filter from the ADS.
The Teflon filter from the dichotomous sampler averaged only 13% of the total nitrate, whereas
the Teflon filter from the ADS averaged 46% of the total nitrate. The authors concluded that
considerable nitrate was lost from the dichotomous sampler filters during handling, which
included weighing and X-ray fluorescence (XRF) measurement in a vacuum.
Kim et al. (1999) also examined nitrate sampling artifacts by comparing denuded and
undenuded quartz and nylon filters, during the PM10 Technical Enhancement Program (PTEP) in
the South Coast Air Basin of California. They observed negative nitrate artifacts (losses) for
most measurements; however, for a significant number of measurements, they observed positive
nitrate artifacts. Kim et al. (1999) pointed out that random measurement errors make it difficult
to measure true amounts of nitrate loss.
Diffusion denuder samplers, developed primarily to measure particle strong acidity
(Koutrakis et al., 1988a,b, 1992), also can be used to study nitrate volatilization. Such
techniques were used to measure loss of particulate nitrate from Teflon filters in seven U.S.
cities (Babich et al., 2000). Measurements were made with two versions of the Harvard-EPA
Annular Denuder System (HEADS). Nitric acid vapor was removed by a Na2CO3-coated
denuder. Particulate nitrate was the sum of nonvolatile nitrate collected on a Teflon filter and
2-70
-------
volatized nitrate collected on a Na2CO3-coated filter downstream of the Teflon filter (full
HEADS) or on a Nylon filter downstream of the Teflon filter (Nylon HEADS). It was found that
the full HEADS (using a Na2CO3 filter) consistently underestimated the total particulate nitrate
by approximately 20% compared to the nylon HEADS. Nonvolatilized nitrate was also
measured on a Teflon filter from collocated Harvard Impactors (His). The PM2 5 HI sampler,
like the PM25 FRM, uses impactors with a 50% cut at 2.5 jim. The HI uses a 37-mm filter and a
flow rate of 10 L/min (0.465 L/min/cm2), while the FRM uses a 47-mm filter and a flow rate of
16.7 L/min (0.481 L/min/cm2). Therefore, the flow rate and pressure drop across the filters are
similar and the loss of nitrate should be similar for both types of samples. Babich et al. (2000)
found significant nitrate losses in Riverside, CA; Philadelphia, PA; and Boston, MA but not in
Bakersfield, CA; Chicago, IL; Dallas, TX; or Phoenix, AZ where measurements were made only
during the winter. Tsai and Huang (1995) used a diffusion denuder to study the positive and
negative artifacts on glass and quartz filters. They found positive artifacts attributed to SO2 and
HNO3 reaction with basic sites on glass fibers and basic particles and negative artifacts attributed
to loss of HNO3 and HC1 due to volatilization of NH4NO3 and NH4C1 and reaction of these
species with acid sulfates.
Eatough et al. (1999a) developed a high-volume diffusion denuder system that combined
diffusion denuder and particle concentrator techniques (see Section 2.2.3.2). In this system, the
particle concentrator reduces the flow through the denuder so that the denuder can be operated
for weeks without a loss of collection efficiency, making the sampler suitable for routine field
sampling. The system was evaluated for the collection of fine particulate sulfate and nitrate in
Riverside, CA (Eatough et al., 1999b). Concentrations of PM25 nitrate obtained from the PC-
BOSS agreed with those obtained using the Harvard-EPA Annular Denuder Sampler, HEADS
(Koutrakis et al., 1988b).
Particulate Organic Compounds
Many semivolatile organic compounds (SVOCs) are of interest because of their possible
health effects. Semivolatile organic compounds include products of incomplete combustion
such as polycyclic aromatic hydrocarbons (PAHs) and polycyclic organic matter, which has been
identified as a hazardous air pollutant. Polycyclic aromatic hydrocarbons also have been
suggested as alternative particulate tracers for automobile emissions, because lead is no longer a
2-71
-------
good tracer for automobile emissions sources due to the phase-out of organo-lead additives in
gasoline (Venkataraman et al., 1994). PAHs also are emitted during biomass burning, i.e.,
burning of cereal crop residues and wood fuels (Jenkins et al., 1996; Roberts and Corkill, 1998).
A number of studies have used absorbing material following quartz filters to determine
phase equilibria of specific organic compounds (Liang et al., 1997; Gundel et al., 1995; Kamens
et al., 1995). Much work has also gone into the development of a theory to help understand the
phase relationships (Yamasaki et al., 1982; Rounds and Pankow, 1990; Pankow, 1987, 1994a,b;
Pankow et al., 1993; Rounds et al., 1993; Odum et al., 1994). The development of a theory
describing phase equilibria of SVOC and the gas/particle partitioning of SVOC on inorganic,
organic, and ambient smog aerosols is ongoing (Liang et al., 1997; Jang et al., 1997; Strommen
and Kamens, 1997; Jang and Kamens, 1998, 1999, 2001; Leach et al., 1999; Kamens et al.,
1999; Kamens and Jaoui, 2001; Chandramouli and Kamens, 2001; Chandramouli et al., 2003).
The mass of OC and EC is usually determined by the analysis of PM collected on a quartz
filter. However, quartz fiber filters have a large specific surface area on which adsorption of
gases can also occur. Possible artifacts associated with adsorption of organic gases onto quartz
filters have been examined in experiments in which two quartz-fiber filters were deployed in
series. The second quartz filter may indicate gaseous volatile organic compounds (VOCs)
adsorbed on both filters (positive artifact), SVOCs evaporated from particles on the first filter
and subsequently adsorbed on the second filter (negative artifact) or a combination of both
effects. Unless the individual compounds are identified, the investigator does not know what to
do with the loading value on the second filter (i.e., to add or subtract from the first filter loading
value). Moreover, even if the individual compounds were identified on the back-up filter, the
decision concerning adding or subtracting the back-up filter loading would not be
straightforward.
The positive quartz filter artifact has been discussed by Gundel et al. (1995) and Turpin
et al. (2000). It is also possible that some SVOCs may desorb from the filter resulting in a
negative artifact (Eatough et al., 1993, 1995; Tang et al., 1994; Gundel et al., 1995; Cui et al.,
1998; Pang et al., 2001; Finn et al., 2001). Semivolatile organic compounds can similarly be lost
from Teflon filters because of volatilization, causing the PM25 mass to be significantly
underestimated (negative artifact). Like particulate nitrates, the FRM for PM2 5 will suffer loss
of SVOC similar to the losses experienced with other simple filter collection systems. Most
2-72
-------
studies that have focused on the positive and negative sampling artifacts associated with SVOCs
have used either diffusion denuder technology (Eatough et al., 1995; Mader et al., 2001) or
placed an adsorbent media, such as a back-up quartz filter or a polyurethane foam adsorbent,
behind the main filter (Wallace and Kites, 1995). Further information on denuder techniques is
given in Appendix 2A.
Using their multichannel diffusion denuder sampling system (BOSS), Eatough et al. (1995)
reported that, for samples collected at the South Coast Air Quality Management District
sampling site at Azusa, CA, changes in the phase distribution of SVOCs could result in an
average loss of 35% of the particulate organic material. Cui et al. (1998) found that losses of
SVOCs from particles in the Los Angeles Basin during the summer were greater during the night
(average = 62%) than during the day (average = 42%).
The percent SVOC lost from the front filter in a filter-denuder system may be greater than
that lost in a filter-only system such as the FRM. In a filter-denuder system, the gas-phase
component of the SVOCs is removed. The absence of the gas-phase causes the gas-particle
equilibrium to shift, so the SVOCs collected on the filter may evaporate more rapidly in a filter-
denuder system than in a filter-only collection system. To determine the fraction of SVOCs lost
from a Teflon filter in a filter-only system, it is necessary to compare the amount measured by a
nondenuder system with that measured by a denuder system. (Similar considerations apply to
the collection of ammonium nitrate. However, in the case of ammonium nitrate, the total
particulate nitrate is easily obtainable from the sum of nitrate on the Teflon front filter and the
back-up Nylon filter. In the case of SVOCs, the existence of the positive artifact [since the
organic denuder is less efficient than the nitric acid denuder], makes it much more difficult to
determine the total OC.) At present, little information is available on the volatilization losses of
SVOCs. However, in one study (Pang et al., 2001), the total mass on denuded and undenuded
filters were compared and found to be identical within error limits (R2 = 0.816, slope = 0.961 ±
0.027 for total mass compared to R2 = 0.940, slope = 0.986 ± 0.020 for sulfate). Pang et al.
(2001) interpreted this result as suggesting that the major cause of loss of SVOCs is the pressure
drop across the filter.
Positive artifacts may occur during sample collection because of the adsorption of gases
onto the filter materials (e.g., Gundel et al., 1995). Using a quartz filter behind a Teflon filter,
Kim et al. (2001a) estimated that on an annual average basis 30% of the PM25 organic carbon
2-73
-------
concentration resulted from positive artifacts. Quartz fiber filters have a larger positive artifact
than Teflon filters because of a greater adsorption of organic vapor onto quartz fibers (Turpin
et al., 1994; Chow et al., 1994a,b, 1996; Eatough et al., 1996; Finn et al., 2001; Kirchstetter
etal.,2001).
Kirchstetter et al. (2001) reported that adsorptive properties of quartz fiber filters vary with
lot number; therefore, front and back-up filters should be taken from the same lot. The literature
suggests that a Teflon filter followed by a quartz back-up filter provides a better estimate of the
adsorption of gases on a quartz fiber front filter than does a quartz filter followed by a quartz
backup, and that the difference between these two adsorption estimates can be substantial for
short durations (Novakov et al., 1997; Kirchstetter et al., 2001; Turpin et al., 2000). The
typically lower organic carbon loadings on concurrently collected quartz followed by quartz
filters relative to Teflon followed by quartz filters are believed to occur because adsorption on
the quartz front filter acts to reduce the gas-phase concentration downstream until adsorption
equilibrium has been achieved in the vicinity of the front quartz filter surface. Because Teflon
filters have little affinity for organic vapors, this equilibrium occurs almost instantaneously for
Teflon filters; and the Teflon-quartz back-up filter is exposed to the ambient concentration of
organic vapors from the beginning of the sampling period. It might be expected that the quantity
of organic vapor adsorbed on quartz filters would depend on the organic composition and would
vary by season and location. However, it is also possible that the quartz possesses a limited
number of adsorption sites that are rapidly occupied, so that the quantity of OC on the back up
filter would be relatively constant and depend on the pretreatment of the quartz.
Combined Measurement of Semivolatile Nitrate and Organic Carbon and Nonvolatile
(Organic Carbon, Elemental Carbon, Nitrate, and Sulfate)
Fine particles in urban atmospheres contain substantial quantities of semivolatile material
(e.g., NH4NO3 and SVOCs) that are lost from particles during collection on a filter. Several
diffusion denuder samplers have been developed for the determination of both NO3 and organic
semivolatile fine particulate components as well as nonvolatile nitrate, organic compounds, and
nonvolatile sulfate (Pang et al., 2001; Eatough et al., 1993). The combination of technology
used in the BOSS diffusion denuder sampler and the Harvard particle concentrator has resulted
in the Particle Concentrator-Brigham Young University Organic Sampling System (PC-BOSS)
for the 24-h integrated collection of PM2 5, including NH4NO3 and SVOC. Modifications of the
2-74
-------
BOSS sampler allow for the determination of these same species on a time scale from a few
hours to weekly (Ding et al., 2002; Eatough et al., 1999a,b, 2001). Episodic studies have been
conducted in Riverside, CA, and Bakersfield, CA, (Obeidi et al., 2002) and Provo, UT (Obeidi
and Eatough, 2002). The average concentration of semivolatile and nonvolatile components in
the three cities, during the study periods, are shown in Figure 2-21.
2.2.3.3 Continuous Measurement of Semivolatile and Nonvolatile Mass
Techniques for the continuous measurement of PM mass are needed both to provide
real-time information on pollution levels (Long et al., 2002) and to reduce the costs involved in
visiting sites to change filters and in the equilibration and weighing of filters. Two techniques
are currently in use. The TEOM is normally operated at 50 °C in order to remove particle-bound
water. However, at 50 °C most semivolatile material is also evaporated. Therefore, the TEOM,
operated at 50 °C, may be considered to measure the mass of nonvolatile PM. Since the filter in
the beta gauge mass monitor is changed more frequently than the filter in the TEOM, the beta
gauge is less sensitive to changes in mass caused by changes in relative humidity. However,
most beta gauge monitors heat the inlet, but do not otherwise control the temperature at the filter.
This heating causes evaporation of a substantial fraction of the particle-bound water and an
unknown fraction of the semivolatile PM. Thus, the beta gauge may be considered to measure
the nonvolatile PM plus a small fraction of the particle-bound water and an unknown fraction of
the semivolatile PM. Three new techniques have been developed to address the issue of lost
semivolatile PM mass: the real-time ambient mass sample (RAMS), the differential TEOM, and
the continuous ambient mass monitor (CAMM).
Real-Time Total Ambient Mass Sampler (RAMS)
The RAMS, a monitor based on diffusion denuder and TEOM monitor technology, has
been developed, validated, and field tested for the real-time determination of total fine PM mass,
including semivolatile PM (Eatough et al., 1999a; Obeidi and Eatough, 2002; Obedi et al., 2002;
Pang et al., 2001). The RAMS measures the total mass of collected particles including
semivolatile species with a TEOM monitor using a "sandwich filter." The sandwich contains a
Teflon coated particle collection filter followed by a charcoal-impregnated filter (GIF) to collect
any semivolatile species lost from the particles during sampling. Because the instrument
2-75
-------
Riverside, CA; September 1997
Semivolatile^
Components/^E
Lost
Ammonium
Nitrate
2.3
Ammonium
Nitrate E|emental Ammonium
Carbon Sulfate
3.1 1.2
2.6
Bakersfield, CA; March"! 998
Semivolatile
Components,.
Lost
Ammoniui
Nitrate tj
0.8
imental
Carbon
1.7
Provo, UT; September 1998
Semivolatile^
Components^^
Nonvolatile
Components
Figure 2-21. Average concentration (jig/m3) of nonvolatile and semivolatile PM
components in three cities.
Source: Obeidi and Eatough (2002); Obeidi et al. (2002).
2-76
-------
measures total mass collected by the sandwich filter, all gas phase compounds that can be
adsorbed by a GIF must be removed from the sampling stream prior to the TEOM monitor.
Laboratory and field validation data indicate that the precision of fine PM mass determination is
better than 10%. The RAMS uses a Nafion dryer to remove particle-bound water from the
suspended particles and a particle concentrator to reduce the amount of gas phase organics that
must be removed by the denuder. Examples of PM2 5 mass as measured by the RAMS, the
TEOM, the PC-BOSS, and an FRM sampler are shown in Figure 2-22. This figure also shows
that the TEOM recorded the PM2 5 mass as being negative for the hours of 16 to 19. This likely
results from the loss of volatile materials from the heated filter.
100
Riverside, CA
13 15 17 19 21 23 0 1
-20
Time of Day
PC-BOSS (Nonvolatile Material) [_] PC-BOSS (Lost From Particles)
- TEOM —B— RAMS FRM PM,
at 35 °C
at 35 °C
"2.5
24 h average
Figure 2-22. 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. (1999).
2-77
-------
Differential TEOM
Patashnick et al. (2001) developed a differential TEOM system that is based on a pair of
TEOM sensors, each of which is preceded by its own electrostatic precipitator (ESP) and
downstream from a common size selective inlet. By alternately switching the ESPs on and off
and out of phase with each other, the two sensors measure "effective mass" that includes both
the nonvolatile component and the volatile component sampled by the TEOM, less the volatile
component that vaporized during the sampling interval. On the sensor side with the ESP turned
on, there is no particle collection on that filter so that only volatilization of previously collected
particles continues. This allows a correction for the effective mass as measured by the first
sensor by subtracting out the volatilization artifact and leaving the nonvolatile and volatile
components of the PM. This system has yet to be well characterized for other biases or
interferences such as reactions on the filters, particle collection efficiency of the ESPs, and
particle and semivolatile material losses.
Continuous Ambient Mass Monitor (CAMM)
Koutrakis and colleagues (Koutrakis et al., 1996; Wang, 1997) developed CAMM,
a technique for the continuous measurement of ambient particulate matter mass concentration
based on the measurement of pressure drop increase with particle loading across a membrane
filter. Recently, Sioutas et al. (1999) examined the increase in pressure drop with increasing
particle loading on Nucleopore filters. They tested filters with two pore diameters (2 and 5 jim)
and filter face velocities ranging from 4 to 52 cm s"1 and examined the effects of relative
humidity in the range of 10 to 50%. They found that, for hygroscopic ammonium sulfate
particles, the change in pressure drop per unit time and concentration was a strong function of
relative humidity, decreasing with increasing relative humidity. These results suggest that
particulate concentration measurements made with the pressure drop method (Koutrakis et al.,
1996) may be subject to additional uncertainties if used in an environment where the ambient
RH is quite variable and the RH where the particles are measured cannot be controlled
accurately. The current version of the CAMM (Wang, 1997) uses a particle concentrator and a
Nafion dryer and frequently moves the filter tape to avoid artifacts due to evaporation of
semivolatile components from the active portion of the filter tape which would occur if the
atmospheric concentration of the semivolatile components decreased.
2-78
-------
The CAMM was recently operated alongside a gravimetric PM method (the Harvard
Impactor, or HI) in seven U.S. cities selected for their distinctly different ambient particulate
compositions and densities. The correlation between the two methods was high, with an
overall r2 of 0.90 and average CAMM/HI ratio of 1.07 (Babich et al., 2000).
2.2.4 U.S. Environmental Protection Agency Monitoring Methods
2.2.4.1 The Federal Reference Methods for Measurement of Equilibrated Mass for PM10,
PM2 5, and PM10 2 5
In 1997, the EPA promulgated new standards for PM25 to address fine-fraction thoracic
particles and retained with minor revisions, the 1987 PM10 standards to continue to address
coarse-fraction thoracic particles (Federal Register, 1997). In partial response to numerous
challenges to these standards, the U.S. Court of Appeals for the District of Columbia Circuit in
American Trucking Association v. EPA, 175 F. 3d 1027 (U.S. Court of Appeals, D.C. Cir. 1999)
found "ample support" for regulating coarse-fraction particles. However, the court revoked the
revised PM10 standards (leaving in effect the 1987 PM10 standards) on the basis of PM10 being a
"poorly matched indicator for coarse particulate pollution" because PM10 includes fine particles.
Consistent with this specific aspect of the Court's ruling, which the EPA did not appeal, the EPA
is now considering use of PM10_2 5 as the indicator for coarse-fraction thoracic particles, in
conjunction with PM2 5 standards that address fine-fraction thoracic particles. Thus, the EPA is
now developing a Federal Reference Method for the measurement of PM10_2 5.
2.2.4.1.1 PMW
The FRM for measuring PM10 is specified in the Code of Federal Regulations (2001a,b).
The PM10 FRM defines performance specifications for samplers in which particles are inertially
separated with a penetration efficiency of 50% at an aerodynamic diameter (Da) of 10 ± 0.5 jim.
The collection efficiency increases to «100% for smaller particles and drops to «0% for larger
particles. Particles are collected on filters and mass concentrations are determined
gravimetrically. Instrument manufacturers are required to demonstrate through field tests a
measurement precision for 24-h samples of ± 5 |ig/m3 for PM10 concentrations below 80 |ig/m3
and 7% above this value. A number of samplers have been designated as PM10 reference
samplers. The TEOM and several beta gauge samplers with 1-h time resolution have been
designated as automated equivalent methods (U.S. Environmental Protection Agency, 2001).
2-79
-------
2.2.4.1.2 PM25
In contrast to the performance-based FRM standard for PM10, the FRM for PM2 5 (Code of
Federal Regulations, 2001a) specifies certain details of the sampler design, as well as of sample
handling and analysis, whereas other aspects have performance specifications (Noble et al.,
2001). The PM2 5 FRM sampler consists of a PM10 inlet/impactor, a PM2 5 impactor with an oil-
soaked impaction substrate to remove particles larger than 2.5 jim Da, and a 47-mm PTFE filter
with a particle collection efficiency greater than 99.7%. The sample duration is 24 h, during
which time the sample temperature is not to exceed ambient temperatures by more than 5 °C.
A schematic diagram of the PM25 FRM sample collection system is shown in Figure 2-23. After
collection, samples are equilibrated for 24 h at temperatures in the range of 20 to 23 °C (± 2 °C)
and at relative humidities in the range of 30 to 40% (± 5%). The equilibration tends to reduce
particle-bound water and stabilizes the filter plus sample weight. Filters are weighed before and
after sampling under the same temperature and relative humidity conditions. For sampling
conducted at ambient relative humidity < 30%, mass measurements at relative humidities down
to 20% are permissible (Code of Federal Regulations, 200la).
The PM10 inlet specified for the PM25 FRM is modified from a previous low flow
rate PM10 inlet that was acceptable in both EPA-designated reference and equivalent PM10
methods. The modification corrects a flaw that was reported for the previous sampler in that
under some meteorological conditions the inlet may allow precipitation to penetrate the inlet.
The modification includes a larger drain hole, a one-piece top plate, and louvers. Tolocka et al.
(200 la) evaluated the performance of this modified inlet in a series of wind tunnel experiments.
The modified inlet was found to provide a size cut comparable to the original inlet, for
both PM2 5 and PM10 sampling. Because the modification did not change the characteristics of
the size cut, the modified inlet may be substituted for the original inlet as part of a reference or
equivalent method for PM10 and PM2 5 (Tolocka et al., 2001a).
WINS Impactor. Design and calibration of the EPA PM2 5 Well Impactor Ninety-Six
(WINS) is given by Peters et al. (2001a). The WINS impactor was designed to be deployed
downstream of the Graseby-Andersen 246B PM10 inlet as part of a sampler operating at a
flow rate of 16.7 L/m. A schematic diagram of the WINS is shown in Figure 2-24. The PM25
inlet consists of a single jet directed toward a round hole with a jet exit impaction surface made
2-80
-------
Ambient
Aerosol
Sampling
Inlet
PM10
Fractionator
Down tube
WINS
PM2.5
Fractionator
Figure 2-23. Schematic diagram of the sample collection portion of the PM2 5
FRM sampler.
Source: Noble et al. (2001).
up of a 37-mm diameter glass fiber filter immersed in 1 ml of low volatility diffusion pump oil
(i.e., the well). Particles not having enough inertia to be removed by the impactor are captured
downstream on the sample collection filter. This design was selected to minimize impactor
overloading that would otherwise result in particle bounce. The oil wicks through the particulate
deposit on the impactor to provide a continuously wetted surface for impaction. The penetration
curve indicated a 50% cutpoint of 2.48 jim Da with a geometric standard deviation of 1.18
(Figure 2-25).
The WINS separator was evaluated for its loading characteristics (Vanderpool et al., 2001)
by monitoring the performance after repeated operation in an artificially generated, high
concentration, coarse-mode aerosol composed of Arizona Test Dust, as well as in the field in
Rubidoux, Phoenix, Philadelphia, Research Triangle Park, and Atlanta. In the wind tunnel
experiments, the WINS performance was found to be a monotonic function of loading.
2-81
-------
PM10 Aerosol from Inlet
Nozzle
Collection Cup
with Antispiil Ring
Impaction Surface:
Filter Immersed in 1 mL
Dow Corning 704
Diffusion Pump Oil
1
'
PM25 Aerosol to Sample Collection Filter
Figure 2-24. Schematic view of the final design of the WINS.
Source: Peters et al. (200la).
A -5% bias in the PM25 measurement resulted from a coarse particulate loading of
approximately 16 mg. This negative bias was due to a slight reduction in the separator cut point.
It was also found that the predictable results from the controlled laboratory experiments could
not be extrapolated to field settings and that the WINS performance was more sensitive to the
impactor loading in the field than it was in experiments with the single component aerosol.
Significant particle bounce was not observed in either the laboratory or the field experiments.
Vanderpool et al. (2001) concluded that their study supports the recommendation that the FRM
WINS wells should be replaced after every 5 days of 24-h operation (U.S. Environmental
Protection Agency, 1998).
A detailed sensitivity study of the WINS impactor was conducted (Vanderpool et al., 2001)
in which the effects on the impactor performance of a number of parameters were examined.
The results of this study are summarized in Table 2-4.
2-82
-------
100
1.0 2.0 3.0 4.0
Aerodynamic Diameter, Da
Figure 2-25. Evaluation of the final version of the WINS.
Source: Peters et al. (200la).
The regulations also allow for Class I, II, and III equivalent methods for PM2 5 (Code of
Federal Regulations, 200 Ic). Class I equivalent methods use samplers with relatively small
deviations from the sampler described in the FRM. Class II equivalent methods include "all
other PM25 methods that are based upon 24-h integrated filter samplers that are subjected to
subsequent moisture equilibration and gravimetric mass analysis." Class III equivalent methods
include non-filter-based methods such as beta attenuation, harmonic oscillating elements, or
nephelometry (McMurry, 2000). As of July 2001, 11 PM2 5 samplers (listed in Table 2-5) had
been tested, leading to the conclusion that the PM10 sampling systems can be designed to
produce concentration measurements that are precise to ± 10%. Cut point tolerances are not
expected to affect the mass concentration for PM2 5 as much as for PM10, because the 2.5 jim Da
cut point generally occurs near the minimum in a mass distribution (e.g., Figure 2-6).
2-83
-------
TABLE 2-4. 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
4%
0.05 micrometers
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
0°C
-20°C
-35°C
± 0.02 micrometers
No effect
-0.07 micrometers
Negligible
2.40 micrometers
No effect
No effect
Need to change WINS
5 days
± 0.4%
< 1.5%
Negligible
NA
No effect
No effect
more frequently than every
Source: Vanderpooletal. (2001).
On the other hand, the PM2 5 mass concentration will be affected by other sampling issues
mentioned, but not discussed extensively, in the previous 1996 PM AQCD (U.S. Environmental
Protection Agency, 1996a). These issues (described earlier in this chapter) include gas/particle,
particle/particle, and particle/substrate interactions for sulfates and nitrates (e.g., Appel et al.,
1984), volatilization losses of nitrates (Zhang and McMurry, 1992), SVOC artifacts (e.g.,
Eatough et al., 1993), and relative humidity effects (e.g., Keeler et al., 1988).
Several studies have been reported in which the FRM was collocated with other PM2 5
samplers in intercomparison studies. During the Aerosol Research and Inhalation Epidemiology
Study (ARIES), several PM2 5 samplers were collocated at a mixed industrial-residential site near
Atlanta, GA (Van Loy et al., 2000). These samplers included a standard PM25 FRM, a TEOM
with Nafion drier, a particulate composition monitor (PCM; Atmospheric Research and
Analysis, Gary, NC), a medium-volume (113 L/min flow rate) fine particle (PM2 5) and SVOC
sampler (i.e., a filter followed by a solid adsorbent) operated by the Desert Research Institute,
2-84
-------
TABLE 2-5. PM25 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 FRM single
Andersen Instruments FRM sequential
Andersen Instruments FRM audit
Rupprecht & Patashnick FRM single
Rupprecht & Patashnick FRM sequential
Rupprecht & Patashnick FRM audit
BGI, Inc. FRM single
BGI, Inc. FRM audit
ThermoEnvironmental Instruments FRM single
URC FRM single
URC 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).
a HEADS sampler, and a dichotomous sampler for coarse PM. The PCM sampler has three
channels, all of which have PM10 cyclone inlets. The first two channels have two denuders
preceding a 2.5-|im WINS impactor and filter packs. The first denuder is coated with Na2CO3
to remove acid gases, and the second is coated with citric acid to remove ammonia. The
third channel has a carbon-coated, parallel-plate denuder preceding the WINS impactor.
Measurements of 24-h mass from the FRM, PCM, and TEOM samplers, as well as
reconstructed PM2 5 mass (RPM), were compared for a 12-month period. The slopes for the
TEOM-FRM, PCM-FRM, and RPM-FRM correlations were 1.01, 0.94, and 0.91, respectively;
whereas the y-intercepts for each were 0.68, 0.04, and 0.98. Particulate sulfate measurements
on the FRM Teflon filter, the PCM Teflon filter, and PCM Nylon filter were nearly identical.
Nitrate results from the three filters were much less consistent, with the FRM collecting
substantially less nitrate than was collected on either the denuded nylon filter or denuder
followed by a Teflon-nylon filter sandwich. Particulate ammonia measurements were also
2-85
-------
compared and showed more scatter than the sulfate measurements but less than the nitrate
measurements.
An intercomparison of both PM10 and PM25 mass measurements was conducted during the
1998 Baltimore PM Study (Williams et al., 2000). PM monitors were collocated at a residential
indoor, residential outdoor, and ambient monitoring site within Baltimore County, MD. PM
samplers included TEOMs, PM2 5 FRMs, cyclone-based inlets manufactured by University
Research Glassware (URG), and Versatile Air Pollution Samplers (VAPSs). The VAPS is a
dichotomous sampler operating at 33 L/min (one coarse particle channel at 3 L/min, and two fine
particle channels at 15 L/min, each). In the configuration employed during this study, one fine
particle channel was operated with a Teflon filter backed by a nylon filter and preceded by
a Na2CO3-coated annular denuder; the second fine particle channel had a quartz filter preceded
by a citric acid-coated annular denuder; and the coarse particle channel had a polycarbonate
filter followed by a Zefluor filter for flow distribution. Differences in PM2 5 mass concentrations
between the samplers, although not large, were attributed to potential particle nitrate losses,
denuder losses, and losses of SVOCs for some samplers. Differences between coarse PM
concentrations, on the other hand, varied widely between the instruments.
In another intercomparison study, Tolocka et al. (200Ib) examined the magnitude of
potential sampling artifacts associated with the use of the FRM by collocating FRMs alongside
other chemical speciation samplers at four U.S. cities. The locations included a high-nitrate,
high-carbon, low-sulfate site (Rubidoux, CA); a high-crustal, moderate-carbon, moderate-nitrate
site (Phoenix); a high-sulfate, moderate-carbon, low-nitrate site (Philadelphia); and a low-PM2 5
mass site (Research Triangle Park, NC). The use of Teflon versus heat-treated quartz filters was
also examined in this study. The Teflon filters collected less nitrate than the heat-treated quartz
filters. Filters in samplers using denuders to remove organic gases collected less organic PM
than filters in samplers without denuders.
Peters et al. (2001b) compiled the results of several field studies in which a number of
FRM and other PM2 5 samplers were intercompared. In addition to the FRM samplers listed in
Table 2-5, other PM2 5 samplers included in the evaluation were the Sierra Instruments
dichotomous sampler, the Harvard Impactor, the IMPROVE sampler, and the Air Metrics
saturation monitor. Results were compiled from PM2 5 field studies conducted in Birmingham,
Denver, Bakersfield, Phoenix, Research Triangle Park, Atlanta, and Rubidoux. Limited studies
2-86
-------
on precision for the non-FRM samplers suggest that the Harvard Impactor and dichotomous
samplers had the lowest coefficients of variation (CVs), with both under 10%. The CV for this
study was calculated by dividing the sample standard deviation by the average concentration.
The IMPROVE samplers had CVs between 10 and 12%, and the Air Metrics samplers had the
highest observed CVs, over 15%. In intercomparisons with FRM samplers, the non-FRM
samplers showed strong linear relationships with the FRM sampler; however, none of the
comparisons passed the current EPA Subpart C equivalent method criteria, which EPA is in the
process of revising.
Detailed information on precision of PM samplers used in monitoring networks may be
found at EPA's Technology Transfer Network website (U.S. Environmental Protection Agency,
2002a).
2.2.4.1.3 PM1025
Measurement techniques for PM10_2 5 are somewhat more complex than those for PM2 5
or PM10, because it is necessary to isolate a size fraction between an upper 50% cut point of
10 |im Da and a lower 50% cut point of 2.5 jim Da for PM10_2 5. EPA is currently developing an
FRM for PM10_2 5. Several candidate techniques are discussed below.
The Difference Method. One approach to measuring PM10_2 5 is to make separate
measurements of PM10 and PM25 and take the difference of the resulting equilibrated masses.
One problem is that, if either the PM25 or the PM10 sampler fails, no PM10_25 measurement can be
obtained. In addition, errors in cut point, flow rate, and filter weights (both before use and after
collection and equilibration of particles) and uncertainties due to the loss of semivolatile
components of PM may occur for each size cut. Careful control of flow rate and equivalent
treatment of PM10 and PM25 filters in terms of pressure drop across the filter and temperature of
the filter during and after collection can improve precision and accuracy. Allen et al. (1999a)
summarized several sampling issues that should be considered in measuring coarse particulate
mass by difference, including the use of identical instrumentation (except cut points), filter
media, filter face velocity, and ambient-filter temperature differences; common flow
measurement devices; use of higher sampler flow rates (a 10 L/min minimum for 24-h sample is
recommended); and avoiding excessive filter loading. The concern, expressed by Allen et al.
2-87
-------
(1999a), that the "pie-plate" inlet required by the final version of the PM25 FRM might have a
different cut point than the "flat-top" inlet of the PM10 FRM, has been addressed by a wind
tunnel study, which found each to have an appropriate PM10 cut point (Tolocka et al., 2001a).
Since the difference method requires weighing two filters, the key to obtaining high
precision in the coarse mass measurement is precise measurements of filter weights. Allen et al.
(1999a) discussed techniques for increasing the precision of the difference method by reducing
errors in filter weights. These include proper temperature and humidity controls, use of a high
quality microbalance, 100% replicate weighings, control of static charge, aging of new filters,
weighing of a sufficient number of laboratory blank filters, and accounting for buoyancy errors
caused by variability in barometric pressure. Allen et al. (1999a) emphasized the necessity of
replicate weighing of filters and of using a third weighing if the difference between the first two
weights exceeds a specified minimum. Lawless et al. (2001) investigated the magnitude of
uncertainties attributed to fluctuations in some of these parameters (humidity, temperature,
drafts, vibration, and electrostatic charges) and recommended methods for improving their
control. Koistinen et al. (1999) and Hanninen et al. (2002) gave an excellent discussion of the
procedures developed to overcome problems associated with gravimetric measurements of PM25
mass in the EXPOLIS (Air Pollution Exposure Distributions Within Adult Urban Populations in
Europe) Study. They discussed factors such as corrections for buoyancy, elimination of static
charge, and increases in the mass of blank filters with time. The establishment of a temperature
and humidity controlled room required for the equilibration and weighing of filters for the FRM
is expensive. Allen et al. (2001) described a more cost-effective technique that uses a chamber
with relative humidity controlled at 34% relative humidity by a saturated aqueous solution
ofMgC!2.
Allen et al. (1999a) recommended that, in reporting precision from collocated samplers
both the CV and the square of the correlation coefficient (r2) be reported. For a study in Boston
with 27 pairs of mass data from collocated PM10 and PM25 using standard weighing methods,
they obtained a CV of 4.7% with an r2 of 0.991 for PM25, a CV of 4.4% with an r2 of 0.994
for PM10, and a CV of 15% with an r2 of 0.88 for PM10_25. By using duplicate weighings and
other techniques suggested for improving precision, they obtained a CV of 1.3% with an r2 of
0.998 for PM2 5 in a study in Chicago with 38 collocated measurements. On the basis of the
improvement in the CV for PM25, they estimated that use of the recommended techniques
-------
for PM10_25 by difference would have yielded a CV of 3.8% with an r2 of 0.98 if they had been
applied in the Chicago study.
Multistage Impaction. A second technique involves the use of impaction to isolate the
size fraction between 2.5 and 10 jim Da. In the impaction process, the air stream is first
accelerated through a small hole (nozzle) or slit. The air stream is directed so that it "impacts"
on a surface. Depending on the velocity and pressure of the air stream, particles smaller than a
certain size will follow the air stream around the impactor surface. Larger particles will impact
on the surface. In practice, impactors have 50% cut points similar to the 50% cut point for the
rejection of larger particles in PM25 and PM10 samples (see Figure 2-7).
Multistage impactors are used to separate particles into several size fractions for the
determination of mass and chemical composition as a function of size (Wang and John, 1988;
Marple et al., 1991). The major problem with the use of impactors to separate the PM10_25 jim Da
fraction of coarse particles (thoracic coarse PM) is bounce. Coarse particles tend to be dry, solid
particles. When they hit a hard surface, they can bounce and be carried away with the air stream
(e.g., Dzubay et al., 1976; Wesolowski et al., 1977; Rao and Whitby, 1978; Cheng and Yeh,
1979; Wang and John, 1987; John and Sethi, 1993). Various techniques have been used to
reduce bounce. One technique is to use a porous substance such as a glass- or quartz-fiber filter
(Chang et al., 1999) material or a polyurethane foam (Breum, 2000; Kavouras and Koutrakis,
2001). However, this technique may result in less precise separation and yield a sample that
must be extracted before chemical analyses can be performed. Another technique is to coat the
impactor with a soft wax or grease (Rao and Whitby, 1977; Turner and Hering, 1987; Pak et al.,
1992). This can cause problems with weighing and chemical analyses, and as the deposit of
particles builds up, incoming particles may not hit the soft surface, but instead hit a previously
collected hard particle and bounce off of it. The WINS impactor discussed earlier uses a filter in
a well of low volatility oil to ensure a wetted surface at all times. However, such a technique,
while appropriate for removing unwanted particles, would not yield a particle sample suitable for
weighing or for chemical analyses.
2-89
-------
Virtual Impaction. The problems of bounce and blow off of particles from impactors,
especially for the collection of large quantities of particles, was addressed by aerosol scientists in
the mid-1960s by the development of what is now known as "virtual" impaction (Hounam and
Sherwood, 1965; Conner, 1966).
In a virtual impactor, a hole is placed in the impaction plate just below the accelerating jet.
Two controlled flows allow a fraction, e.g., 10% (or another predetermined fraction, typically 5
to 20%), of the air containing the coarse particles to go through the hole and through a filter
(minor flow). A 10% minor flow gives a coarse channel enrichment factor of 10. The remaining
fraction (e.g., 90% of the airflow) containing the fine particles follows a different path and goes
through a second filter (major flow). The upper cut point is usually set by the inlet (e.g.,
10 |im Da). The flow rates, pressures, and distance from the nozzle to the virtual impactor
surface can be varied to direct particles with an Da greater than the lower cut point (i.e.,
> 2.5 |im) to go through the hole and be collected on the first filter and to direct smaller particles
(i.e., < 2.5 |im) to flow around the impactor and be collected on the second filter. Large particles
"impact" into the hole with a small amount of the air flow. The smaller particles follow the
major air flow around the impactor plate. This technique overcomes the problem of bounce.
An example of the separation into fine and coarse particles is shown in Figure 2-26.
The usefulness of this technique for collecting samples of fine and coarse particles for
chemical analysis was recognized by the EPA in the mid-1970s and led to the development of
the now well-known "dichotomous sampler" (a virtual impactor that separates particles into two
size fractions) and an associated XRF analyzer (Dzubay and Stevens, 1975; Loo et al., 1976;
Jaklevic et al., 1977; Dzubay et al., 1977). The dichotomous sampler was originally developed
for use in the Regional Air Monitoring Study (RAMS), part of the Regional Air Pollution Study
(RAPS), conducted in St. Louis, MO in the mid-1970s. Dichotomous samplers were operated at
10 RAMS sites from March 1975 to March 1977; and 33,695 filters were collected and analyzed
by XRF with an overall sampling efficiency of 97.25% (Strothmann and Schiermeier, 1979; Loo
et al., 1976; Loo et al., 1978; Dzubay, 1980; Lewis and Macias, 1980). Dichotomous samplers
were a novel concept at that time. Concern over particle losses and other problems at cut point
sizes < 2.5 jim Da influenced the decision to choose 2.5 instead of 1.0 as the cut point diameter.
Since the use of the dichotomous sampler in RAPS, considerable progress has been made
in the theory and practice of designing virtual impactors, especially in how to reduce losses and
2-90
-------
Coarse Particle Filter
Fine Particle Filter
Figure 2-26. Schematic diagram showing the principle of virtual impaction. The initial
flow, Q0, is split into a minor flow, Qt, which carries the larger particles that
impact into the hole, to the coarse particle filter and a major flow, Q2, which
carries the smaller particles that can follow the airflow, to the fine particle
filter.
Source: Loo etal. (1976).
provide a sharp cut (Masuda et al., 1979; Marple and Chien, 1980; Chen at al., 1985, 1986; Loo
and Cork, 1988). Now virtual impactors, with rectangular slits or round holes, are used to
(a) provide cut point sizes as low as 0.15 jim Da and (b) concentrate coarse, accumulation, and
ultrafme mode particles for use in health studies (Solomon et al., 1983; Marple et al., 1990;
Sioutas et al., 1994a,b,c). Dichotomous samplers were also used in a national network to
measure PM2 5 and PM10_2 5 in the Harvard Six City Study (Spengler and Thurston, 1983;
Dockery et al., 1993) and the Inhalable Particulate Network (Suggs and Burton, 1983).
A trichotomous high volume sampler has also been developed that provides samples of particles
< 1.0 |im, 1.0 |im to 2.5 jim, and > 2.5 jim (Marple and Olsen, 1995). This sampler was
intended to aid in the study of the composition of the intermodal mass in the range of 1.0 to
2.5 |imDa.
2-91
-------
More recently, two dichotomous sequential PM air samplers were collocated with a manual
FRM air sampler and operated for over a year at a waterfront site on Tampa Bay (Poor et al.,
2002). The FRM sampler was alternately configured as a PM2 5, then as a PM10 sampler. For the
dichotomous sampler measurements, daily 24-h integrated PM2 5 and PM10_2 5 ambient air
samples were collected at a total flow rate of 16.7 L/min. As was the case for earlier versions of
the dichotomous sampler, a virtual impactor split the air into flow rates of 1.67 and 15.01 L/min
and collected PM10_2 5 and PM2 5 on 47-mm diameter Teflon filters. Between the two
dichotomous air samplers, the average concentration, relative bias, and relative precision
for PM25 were 13.3 |ig/m3, 0.02% and 5.2% (n = 282), and for PM10.25 were 12.3 |ig/m3,
3.9% and 7.7% (n = 282). Federal Reference Method measurements were alternate-day 24-h
integrated PM2 5 or PM10 ambient air samples collected onto 47-mm diameter PTFE filters at a
flow rate of 16.71 L/min. Between a dichotomous and a PM2 5 FRM air sampler, the average
concentration, relative bias, and relative precision were 12.4 |ig/m3, -5.6%, and 8.2% (n = 43).
Between a dichotomous and a PM10 FRM air sampler, the average concentration, relative bias
and relative precision were 25.7 |ig/m3, -4.0%, and 5.8% (n = 102). The PM25 concentration
measurement standard errors for the two dichotomous and one FRM samplers were 0.95, 0.79,
and 1.02 |ig/m3; and for the PM10 sampler, the standard errors were 1.06, 1.59, and 1.70 |ig/m3.
The authors (Poor et al., 2002) concluded that their results indicated that "the dichotomous
samplers have superior technical merit" and demonstrate "the potential for the dichotomous
sequential air sampler to replace the combination of the PM2 5 and PM10 FRM air samplers,
offering the capability of making simultaneous, self-consistent determinations of these
particulate matter fractions in a routine ambient monitoring mode."
The dichotomous sampler provides two separate samples. However, a fraction of the
smaller particles, equal to the minor flow, will go through the virtual impaction opening with the
air stream and be collected on the coarse particle filter. In the dichotomous sampler used in the
RAPS program, 10% of the fine particles were collected with the coarse particles. Thus, in order
to determine the mass or composition of the coarse particles, it is necessary to determine the
mass and composition of the fine particles and subtract the appropriate fraction from the mass or
composition of the particles collected on the coarse particle filter. Allen et al. (1999b) discussed
potential errors in the dichotomous sampler caused by uncertainties in the coarse mass channel
enrichment factor. Virtual impactors have also been designed with a clean air jet in the center of
2-92
-------
the round nozzle. This reduces the contamination of coarse particles by fine particles, while
maintaining low losses and sharp cuts (Chen and Yeh, 1987; Chein and Lundgren, 1993). The
fine particle intrusion into the coarse particle sample can also be reduced by operating two
virtual impactors in series (Dzubay and Stevens, 1975).
Aerosol physicists have also conducted theoretical and experimental investigations of
virtual impaction using slits instead of round holes (Forney et al., 1978, 1982; Ravenhall et al.,
1978; Masuda andNakasita, 1988; Sioutas et al., 1994b,c,d; Ding and Koutrakis, 2000). The slit
virtual impactor permits a much higher flow rate than round hole virtual impactors and resolves
problems that occur with multihole virtual impactors (Marple et al., 1990; Fang et al., 1991).
The slit technique has been used to develop virtual impaction systems for concentrating particles
in the size range 0.1 to 2.5 |im Da for exposure studies using animals and humans (Sioutas et al.,
1995a, b). The slit impactor can also be used to concentrate coarse particles for measurement
(Misra et al., 2001) or exposure studies (Chang et al., 2002). In addition, ultrafine particles
(> 0.1 jim) can be concentrated by first separating ultrafine particles from larger particles, adding
water vapor to saturate the air containing the ultrafine particles, cooling the air to cause
supersaturation and growth of the ultrafine particles into the 1.0 to 4.0 |im size range, then
concentrating these particles with a slit virtual impactor, and finally, heating the air to return the
particles to their original size (Sioutas and Koutrakis, 1996; Sioutas et al., 1999; Sioutas et al.,
2000; Kim et al., 2001b,c; Geller et al., 2002).
2.2.5 Speciation Monitoring
Speciation Network and Monitoring
In addition to FRM sampling to determine compliance with PM standards, the EPA
requires states to conduct chemical Speciation sampling primarily to determine source categories
and trends (Code of Federal Regulations, 200 Ib). Source category apportionment calculations
are discussed in Chapter 3. A PM2 5 chemical Speciation Trends Network (STN) has been
deployed that consists of 54 core National Ambient Monitoring Stations (NAMS) and
approximately 250 State and Local Air Monitoring Stations (SLAMS). In addition, over
100 IMPROVE (Interagency Monitoring of Protected Visual Environments) samplers located at
regional background and transport sites can be used to fulfill SLAMS requirements. The overall
goal of the speciation program is "to provide ambient data that support the Nation's air quality
2-93
-------
program objectives" (U.S. Environmental Protection Agency, 1999). Information and reports on
the EPA's speciation monitoring program may be found on the EPA's Technology Transfer
Network at http://www.epa.gov/ttn/amtic/pmspec.html. The NAMS speciation sites will provide
routine chemical speciation data to be used in developing annual and seasonal aerosol
characterization, air quality trends analysis, and emission control strategies. The SLAMS
speciation sites will further support the NAMS network and provide information for
development of state implementation plans (SIPs).
At both NAMS and SLAMS sites, aerosol samples will be collected for the analysis of
trace elements, ions (sulfate, nitrate, ammonium, sodium, and potassium), and total carbon (TC).
The NAMS speciation sites will operate on a l-in-3-day schedule, with 10 of these sites
augmented with continuous speciation analyses for everyday operation. The SLAMS speciation
sites will generally operate on a l-in-6-day basis; however, many sites may be operated on a 1-
in-3-day basis in locations where increased data collection is needed. There are several samplers
that are suitable for use in the NAMS/SLAMS network. These samplers include an inlet cut
point with size cut characteristics comparable to the WINS FRM, proven denuder technology for
nitrate, and sampler face velocity and sample volume similar to that of the FRM. The current
samplers include three filters: (1) Teflon for equilibrated mass and elemental analysis by energy
dispersive X-ray fluorescence (EDXRF), (2) a nitric acid denuded nylon filter for ion analysis
(ion chromatography), (3) a quartz-fiber filter for elemental and organic carbon. Elemental
carbon and OC are determined by thermal-optical analysis via a modification of the NIOSH
(National Institute for Occupational Safety and Health) method 5040 (thermal-optical
transmission [TOT]). However, no corrections are made for positive artifacts caused by
adsorption on organic gases or the quartz filters or negative artifacts due to the evaporation of
SVOCs from collected particles.
Since 1987, the IMPROVE network has provided measurements of ambient PM and
associated light extinction in order to quantify PM chemical components that affect visibility at
Federal Class 1 areas that include designated national parks, national monuments, and
wilderness areas. Management of this network is a cooperative effort between the EPA,
federal land management agencies, and state governments. The IMPROVE program has
established analytical protocols for measurements of ambient concentrations of PM10, PM25,
2-94
-------
sulfates, nitrates, organic and elemental carbon, crustal material, and a number of other elements.
Information on the IMPROVE program may be found at http://vista.cira.colostate.edu/ improve.
IMPROVE aerosol monitoring consists of a combination of particle sampling and sample
analysis. The IMPROVE sampler, which collects two 24-h samples per week, simultaneously
collects one sample of PM10 on a Teflon filter, and three samples of PM2 5 on Teflon, nylon, and
quartz filters. PM10 mass concentrations are determined gravimetrically from the PM10 filter
sample, while PM2 5 mass concentrations are determined gravimetrically from the PM2 5 Teflon
filter sample. The PM2 5 Teflon filter sample is also used to determine concentrations of selected
elements using PIXE, XRF, and PESA. The PM2 5 nylon filter sample, which is preceded by a
denuder to remove acidic gases, is analyzed to determine nitrate and sulfate aerosol
concentrations using 1C. Finally, the PM2 5 quartz filter sample is analyzed for OC and EC using
the thermal-optical reflectance (TOR) method. Corrections are made for positive artifacts but
not for negative artifacts.
The STN and the IMPROVE networks represent a major advance in the measurement of
nitrate, because the combination of a denuder to remove nitric acid vapor and a Nylon filter to
adsorb nitric acid vapor volatilizing from the collected ammonium nitrate particles overcomes
the loss of nitrate from Teflon filters. However, different techniques used for measurement of
OC and EC lead to significant differences between their measurements when the two techniques
are intercompared (Chow et al., 2001). IMPROVE yields higher EC and lower OC, although
there is good agreement for TC. Another difference arises from the treatment of the positive
artifacts due to the absorption of organic gases by the quartz filters used in IMPROVE and STN
samplers. More information on these differences is given in Section 2.2.7 and details are
discussed in Appendix 2B.
Several of the PM2 5 size selectors developed for use in the EPA National PM2 5 STN were
recently evaluated by comparing their penetration curves under clean room experiments with
that of the WINS impactor (Peters et al., 2001c). The corresponding speciation monitors were
then compared to the FRM in four cities. The PM2 5 inlets tested were the SCC 2.141 cyclone
(6.7 L/min) that is in the Met One Instruments SASS sampler, the SCC 1.829 cyclone
(5.0 L/min) that is proposed for use in the Rupprecht and Patashnik real-time sulfate/nitrate
monitor, the AN 3.68 cyclone (24.0 L/min) that is in the Andersen RAAS, and the spiral
separator (7.0 L/min) that was previously in the Met One SASS. The cutpoints of the SCC
2-95
-------
cyclones compared reasonably well with the WINS (2.52 and 2.44 jim for the SCC 2.141 and
SCC 1.829, respectively, at their design flow rates), but both demonstrated a tail extending into
the coarse-particle mode. The AN inlet had the sharpest cut point curve, but the 50% cut point
diameter was 2.7 jam Da at its design flow rate. The spiral inlet had the shallowest cut point
curve: the 50% cut point was 2.69 and 2.67 jim Da for an ungreased and greased inlet,
respectively. The speciation samplers were also compared to the FRM sampler with WINS inlet
under ambient conditions in four cities. The Andersen RAAS equipped with the AN 3.68
cyclone compared well to the FRM in all four cities when compared on the basis of PM25 mass,
sulfate, and crustal concentrations. Greasing the spiral inlet in the Met One sampler improved
the performance of that sampler, which tended to give much higher PM25 concentrations than the
FRM in cities with high crustal PM.
2.2.6 Inorganic Elemental Analyses
In addition to the lighter elements (hydrogen, carbon, oxygen and nitrogen), the following
40 heavier elements are commonly found in ambient air samples: sodium, magnesium,
aluminum, silicon, phosphorus, sulfur, chlorine, potassium, calcium, titanium, vanadium,
chromium, manganese, iron, cobalt, nickel, copper, zinc, gallium, arsenic, selenium, bromine,
rubidium, strontium, yttrium, zirconium, molybdenum, palladium, silver, cadmium, indium, tin,
antimony, barium, lanthanum, gold, mercury, thallium, lead, and uranium. These heavier
elements are often indicators of air pollution sources, and several of them are considered to be
toxic (such as transition metals, water-soluble metals, and metals in certain valence states [e.g.,
Fe(II), Fe(III), Cr(III), Cr(VI), As(III), As(V)]). Various measurement methods for inorganic
elements are listed in Table 2-6. These methods differ with respect to detection limits, sample
preparation, and cost (Chow, 1995). EDXRF and PIXE are the most commonly applied methods
because they quantify more than 40 detectable elements, are nondestructive, and are relatively
inexpensive. Both were discussed in the previous 1996 PM AQCD (U.S. Environmental
Protection Agency, 1996a). TRXRF and S-XRF are newer techniques with lower detection
limits. AAS, ICP-AES, and ICP-MS are also appropriate for ionic measurements of elements
that can be dissolved. PESA provides a means for measuring elements with lower atomic
numbers (i.e., the elements from hydrogen to carbon). More detailed information on each
technique is given in Appendix 2B. 1.
2-96
-------
TABLE 2-6. MEASUREMENT METHODS FOR INORGANIC ELEMENTS
# Acronym
Full Name
Comments
1 EDXRF Energy Dispersive X-ray Fluorescence
2 S-XRF Synchrotron Induced X-ray Fluorescence
3 PIXE Proton Induced X-ray Emission
4 PESA Proton (or particle) Elastic
Scattering Analysis
5 TRXRF Total Reflection X-ray Fluorescence
INAA Instrumental Neutron Activation Analysis
AAS Atomic Absorption Spectrophotometry
ICP-AES Inductively Coupled Plasma with
Atomic Emission Spectroscopy
ICP-MS Inductively Coupled Plasma with
Mass Spectroscopy
10
SEM Scanning Electron Microscopy
Heavier elements
Heavier elements; lower detection
limits than EDXRF
Heavier elements; lower detection
limits than EDXRF
Lighter elements
Heavier elements; lower detection
limits than EDXRF
Many elements; sensitivity different
than EDXRF
Many elements that can be dissolved
Many elements that can be dissolved
Many elements that can be dissolved
Heavier elements
2.2.7 Elemental and Organic Carbon in Particulate Matter
Ambient particles from combustion sources contain carbon in several chemically and
optically distinct forms. Health- and visibility-related studies of these particles require
information about the relative contributions to total particle mass by these different forms of
carbon. With the exception of carbonate-based carbon, however, a clear classification scheme
has not yet been established to distinguish OC, light-absorbing carbon, black carbon, soot and
EC. The absence of clear, physically-based definitions results in confusion in the interpretation
of speciation data. For example, depending on the radiation wavelength specified, "light-
absorbing" carbon can include compounds that volatilize without oxidation. "Black" carbon
includes various mixtures containing "elemental" (graphitic) carbon; partially degraded,
oxidized graphitic fragments; and partially oxidized amorphous aromatic carbon. For studying
2-97
-------
visibility reduction, a measurement of light-absorbing carbon may be more useful than one of
EC. For source apportionment by receptor models, several consistent, but distinct, fractions of
carbon in both source and receptor samples are desired, regardless of their light-absorbing or
chemical properties. Differences in the ratios of the carbon concentrations in these fractions
form part of the source profiles that distinguish the contribution of one source from another
(Watson et al., 1994a,b).
Three method-dependent operational classes of carbon are commonly measured in ambient
aerosol samples collected on quartz fiber filters: (1) semivolatile organic or non-visible light-
absorbing carbon, termed "organic carbon (OC)"; (2) elemental carbon, soot, black carbon, or
light-absorbing carbon, termed "elemental carbon (EC)"; and (3) carbon present as K2CO3,
Na2CO3, MgCO3, CaCO3, termed "carbonate carbon (CC)." The sum of OC, EC, and CC in PM
gives the total carbon (TC).
The thermal-optical reflectance (TOR), thermal-optical transmission (TOT), and thermal-
manganese oxidation (TMO) methods are most commonly used for the analysis of OC and EC in
atmospheric PM. In thermal separation methods, OC is vaporized and the EC remaining on the
filter is then oxidized to CO2 and quantified by nondispersive infrared detection, by
electrochemical techniques or by reducing the CO2 to CH4 and the quantifying CH4 via flame
ionization detection (FID). Organic carbon that does not vaporize below 550 °C can pyrolyze at
higher temperatures to form additional black carbon. Thermal optical methods must correct for
this effect in order to correctly distinguish OC from EC. The various methods give similar
results for TC, but not for EC or OC, due to differing assumptions and analytical procedures
regarding the thermal behavior of ambient aerosol carbon. These methods are discussed in detail
in Appendix 2B.2.
Carbonate carbon can be determined thermally, or on a separate filter section by measuring
the carbon dioxide (CO2) evolved after acidification (Johnson et al., 1980). It is usually on the
order of 5% or less of TC for ambient particulate samples collected in urban areas (Appel, 1993).
The forms of carbon present in natural materials that may lead to the formation of
atmospheric aerosol tend to be poorly defined. Thus, the pyrolysis products of these materials
during thermal-optical analysis cannot be predicted. The Geochemical Society convened an
international steering committee in 1999 to define a set of representative black carbon and black
carbon-containing benchmark materials to be used to support ambient aerosol sample analysis.
2-98
-------
Such materials can be used to provide thermal-optical "fingerprints" for deducing primary
aerosol sources, and to establish characteristic analytic interferences or artifacts associated with
such sources. The materials recommended to date include n-hexane soot, lignocellulosic chars,
soils, marine sediments, and the NIST urban dust standard reference material (SRM 1649a). The
committee has also recommended a set of standard materials that may potentially interfere with
black carbon analyses: shale, natural organic matter, coals, and melanoidin (an amino
acid-based material). These recommendations are discussed on the steering committee's
website: http://www.du.edu/~dwismith/bcsteer.html.
2.2.8 Ionic Species
Ion chromatography is widely used for analyzing ionic species in the water-soluble portion
of suspended PM. Ion chromatography is the method of choice for the measurement of sulfate,
nitrate, ammonium, sodium, and potassium ions for the NAMS program. Aerosol strong
acidity, H+, is determined by titrating on aqueous solution of PM that is collected after passage
through a series of annular denuders to remove acid and basic gases with back-up filters to
collect the NH3 and HNO3 that volatilize from the PM during collection. The 1996 PM AQCD
(U.S. Environmental Protection Agency, 1996a) discussed measurement of ions by ion
chromatography (Section 4.3.3.1) and of strong acidity (Sections 3.3.1.1 and 4.3.3.1); so,
no further details will be discussed here.
2.2.9 Continuous Monitoring
The EPA expects that many local environmental agencies will operate continuous PM
monitors. All currently available continuous measurements of suspended particle mass share the
problem of dealing with semivolatile PM components. So as not to include particle-bound water
as part of the mass, the particle-bound water must be removed by heating or dehumidification.
However, heating also causes the loss of ammonium nitrate and SVOCs. A variety of potential
candidates for the continuous measurement of particle mass and related properties are listed in
Table 2-7. These techniques are discussed in more detail in Appendix 2B.3.
2-99
-------
TABLE 2-7. METHODS FOR CONTINUOUS MEASUREMENT OF PM MASS,
PM COMPONENTS, ETC.
# Acronym
Name
Comments
TEOM
2 TEOM
with SES
4 FDMS
5 RAMS
6 CAMM
7 BAM
9 CCPM
10 —
11 —
12 —
13 —
14 —
15
16
BAD
Tapered Element
Oscillating Microbalance
TEOM with Sample
Equilibration System
Differential TEOM
Filter Dynamics
Measurement System
Real-time total Ambient
Mass Sampler
Continuous Ambient
Mass Monitor
Beta gauge
Piezoelectric
microbalance
Continuous Coarse
Particle Monitor
Semi-continuous EC/OC
Semi-continuous nitrate
Semi-continuous sulfate
Continuous ion
chromatography of
water-soluble ions
Mass spectroscopy of
individual particles
Electrical Aerosol
Detector
Integrating nephelometer
Measures only nonvolatile components of PM.
By using dehumidification instead of heat for drying and a lower
temperature for stabilization, includes some of the semivolatile
components of PM.
Research instrument designed to measure both the nonvolatile
and semivolatile components of PM.
Commercial version of the differential TEOM.
Research instrument designed to measure total PM using
denuders and an adsorptive filter.
Measures total PM by pressure drop across a frequently
changed filter.
Measures PM mass by beta attenuation. Unless dried by heat or
dehumidification will also measure particle-bound water.
Measures mass by change in resonance frequency when particles
are deposited on a crystal.
Virtual impaction is used to concentrate PM10_2 5 which is then
measured by a TEOM.
Several commercially available instruments automate the thermal
technique for EC/OC and provide hourly measurements.
Collection of PM followed by flash vaporization and
determination of NOX provides 10 minute measurements of
paniculate nitrate.
Several techniques are available in which paniculate sulfate is
converted to SO2 which is measured by a pulsed fluorescence
analyzer.
Particles are grown by mixing with water vapor, collected in
water, and injected into an automatic ion chromatography.
Single particles are evaporated, ionized and the components
analyzed by mass spectroscopy. Several different systems are in
use in research studies.
This instrument measures charge collected by particles and gives
a continuous signal that is proportional to the integral of the
particle diameter.
Light scattering by suspended particles, collected over a large
solid angle, provides an indicator for particle mass including
particle-bound water unless the air sample is dried by heating or
dehumidification.
2-100
-------
2.2.10 Measurements of Individual Particles
Recently, several researchers have developed instruments for real-time in situ analysis of
single particles (e.g., Noble and Prather, 1996; Gard et al., 1997; Johnson and Wexler, 1995;
Silva and Prather, 1997; Thomson and Murphy, 1994). Although the technique varies from one
laboratory to another, the underlying principle is to fragment each particle into ions, using either
a high-power laser or a heated surface and, then, a time-of-flight mass spectrometer (TOFMS) to
measure the ion fragments in a vacuum. Each particle is analyzed in a suspended state in the air
stream (i.e., without collection), avoiding sampling artifacts associated with impactors and
filters. The technique is called aerosol time-of-flight mass spectrometry (ATOFMS).
By measuring both positive and negative ions from the same particle, information can be
obtained about the composition of individual particles of known aerodynamic diameter. This
information is especially useful in determining the 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. Four systems
for measurement of single particles using mass spectroscopy are reviewed by Middlebrook et al.
(2003). An example of the type of information that can be determined is shown in Figure 2-27.
Because particles are analyzed individually, biases in particle sampling (including biases in
the efficiency of particle transmission into the sensor chamber as a function of size; particle size
measurement, and detection of particles prior to fragmentation) represent a major challenge for
these instruments. Moreover, laser ablation has a relatively large variability in ion yields, i.e.,
identical samples would yield relatively large differences in mass spectrometer signals
(Thomson and Murphy, 1994). Therefore, quantitation is inherently difficult (Murphy and
Thomson, 1997). Quantitation will be even more challenging for complex organic mixtures
because of the following two reasons: (1) a large number of fragments are generated from each
molecule, and (2) ion peaks for organics can be influenced or obscured by inorganic ions
(Middlebrook et al., 1998). Nonetheless, scientists have been successful in using these
techniques to identify the presence of organics in atmospheric particles and laboratory-generated
particles (i.e., as contaminants in laboratory-generated sulfuric acid droplets) as well as the
identification of specific compound classes such as PAHs in combustion emissions (Castaldi and
Senkan, 1998; Hinz et al., 1994; Middlebrook et al., 1998; Murphy and Thomson, 1997;
2-101
-------
o
o
(0
Q.
0)
JJJ
0)
a:
Organic
Marine
Soil
0.3 0.40.5 0.6 0,7 0.8 0.91.0
2,0
3.0 40
Aerodynamic Diameter, Da (|jm)
Figure 2-27. Size distribution of particles divided by chemical classification into organic,
marine, and soil (or crustal).
Source: Noble and Prather (1998).
Neubauer et al., 1998; Noble and Prather, 1998; Reilly et al., 1998; Silva and Prather, 1997).
A new multivariate technique for calibrating of ATOFMS using microorifice impactors shows
promise for simplifying the calibration process (Fergenson et al., 2001). This calibration
technique has been applied to gasoline and diesel particles to demonstrate the feasibility of using
this technique for the source apportionment of gasoline and diesel particles in an atmospheric
mixture (Song et al., 2001).
Previously, ATOFMS systems have only been able to characterize particles larger than
-0.2 to 0.3 |im in diameter. Recently, Wexler and colleagues (Carson et al., 1997; Ge et al.,
1998) developed an ATOFMS instrument that is able to size, count, and provide chemical
composition on individual particles ranging in size from 10 nm to 2 |im.
2-102
-------
2.2.11 Low Flow Filter Samples for Multiday Collection of
Particulate Matter
For some purposes, such as demonstrating the attainment of an annual standard or as an
exposure indicator for epidemiologic studies of chronic health effects, 24-h measurements are
not essential. Annual or seasonal averages may be adequate and multiday sampling techniques
can result in lower costs for weighing, chemical analysis, and travel time to change filters. The
multiday sampler serves a second purpose. Most commercially available samplers are optimized
for collecting 24-h samples of PM concentrations as found in the United States, Europe, or
Japan, but many other parts of the world have significantly higher PM concentrations. Under
such conditions, the 16.7 L/min flow through 37- or 47-mm diameter filters may overload the
filter, preventing the sampler from maintaining the prescribed flow rate for 24 h. A low-flow
sampler with a 0.4 L/min flow rate and a 47-mm diameter filter has been designed by Aerosol
Dynamics, Inc. With this sampler, the sample collection time can be chosen to suit the ambient
concentration level. This sampler, with a 1-week collection period, has been used to characterize
PM2 5 in Beijing, PRC (He et al., 2001). It is also being used with a 2-week collection period in a
chronic epidemiologic study in southern California (Gauderman, et al., 2000).
The low-flow sampler, as described by He et al. (2001), has three PM2 5 channels. One
channel collects PM on a Teflon filter for gravimetric mass measurement and elemental analysis
by XRF. A second channel collects PM on a quartz filter for OC and EC analysis. A denuder
may be added to the second channel to remove organic gases, as may a backup filter to collect
SVOCs. The third channel uses a carbonate denuder to remove acid gases such as HNO3 and
SO2, a Teflon filter to collect PM for analysis of ions by 1C, and a nylon filter to collect
volatilized nitrate. The Teflon filter can also be weighed prior to extraction; thus, the multiday
sampler can provide the information needed for source apportionment by chemical mass balance
techniques (Watson et al., 1990a,b; U.S. Environmental Protection Agency, 2002b).
For monitoring sites with high day-to-day variability in PM concentrations, a sample
integrated over a week may provide a more accurate measurement of the annual average than
can be obtained by l-in-3- or l-in-6-day sampling schedules. Daily PM data from Spokane, WA
were resampled to simulate common sampling schedules and the error due to less-than-daily
sampling was computed (Rumburg et al., 2001). The error in the annual mean concentration
for PM25, expressed as a percentage difference from the daily sampling mean, was 1.7, 3.4, and
2-103
-------
7.7% for l-in-2-, l-in-3, and l-in-6-day sampling, respectively. These results may or may not be
representative of other locations and time periods.
2.3 SUMMARY AND KEY POINTS
2.3.1 Chemistry and Physics of Atmospheric Particles
Atmospheric particles originate from a variety of sources and possess a range of
morphological, chemical, physical, and thermodynamic properties. The composition and
behavior of aerosols are linked with those of the surrounding gas. An aerosol may be defined as
a suspension of solid or liquid particles in air and includes both the particles and all vapor or gas
phase components of air. However, the term aerosol is also often used to refer to the suspended
particles only.
A complete description of atmospheric PM would include an accounting of the chemical
composition, morphology, and size of each particle, and the relative abundance of each particle
type as a function of particle size. Recent developments in single particle analysis techniques
are bringing this description closer to reality.
The diameter of a spherical particle may be determined geometrically, from optical or
electron microscopy, by light scattering and Mie theory, or by a particle's behavior (e.g.,
electrical mobility or its aerodynamic behavior). However, the various types of diameters may
be different, and atmospheric particles often are not spherical. Therefore, particle diameters are
described by an "equivalent" diameter. Aerodynamic diameter, Da (the diameter of a unit
density sphere that would have the same terminal settling velocity as the real particle), and the
Stokes diameter, Dp (the diameter of a sphere of the same density as the particle that would have
the same aerodynamic resistance or drag), are the most widely used equivalent diameters.
Atmospheric size distributions show that most atmospheric particles are quite small,
< 0.1 |im in diameter; whereas most of the particle volume (and therefore most of the mass) is
found in particles > 0.1 jim. An important feature of the mass or volume size distributions of
atmospheric particles is their multimodal nature. Volume distributions, measured in ambient air
in the United States, almost always have a minimum between 1.0 and 3.0 jim. The portion of the
size distribution that contains particles that are mostly larger than the minimum is called
"coarse" particles or the "coarse" mode. The portion of the size distribution that contains
2-104
-------
particles that are mostly smaller than the minimum is called "fine" particles and includes several
modes. "Accumulation mode" refers to that portion of fine particles with diameters above about
0.1 |im. The portion of fine particles with diameters < 0.1 jim are usually called "ultrafme"
by toxicologists and epidemiologists and "nanoparticles" by aerosol physicists and material
scientists. In the number distribution of ultrafme particles, particles in the size range < 0.01 are
called the nucleation mode and particles between 0.01 and 0.1 are called the Aitken mode. The
Aitken mode can be observed as a separate mode in mass or volume distributions only in clean
or remote areas or near sources of new particle formation by nucleation.
The aerosol community uses three different approaches or conventions in the classification
of particles by size: (1) modes, based on the observed size distributions and formation
mechanisms; (2) dosimetry or occupational sizes, based on the entrance into various
compartments of the respiratory system; and (3) cut point, usually based on the 50% cut point of
the specific sampling device, including legally specified, regulatory sizes for air quality
standards. Over the years, the terms fine and coarse as applied to particle sizes have lost their
original precise meanings. In any given article, therefore, the meaning of fine and coarse, unless
defined, must be inferred from the author's usage. In particular, PM25 and fine particles are not
equivalent. PM25 refers to the aggregate sample of PM that is collected following a size-
selective inlet with a specified penetration as a function of size and a 50% cutpoint of 2.5 jim Da.
It may also be used to refer to the number (or other measure) of particles suspended in the
atmosphere that would be collected by such a sampler). PM10 is defined similarly. PM10_2 5
refers to the sample that would be collected if the PM2 5 component could be removed from
a PM10 sample.
Several processes influence the formation and growth of particles. New particles may be
formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
material condenses onto existing particles. Particles may also grow by coagulation as two
particles combine to form one. Gas phase material condenses preferentially on smaller particles,
and the rate constant for coagulation of two particles decreases as the particle size increases.
Therefore, nucleation-mode particles grow into Aitken-mode particles and Aitken-mode
particles grow into the accumulation mode, but growth of accumulation-mode particles into the
coarse mode size range is unusual.
2-105
-------
The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
ions; particle-bound water; elemental or black carbon; a great variety of organic compounds; and
crustal material. Atmospheric PM contains a large number of elements in various compounds
and concentrations and hundreds of specific organic compounds. Particulate material can be
primary or secondary. It is designated primary if it is in the same chemical form in which it was
emitted into the atmosphere. It is called secondary PM if it is formed by chemical reactions in
the atmosphere. Primary coarse particles are usually formed by mechanical processes; whereas
primary fine particles are emitted from sources either directly as particles or as vapors that
rapidly condense to form particles. Water-soluble and potentially toxic gases such as SO2, H2O2,
and formaldehyde, which in the absence of particles would be removed by deposition to the
mucus on the upper airways of the respiratory tract, may dissolve in particle-bound water and be
carried into the air exchange regions of the lungs.
Most of the sulfate and nitrate and a portion of the organic compounds in atmospheric
particles are secondary. Secondary aerosol formation depends on numerous factors including the
concentrations of precursors; the concentrations of other gaseous reactive species such as ozone,
hydroxyl radical, peroxy radicals, and hydrogen peroxide; atmospheric conditions, including
solar radiation and relative humidity; and the interactions of precursors and preexisting particles
within cloud or fog droplets or on or in the liquid film on solid particles. As a result, it is
considerably more difficult to relate ambient concentrations of secondary species to sources of
precursor emissions than it is to identify the sources of primary particles.
The lifetimes of particles vary with particle size. Coarse particles can settle rapidly from
the atmosphere within minutes or hours and normally travel only short distances. However,
when mixed high into the atmosphere, as in dust storms, the smaller-sized, coarse-mode particles
may have longer lives and travel greater distances. Accumulation-mode particles are kept in
suspension by normal air motions and have a lower deposition velocity than coarse-mode
particles; they can be transported thousands of kilometers and remain in the atmosphere for a
number of days. They are removed from the atmosphere primarily by cloud processes. Dry
deposition rates are expressed in terms of a deposition velocity that varies with the particle size,
reaching a minimum between 0.1 and 1.0 jim Da.
Particulate matter is a factor in acid deposition. Particles serve as cloud condensation
nuclei and contribute directly to the acidification of rain. In addition, the gas-phase species that
2-106
-------
lead to the dry deposition of acidity are also precursors of particles. Therefore, reductions in SO2
and NOX emissions will decrease both acid deposition and PM concentrations, but not
necessarily in a linear fashion. Sulfuric acid, ammonium nitrate, and organic particles also are
deposited on surfaces by dry deposition and can contribute to ecological effects.
2.3.2 Measurement of Atmospheric Particles
The decision by the EPA to revise the PM standards by adding daily and yearly standards
for PM2 5 has led to a renewed interest in the measurement of atmospheric particles and to a
better understanding of the problems in their precise and accurate measurements. Unfortunately,
it is very difficult to measure and characterize particles suspended in the atmosphere.
Particulate matter monitoring is used to develop information to guide the implementation
of standards (i.e., to identify particles sources, to determine whether or not a standard has been
attained) and to determine health, ecological, and radiative effects. Federal Reference Methods
(FRM) specify techniques for measuring PM10 and PM2 5. Particles are collected on filters and
mass concentrations are determined gravimetrically. The PM10 FRMs consist of low flow rate
(1 L/h) and high flow rate (67.92 L/h) samplers, each with a PM10 inlet/impactor. The high flow
rate sampler uses a 20.3 x 35.4 cm2 quartz filter whereas the low flow rate sampler uses a 47-mm
Teflon filter, each with a collection efficiency of > 99% as described in Appendix M to 40 CFR,
Part 50 (Code of Federal Regulations, 200 le). The PM25 FRM is similar to the low flow rate
PM10 FRM except that it includes a PM2 5 impactor with an oil-covered impaction substrate to
remove particles > 2.5 jim and a filter with a collection efficiency of greater than 99.7% as
described in Appendix L, 40 CFR, Part 50 (Code of Federal Regulations, 200 le). Both
techniques provide relatively precise (± 10%) methods for determining the mass of material
remaining on a filter after equilibration.
Despite considerable progress in measuring the atmospheric PM mass concentration,
numerous uncertainties continue to exist as to the relationship between the mass and composition
of material remaining on the filter as measured by the FRMs and the mass and composition of
material that exists in the atmosphere as suspended PM. There is no reference standard for
particles suspended in the atmosphere, nor is there an accepted way to remove particle-bound
water without losing some of the semivolatile components of PM such as ammonium nitrate and
semivolatile organic compounds. It also is difficult to cleanly separate fine and coarse PM. As a
2-107
-------
result, EPA defines accuracy for PM measurements in terms of agreement of a candidate sampler
with a reference sampler. Therefore, intercomparisons of samplers become very important in
determining how well various samplers agree and how various design choices influence what is
actually measured.
Current filtration-based mass measurements lead to significant evaporative losses of a
variety of semivolatile components (i.e., species that exist in the atmosphere in dynamic
equilibrium between the condensed phase and gas phase) during and possibly after collection.
Important examples include ammonium nitrate, SVOCs, and particle-bound water. Loss of these
components may significantly affect the quality of the measurement and can lead to either
positive or negative sampling artifacts. Negative artifacts resulting from loss of ammonium
nitrate and S VOC may occur during sampling because of changes in temperature, relative
humidity, or the composition of the aerosol, or because of the pressure drop across the filter.
Negative artifacts also may occur during handling and storage because of evaporation. Positive
artifacts occur when gas-phase compounds (H2O, HNO3, SO2, and organic compounds) absorb
onto or react with filter media or collected PM, or when some particle-bound water is not
removed.
Sampling systems for semivolatile PM components make use of denuders to remove the
gas-phase fraction and absorptive filters to remove the condensed phase while retaining any
material that subsequently evaporates from the collected PM. The loss of particulate nitrate may
be determined by comparing nitrate collected on a Teflon filter to that collected on a nylon filter
(which absorbs nitric acid as it evaporates from ammonium nitrate particles) preceded by a
denuder to remove gas-phase nitric acid. In two studies in southern California, the PM2 5 mass
lost because of volatilization of ammonium nitrate was found to represent 10 to 20% of the
total PM2 5 mass and almost a third of the nitrate. Denuder/absorptive filter sampling systems
also have been developed for measuring particulate phase organic compounds. This technique is
an improvement over the filter/adsorbent collection method. However, the denuder systems
currently discussed in the literature are not straightforward in their use, and the user must have a
thorough understanding of the technology. The FRM for PM25 will likely suffer loss of
particulate nitrates and SVOCs, similar to the losses experienced with other single-filter
collection systems.
2-108
-------
It is generally desirable to collect and measure ammonium nitrate and SVOCs as part of
PM mass. However, it is usually desirable to remove the particle-bound water before
determining the mass. In some situations, it may be important to know how much of the
suspended particle's mass or volume results from particle-bound water. Calculations and
measurements indicate that aerosol water content is strongly dependent on relative humidity and
composition. Particle-bound water can represent a significant mass fraction of the PM
concentration at relative humidities above 60%. A substantial fraction of accumulation-mode
PM is hygroscopic or deliquescent. The more hygroscopic particles tend to contain more
sulfates, nitrates, and secondary organic compounds, while the less hygroscopic particles tend to
contain more elemental carbon, primary organic compounds, and crustal components. Fresh,
submicrometer-size soot particles may tend to shrink with increasing relative humidity because
of a structural change. The effects of relative humidity on the sorption of semivolatile organic
compounds on particles are not well understood. The amount of water sorbed to an atmospheric
aerosol may be affected by the presence of an organic film on the particle, which can impede the
transport of water across the surface.
Fine and coarse particles differ not only in formation mechanisms and size, but also in
sources, composition, and chemical, physical, and biological properties. Fine and coarse
particles overlap in the intermodal size range (1 to 2.5 jim Da). As relative humidity increases,
fine particles grow into this size range; as relative humidity decreases, more coarse particles may
be suspended in this size range. It is desirable to measure fine PM and coarse PM separately in
order to properly allocate health effects to either fine PM or coarse PM as well as to correctly
determine sources by factor analysis or chemical mass balance. The selection of a cut point of
2.5 |im as a basis for EPA's 1997 NAAQS for fine particles (Federal Register, 1997) and its
continued use in many health effects studies reflects the importance placed on more complete
inclusion of accumulation-mode particles, while recognizing that intrusion of coarse-mode
particles can occur under some conditions with this cut point.
In addition to FRM sampling of equilibrated mass to determine compliance with PM
standards, EPA requires states to conduct speciation sampling, primarily to determine source
categories and trends. The current speciation samplers collect PM2 5 on three filters:
(1) a Teflon filter for gravimetric determination of mass and for analysis of heavy elements by
X-ray fluorescence; (2) a nylon filter preceded by a nitric acid denuder for artifact-free
2-109
-------
determination of nitrate and measurement of other ionic species by ion chromatography; and
(3) a quartz filter for measurement of elemental carbon (EC) and organic carbon (OC).
In addition, IMPROVE samplers provide information on regional PM background and transport.
IMPROVE samplers, in addition to the three types of filters collected by the speciation samplers,
also collect a PM10 sample. The IMPROVE and speciation networks use slightly different
methods for determination of EC and OC. The two methods agree on total carbon (TC) but
differ in the split of TC into EC and OC. The two methods also differ in their correction for
positive artifacts due to absorption of volatile organic compounds on the quartz filters. Neither
EC/OC method provides for any correction for negative artifacts due to the evaporation of
SVOCs from the collected particles.
The EPA expects that monitoring agencies will operate continuous PM monitors and is in
the process of providing guidance regarding appropriate continuous monitoring techniques.
All currently available techniques for continuous measurements of suspended particle mass, e.g.,
the integrating nephelometer, the beta-absorption monitor, and the Tapered Element Oscillating
Microbalance (TEOM), share the problem of dealing with semivolatile PM components: that is,
in order to not include particle-bound water as part of the mass, the particle-bound water must be
removed by heating or dehumidification; however, heating causes ammonium nitrate and
SVOCs to evaporate. The TEOM monitor operates at a constant, but higher than ambient,
temperature to remove particle-bound water, whereas the FRM is required to operate at no more
than 5 °C above the ambient temperature. Subsequently, much of the particle-bound water is
removed during equilibration at 40% relative humidity. This difference in techniques for the
removal of particle-bound water causes differences in the measured mass concentration between
the TEOM and FRMs.
Several new techniques for continuous PM mass measurements are currently being field
tested. The RAMS measures the total mass of collected particles including semivolatile species
with a TEOM monitor using a "sandwich filter." The sandwich contains a Teflon-coated
particle-collection filter followed by a charcoal-impregnated filter to collect any semivolatile
species lost from the particles during sampling. The RAMS uses a Nafion dryer to remove
particle-bound water from the suspended particles and a particle concentrator to reduce the
quantity of gas phase organic compounds that must be removed by the denuder. The CAMM
estimates ambient PM mass by measuring of the increase in the pressure drop across a
2-110
-------
membrane filter caused by particle loading. It also uses a Nafion dryer to remove particle-bound
water. A new differential TEOM offers the possibility of measuring both nonvolatile and
semivolatile PM mass. In addition to continuous mass measurement, a number of techniques for
continuous measurement of sulfate, nitrate, or elements are being tested.
2.3.3 Key Points
Fine and Coarse Particles
Particle size distributions show that atmospheric particles exist in two classes, fine
particles and coarse particles. Fine and coarse particles are defined primarily in terms of their
formation mechanisms and size; and they also differ in sources, chemical composition, and
removal processes (see Table 2-2). Subsequent chapters will show that fine and coarse particles
also differ with regard to aspects of concentration, exposure, dosimetry, toxicology, and
epidemiology.
These differences support the setting of separate standards for fine and coarse particles.
Fine and coarse particles overlap in the size range between 1 and 3 jim Da where PM
concentrations are at a minimum. Coarse particles are generally larger than this minimum and
are generally formed by mechanical processes. Coarse particles and coarse-mode particles are
equivalent terms. Fine particles are generally smaller than the minimum and are generally
formed by coagulation and condensation of gases. Fine particles are subdivided into
accumulation, Aitken, and nucleation modes. In earlier texts, nuclei mode referred to the size
range now split into the Aitken and nucleation modes (see Figures 2-5 and 2-6).
Measurement of Mass and Composition
Nonvolatile Paniculate Matter. Analytical techniques exist for measurement of the mass
and chemical composition of PM retained on a filter (nonvolatile mass) in terms of elements
(except carbon) and certain key ions (sulfate, nitrate, hydrogen, and ammonium). Acceptable
measurements can be made of the TC retained on a filter. However, the split into OC and EC
depends on the operational details of the analytical methods and varies somewhat among
methods. Determination of the various organic compounds in the organic carbon fraction
remains a challenge.
2-111
-------
Semivolatile Paniculate Matter. Important components of atmospheric PM (particle-
bound water, ammonium nitrate, and many organic compounds) are termed semivolatile because
significant amounts of both the gaseous and condensed phases may exist in the atmosphere in
equilibrium. While particle-bound water is not considered to be a pollutant, it can significantly
influence the size and light scattering and absorbing properties of particles and may also act as a
carrier to convey dissolved gases or reactive species into the lungs. Most of the particle-bound
water is removed by heating the particles or by equilibration of the collected particles at a low
relative humidity (40%) for 24 hours. However, these processes also cause the loss of other
semivolatile components. Semivolatile components also evaporate from the filter during
sampling due to the pressure drop across the filter or due to a reduction in the atmospheric
concentration during the sampling time. Filter collection and equilibration techniques for PM,
such as prescribed by the FRM, lose a fraction of the semivolatile PM. Continuous methods
must dry the PM to remove particle-bound water. If heating is used to dry the particles, more of
the semivolatile components may be removed than are lost in filter sampling. Collection and
retention of ammonium nitrate and SVOCs represents a major challenge in the effort to move to
the continuous measurement of PM mass. The use of diffusion dryers, which dehumidify the air
stream without heating, represents a promising approach. Uncertainty in the efficiency of the
retention of ammonium nitrate and organic compounds on filters also impacts source category
attribution and epidemiologic studies.
Separation of Fine and Coarse Particulate Matter
Satisfactory techniques are available to separate fine particles from coarse particles and to
collect the fine particles on a filter. However, no consensus exists as yet on the best technique
for collecting a coarse particle sample for determination of mass and composition. Candidates
include multistage impaction, virtual impaction, and difference (subtracting PM2 5 mass or
composition from PM10 mass or composition). Advances in the theory and practice of virtual
impaction suggest that it would be possible to design virtual impactors with much less than the
10% of fine PM collected in the coarse PM sample as is now the case for the dichotomous
samplers used in air quality studies and with penetration curves as sharp as those used in the
current FRM for PM2 5.
2-112
-------
REFERENCES
Aiginger, H. (1991) Historical development and principles of total reflection x-ray fluorescence analysis (TXRF).
Spectrochim. ActaPartB 46: 1313-1321.
Aiginger, H.; Streli, C. (1997) Total reflection x-ray fluorescence spectrometry (TXRF) for trace element analysis.
Spectrosc. Eur. 9: 16-21.
Allen, G.; Sioutas, C.; Koutrakis, P.; Reiss, R.; Lurmann, F. W.; Roberts, P. T. (1997) Evaluation of the TEOM®
method for measurement of ambient paniculate mass in urban areas. J. Air Waste Manage. Assoc.
47: 682-689.
Allen, G. A.; Oh, J. A.; Koutrakis, P.; Sioutas, C. (1999a) Techniques for high-quality ambient coarse paniculate
mass measurements. J. Air Waste Manage. Assoc. 49(special issue): PM133-PM141.
Allen, J. O.; Paschkewitz, J. S.; Plummer, E. F.; Lafleur, A. L.; Sarofim, A. F.; Smith, K. A. (1999b) Absorption of
semi-volatile compounds in oiled impaction substrates: measurement of pyrene absorption. Aerosol Sci.
Technol. 30: 16-29.
Allen, G. A.; Lawrence, J.; Koutrakis, P. (1999c) Field validation of a semi-continuous method for aerosol black
carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
southwestern PA. Atmos. Environ. 33: 817-823.
Allen, R.; Box, M.; Larson, T. V.; Liu, L. J. (2001) A cost-effective weighing chamber for paniculate matter filters.
J. Air Waste Manage. Assoc. 51: 1650-1653.
American Conference of Governmental Industrial Hygienists. (1994) 1994-1995 threshold limit values for chemical
substances and physical agents and biological exposure indices. Cincinnati, OH: American Conference of
Governmental Industrial Hygienists.
Andrews, E.; Larson, S. M. (1993) Effect of surfactant layers on the size changes of aerosol particles as a function of
relative humidity. Environ. Sci. Technol. 27: 857-865.
Andrews, E.; Saxena, P.; Musarra, S.; Hildemann, L .M.; Koutrakis, P.; McMurry, P. H.; Olmez, I.; White, W. H.
(2000) Concentration and composition of atmospheric aerosols from the 1995 SEAVS experiment and a
review of the closure between chemical and gravimetric measurements. J. Air Waste Manage. Assoc.
50: 648-664.
Appel, B. R. (1993) Atmospheric sample analysis and sampling artifacts. In: Willeke, K.; Baron, P. A., eds. Aerosol
measurement: principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 233-259.
Appel, B. R.; Tokiwa, Y.; Haik, M. (1981) Sampling of nitrates in ambient air. Atmos. Environ. 15: 283-289.
Appel, B. R.; Tokiwa, Y.; Haik, M.; Kothny, E. L. (1984) Artifact paniculate sulfate and nitrate formation on filter
media. Atmos. Environ. 18: 409-416.
Arnott, W. P.; Moosmtiller, H.; Rogers, C. F.; Jin, T.; Bruch, R. (1999) Photoacoustic spectrometer for measuring
light absorption by aerosol; instrument description. Atmos. Environ. 33: 2845-2852.
Arnott, W. P.; Moosmtiller, H.; Walker, J. W. (2000) Nitrogen dioxide and kerosene-flame soot calibration of
photoacoustic instruments for measurement of light absorption by aerosols. Rev. Sci. Instruments
71:4545-4552.
Ayers, G. P.; Keywood, M. D.; Gras, J. L. (1999) TEOM vs. manual gravimetric methods for determination of PM2 5
aerosol mass concentrations. Atmos. Environ. 33: 3717-3721.
Babich, P.; Davey, M.; Allen, G.; Koutrakis, P. (2000) Method comparisons for paniculate nitrate, elemental carbon,
and PM2 5 mass in seven U.S. cities. J. Air Waste Manage. Assoc. 50: 1095-1105.
Bacon, J. R.; Chenery, S. R. N.; Ellis, A. T.; Fisher, A.; McMahon, A. W.; Potts, P. J.; Williams, J. G. (1995)
Atomic spectrometry update-atomic mass spectrometry and x-ray fluorescence spectrometry. J. Anal. At.
Spectrom. 10: 253R-310R.
Baldwin, D. P.; Zamzow, D. S.; D'Silva, A. P. (1994) Aerosol mass measurement and solution standard additions for
quantitation in laser ablation-inductively coupled plasma atomic emission spectrometry. Anal. Chem.
66: 1911-1917.
Ballach, J.; Hitzenberger, R.; Schultz, E.; Jaeschke, W. (2001) Development of an improved optical transmission
technique for black carbon (BC) analysis. Atmos. Environ. 35: 2089-2100.
Baron, P. A.; Willeke, K., eds. (2002) Aerosol measurement: principles, techniques, and applications. 2nd ed.
New York, NY: John Wiley & Sons.
Benner, R. L.; Stedman, D. H. (1989) Universal sulfur detection by chemiluminescence. Anal. Chem.
61: 1268-1271.
Benner, R. L.; Stedman, D. H. (1990) Field evaluation of the sulfur chemiluminescence detector. Environ. Sci.
Technol. 24: 1592-1596.
2-113
-------
Bergin, M. H.; Ogren, J. A.; Schwartz, S. E.; Mclnnes, L. M. (1997) Evaporation of ammonium nitrate aerosol in a
heated nephelometer: implications for field measurements. Environ. Sci. Technol. 31: 2878-2883.
Berner, A. (1989) Haze and its relation to atmospheric accumulation aerosol. Sci. Total Environ. 86: 251-263.
Bertoni, G.; Febo, A.; Perrino, C.; Possanzini, M. (1984) Annular active diffuse sampler: a new device for the
collection of organic vapours. Ann. Chim. (Rome) 74: 97-104.
Billiet, J.; Dams, R.; Hoste, J. (1980) Multielement thin film standards for XRF analysis. X-Ray Spectrom.
9:206-211.
Binkowski, F. S.; Ching, J. K. S. (1995) Modeling fine paniculate mass and visibility using the EPA regional
paniculate model. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research
and Development, National Exposure Research Laboratory; report no. EPA/600/A-95/143. Available from:
NTIS, Springfield, VA; PB96-139100.
Binkowski, F. S.; Shankar, U. (1995) The regional paniculate matter model. I. Model description and preliminary
results. J. Geophys. Res. [Atmos.] 100: 26,191-26,209.
Birch, M. E. (1998) Analysis of carbonaceous aerosols: interlaboratory comparison. Analyst (Cambridge, UK)
123: 851-857.
Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of paniculate diesel
exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
Birmili, W.; Wiedensohler, A. (1998) The influence of meteorological parameters on ultrafine particle production at
a continental site. J. Aerosol Sci. 29(suppl. 1): S1015-S1016.
Bowers, W. D.; Chuan, R. L. (1989) Surface acoustic-wave piezoelectnc crystal aerosol mass monitor. Rev. Sci.
Instrum. 60: 1297-1302.
Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1991) Indoor and outdoor concentrations of inorganic acidic
aerosols and gases. J. Air Waste Manage. Assoc. 41: 171-181.
Breum, N. O. (2000) The dust holding capacity of porous plastic foam used in particle size-selective sampling.
J. Aerosol Sci. 31:379-385.
Brock, J. R.; Durham, J. L. (1984) Aqueous aerosol as a reactant in the atmosphere. In: Calvert, J. G., ed. SO2, NO
andNO2 oxidation mechanisms: atmospheric considerations. Boston, MA: Butterworth Publishers;
pp. 209-249. (Teasley, J. L, ed. Acid precipitation series: v. 3).
Brook, J. R.; Dann, T. F. (1999) Contribution of nitrate and carbonaceous species to PM25 observed in Canadian
cities. J. Air Waste Manage. Assoc. 49: 193-199.
Brorstrom, E.; Grennfelt, P.; Lindskog, A. (1983) The effect of nitrogen dioxide and ozone on the decomposition of
particle-associated polycyclic aromatic hydrocarbons durtng sampling from the atmosphere. Atmos. Environ.
17: 601-605.
Bunz, H.; Koyro, M.; Mohler, O.; Saathoff, H. (1996) Temperature dependence of the redistribution mechanisms for
volatile components of atmospheric aerosol particles. J. Aerosol Sci. 27(suppl. 1): S653-S654.
Cachier, H.; Bremond, M.-P.; Buat-Menard, P. (1989) Determination of atmospheric soot carbon with a simple
thermal method. Tellus 41B: 379-390.
Cadle, S. H.; Groblicki, P. J. (1982) An evaluation of methods for the determination of organic and elemental carbon
in paniculate samples. In: Wolff, G. T.; Klimisch, R. L., eds. Paniculate carbon: atmospheric life cycle.
New York, NY: Plenum Press; pp. 89-109.
Cadle, S. H.; Mulawa, P.A. (1990) Atmospheric carbonaceous species measurement methods comparison study:
General Motors results. Aerosol Sci. Technol. 12: 128-131.
Cahill, T. A. (1981) Innovative aerosol sampling devices based upon PIXE capabilities. In: Johansson, S. A. E., ed.
Particle induced x-ray emission and its analytical applications: proceedings of the 2nd international
conference on PIXE and its analytical applications; June, 1980; Lund, Sweden. Amsterdam, The Netherlands:
North-Holland Publishing Co. Nucl. Instr. Meth. Phys. Res. 181: 473-480.
Cahill, T. A. (1985) Proton-induced X-ray emission (PIXE). In: Res, J. P., ed. Metals Handbook: Materials
Characterization. Cleveland, OH: American Society for Metals.
Cahill, T. A. (1990) Particle-induced x-ray emission. In: Mills, K., ed. Metals Handbook; v. 10, p. 102. 9th ed.
Cleveland, OH: American Society for Metals.
Cahill, T. A.; Eldred, R. A.; Motallebi, N.; Malm, W. C. (1989) Indirect measurement of hydrocarbon aerosols
across the United States by nonsulfate hydrogen-remaining gravimetric mass correlations. Aerosol Sci.
Technol. 10: 421-429.
Cahill, T. A.; Wilkinson, K.; Schnell, R. (1992) Composition analyses of size-resolved aerosol samples taken from
aircraft downwind of Kuwait, spring 1991. J. Geophys. Res. [Atmos.] 97: 14,513-14,520.
2-114
-------
Cahill, T. A.; Wakabayashi, P.; James, T. A. (1996) Chemical states of sulfate at Shenandoah National Park during
summer, 1991. Nucl. lustrum Methods Phys. Res. B 109/110: 542-547.
Camp, D. C.; Stevens, R. K.; Cobourn, W. G.; Husar, R. B.; Collins, J. F.; Huntzicker, J. J.; Husar, J. D.; Jaklevic,
J. M.; McKenzie, R. L.; Tanner, R. L.; Tesch, J. W. (1982) Intercomparison of concentration results from fine
particle sulfur monitors. Atmos. Environ. 16: 911-916.
Campbell, A. J.; Humayun, M. (1999) Trace element microanalysis in iron meteorites by laser ablation ICPMS.
Anal. Chem. 71:939-946.
Carson, P. G.; Johnston, M. V.; Wexler, A. S. (1997) Laser desorption/ionization of ultra fine aerosol particles.
Rapid Commun. Mass Spectrom. 11: 993-996.
Castaldi, M. J.; Senkan, S. M. (1998) Real-time, ultrasensitive monitoring of air toxics by laser photoionization
time-of-flight mass spectroscopy. J. Air Waste Manage. Assoc. 48: 77-81.
Chandramouli, B.; Kamens, R. (2001) The photochemical formation and gas-particle partitioning of oxidation
products of decamethyl cyclopentasiloxane and decamethyl tetrasiloxane in the atmosphere. Atmos. Environ.
35: 87-95.
Chandramouli, B.; Jang, M.; Kamens, R. M. (2003) Gas-particle partitioning of semi-volatile organics on organic
aerosols using a predictive activity coefficient model: analysis of the effects of parameter choices on model
performance. Atmos. Environ. 37: 853-864.
Chang, M.; Kim, S.; Sioutas, C. (1999) Experimental studies on particle impaction and bounce: effects of substrate
design and material. Atmos. Environ. 33: 2313-2322.
Chang, M. C.; Sioutas, C.; Kim, S.; Gong, H., Jr.; Linn, W. S. (2000) Reduction of nitrate losses from filter and
impactor samplers by means of concentration enrichment. Atmos. Environ. 34: 85-98.
Chang, M.-C.; Geller, M. D.; Sioutas, C.; Fokkens, P. H. B.; Cassee, F. R. (2002) Development and evaluation of a
compact, highly efficient coarse particle concentrator for lexicological studies. Aerosol Sci. Technol.
36: 492-501.
Chein, H.; Lundgren, D. A. (1993) A virtual impactor with clean air core for the generation of aerosols with narrow
size distributions. Aerosol Sci. Technol. 18: 376-388.
Chen, B. T. (1993) Instrument calibration. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles,
techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 493-520.
Chen, B. T.; Yeh, H. C. (1987) An improved virtual impactor: design and performance. J. Aerosol Sci. 18: 203-214.
Chen, B. T.; Yeh, H. C.; Cheng, Y. S. (1985) A novel virtual impactor: calibration and use. J. Aerosol Sci.
16: 343-354.
Chen, B. T.; Yeh, H. C.; Cheng, Y. S. (1986) Performance of a modified virtual impactor. Aerosol Sci. Technol.
5: 369-376.
Cheng, Y.-S.; Yeh, H.-C. (1979) Particle bounce in cascade impactors. Environ. Sci. Technol. 13: 1392-1396.
Chow, J. C. (1995) Measurement methods to determine compliance with ambient air quality standards for suspended
particles. J. Air Waste Manage. Assoc. 45: 320-382.
Chow, J. C.; Watson, J. G. (1999) Ion chromatography in elemental analysis of airborne particles. In:
Landsberger, S.; Creatchman, M., eds. Elemental Analysis of Airborne Particles. Newark, NJ: Gordon and
Breach Science Publishers; pp. 97-137.
Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere;
April 1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
Chow, J. C.; Fujita, E. M.; Watson, J. G.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994a) Evaluation of filter-based
aerosol measurements during the 1987 Southern California Air Quality Study. Environ. Monit. Assess.
30: 49-80.
Chow, J. C.; Watson, J. G.; Fujita, E. M.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994b) Temporal and spatial
variations of PM2 5 and PM10 aerosol in the Southern California Air Quality Study. Atmos. Environ.
28: 2061-2080.
Chow, J. C.; Watson, J. G.; Lu, Z.; Lowenthal, D. H.; Frazier, C. A.; Solomon, P. A.; Thuillier, R. H.; Magliano, K.
(1996) Descriptive analysis of PM25 and PM10 at regionally representative locations during
SJVAQS/AUSPEX. In: Parrish, D.; Trainer, M.; Rao, S. T.; Solomon, P. A., eds. Air and Waste Management
Association international specialty conference on regional photochemical measurements and modeling, part 2;
November 1993; San Diego, CA. Atmos. Environ. 30: 2079-2112.
2-115
-------
Chow, J. C; Zielinska, B.; Watson, J. G.; Fujita, E. M; Richards, H. W.; Neff, W. D.; Dietrich, D.; Hering, S. V.
(1998) Northern Front Range Air Quality Study. Volume A: Ambient measurements. Fort Collins, CO:
Colorado State University, Cooperative Institute for Research in the Atmosphere. Available:
http://www.nfraqs.colostate.edu/dri.html [4 February, 2002].
Chow, J. C.; Watson, J. G.; Crow, D.; Lowenthal, D. H.; Merrifield, T. (2000) Comparison of IMPROVE and
NIOSH carbon measurements. Aerosol Sci. Technol. 34: 23-34.
Chow, J. C.; Watson, J. G.; Crow, D.; Lowenthal, D. H.; Merrifield, T. (2001) Comparison of IMPROVE and
NIOSH carbon measurements. Aerosol Sci. Technol. 34: 23-34.
Claiborn, C. S.; Finn, D.; Larson, T. V.; Koenig, J. Q. (2000) Windblown dust contributes to high PM2 5
concentrations. J. Air Waste Manage. Assoc. 50: 1440-1445.
Clark, W. E.; Whitby, K. T. (1975) Measurements of aerosols produced by the photochemical oxidation of SO2 in
air. J. Colloid Interface Sci. 51: 477-490.
Clarke, A. D. (1992) Atmospheric nuclei in the remote free-troposphere. J. Atmos. Chem. 14: 479-488.
Clarke, A. D.; Varner, J. L.; Eisele, F.; Mauldin, R. L.; Tanner, D.; Litchy, M. (1998) Particle production in the
remote marine atmosphere: cloud outflow and subsidence during ACE 1. J. Geophys. Res. A
103: 16,397-16,409.
Cobourn, W. G.; Husar, R. B. (1982) Diurnal and seasonal patterns of paniculate sulfur and sulfuric acid in
St. Louis, July 1977-June 1978. Atmos. Environ. 16: 1441-1450.
Cobourn, W. G.; Husar, R. B.; Husar, J. D. (1978) Continuous in situ monitoring of ambient paniculate sulfur using
flame photometry and thermal analysis. In: Husar, R. B.; Lodge, J. P., Jr.; Moore, D. J., eds. Sulfur in the
atmosphere: proceedings of the international symposium; September 1977; Dubrovnik, Yugoslavia. Atmos.
Environ. 12: 89-98.
Code of Federal Regulations. (200 la) Appendix J to Part 50-reference method for the determination of paniculate
matter as PM10 in the atmosphere. C. F. R. 40: pt. 50, app. J.
Code of Federal Regulations. (200 Ib) Ambient air monitoring reference and equivalent methods. C. F. R. 40: §53.
Code of Federal Regulations. (200 Ic) Appendix L to Part 50-reference method for the determination of fine
paniculate matter as PM2 5 in the atmosphere. C. F. R. 40: pt. 50, app. L.
Code of Federal Regulations. (2001d) Ambient air quality surveillance. C. F. R. 40: §58.
Code of Federal Regulations. (200 le) Appendix M to Part 50-reference method for the determination of paniculate
matter as PM10 in the atmosphere. C. F. R. 40: pt. 50, app. M.
Conner, W. D. (1966) An inertial-type particle separator for collecting large samples. J. Air Pollut. Control Assoc.
16: 35-38.
Conner, T. L.; Norris, G. A.; Landis, M. S.; Williams, R. W. (2001) Individual particle analysis of indoor, outdoor,
and community samples from the 1998 Baltimore particulate matter study. Atmos. Environ. 35: 3935-3946.
Gotham, W. E.; Bidleman, T. F. (1992) Laboratory investigations of the partitioning of organochlorine compounds
between the gas phase and atmospheric aerosols on glass fiber filters. Environ. Sci. Technol. 26: 469-478.
Countess, R. J. (1990) Interlaboratory analyses of carbonaceous aerosol samples. Aerosol Sci. Technol. 12: 114-121.
Courtney, W. J.; Shaw, R. W.; Dzubay, T. G. (1982) Precision and accuracy of a P gauge for aerosol mass
determinations. Environ. Sci. Technol. 16: 236-239.
Coutant, R. W.; Callahan, P. J.; Chuang, J. C.; Lewis, R. G. (1992) Efficiency of silicone-grease-coated denuders for
collection of polynuclear aromatic hydrocarbons. Atmos. Environ. Part A 26: 2831-2834.
Covert, D. S.; Kapustin, V. N.; Quinn, P. K.; Bates, T. S. (1992) New particle formation in the marine boundary
layer. J. Geophys. Res. [Atmos.] 97: 20,581-20,589.
Covert, D. S.; Wiedensohler, A.; Aalto, P.; Heintzenberg, J.; McMurry, P. H.; Leek, C. (1996) Aerosol number size
distributions from 3 to 500 nm diameter in the arctic marine boundary layer during summer and autumn.
Tellus 48B: 197-212.
Cui, W.; Machir, J.; Lewis , L.; Eatough, D. J.; Eatough, N. L. (1997) Fine particulate organic material at Meadview
during the project MOHAVE summer intensive study. J. Air Waste Manage. Assoc. 47: 357-369.
Cui, W.; Eatough, D. J.; Eatough, N. L. (1998) Fine particulate organic material in the Los Angeles Basin -1:
assessment of the high-volume Brigham Young University organic sampling system, BIG BOSS. J. Air
Waste Manage. Assoc. 48: 1024-1037.
Dagnall, R. M.; Thompson, K. C.; West, T. S. (1967) Molecular-emission in cool flames. Part I. the behaviour of
sulphur species in a hydrogen-nitrogen diffusion flame and in a shielded air-hydrogen flame. Analyst
92: 506-512.
Dams, R.; Robbins, J. A.; Rahn, K. A.; Winchester, J. W. (1970) Nondestructive neutron activation analysis of air
pollution particulates. Anal. Chem. 42: 861-867.
2-116
-------
Dane, A. D.; Timmermans, P. A. M; Van Sprang, H. A.; Buydens, L. M. C. (1996) A genetic algorithm for
model-free x-ray fluorescence analysis of thin film. Anal. Chem. 68: 2419-2425.
Davis, D. W.; Reynolds, R. L.; Tsou, G. C.; Zafonte, L. (1977) Filter attenuation corrections for x-ray fluorescence
analysis of atmospheric aerosols. Anal. Chem. 49: 1990-1993.
Davis, C. S.; Fellin, P.; Otson, R. (1987) A review of sampling methods for polyaromatic hydrocarbons in air. J. Air
Pollut. Control Assoc. 37: 1397-1408.
DeBoer, D. K. G.; Leenaers, A. J. G.; Van Den Hoogenhof, W. W. (1995) Glancing-incidence x-ray analysis of
thin-layered materials: a review. X-Ray Spectrom. 24: 91-102.
Ding, Y.; Koutrakis, P. (2000) Development of a dichotomous slit nozzle virtual impactor. J. Aerosol Sci.
31: 1421-1431.
Ding, Y.; Lee, M. L.; Eatough, D. J. (1998a) The determination of total nitrite and 7V-nitroso compounds in
atmospheric samples. Int. J. Environ. Anal. Chem. 69: 243-255.
Ding, Y.; Cui, W.; Lee, M. L.; Eatough, D. J. (1998b) Fine paniculate jV-nitroso and nitrite organic compounds in
the atmosphere. Appl. Occup. Environ. Hyg. 13: 432-438.
Ding, Y.; Pang, Y.; Eatough, D. J.; Eatough, N. L.; Tanner, R. L. (2002) High volume diffusion denuder sampler for
the routine monitoring of fine paniculate matter: II. field evaluation of the PC BOSS. Aerosol Sci. Technol.:
36: 383-396.
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.
(1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329: 1753-1759.
Durham, J. L.; Wilson, W. E.; Bailey, E. B. (1978) Application of an SO2-denuder for continuous measurement of
sulfur in submicrometric aerosols. Atmos. Environ. 12: 883-886.
Dzubay, T. G. (1980) Chemical element balance method applied to dichotomous sampler data. Ann. N. Y. Acad.
Sci. 338: 126-144.
Dzubay, T. G.; Stevens, R. K. (1975) Ambient air analysis with dichotomous sampler and X-ray fluorescence
spectrometer. Environ. Sci. Technol. 9: 663-668.
Dzubay, T. G.; Hines, L. E.; Stevens, R. K. (1976) Particle bounce errors in cascade impactors. Atmos. Environ.
10: 229-234.
Dzubay, T. G.; Stevens, R. K.; Peterson, C. M. (1977) Application of the dichotomous sampler to the
characterization of ambient aerosols. In: Dzubay, T. G., ed. X-ray fluorescence analysis of environmental
samples. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.
Easter, R. C.; Peters, L. K. (1994) Binary homogeneous nucleation: temperature and relative humidity fluctuations,
nonlinearity, and aspects of new particle production in the atmosphere. J. Appl. Meteorol. 33: 775-784.
Eatough, D. J. (1999) BOSS, the 5righam Young University Organic Sampling System: determination of paniculate
carbonaceous material using diffision denuder sampling technology. In: Lane, D., ed. Gas and particle phase
measurements of atmospheric organic compounds. New York, NY: Gordon and Breach Science Publishers.
(Advances in environmental, industrial and process control technologies: v. 2).
Eatough, D. J.; Wadsworth, A.; Eatough, D. A.; Crawford, J. W.; Hansen, L. D.; Lewis, E. A. (1993) A
multiple-system, multi-channel diffusion denuder sampler for the determination of fme-particulate organic
material in the atmosphere. Atmos. Environ. Part A 27: 1213-1219.
Eatough, D. J.; Tang, H.; Cui, W.; Machir, J. (1995) Determination of the size distribution and chemical composition
of fine paniculate semi-volatile organic material in urban environments using diffusion denuder technology.
In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution and human
mortality and morbidity, part II; January 1994; Irvine, CA. Inhalation Toxicol. 7: 691-710.
Eatough, D. J.; Eatough, D. A.; Lewis, L.; Lewis, E. A. (1996) Fine paniculate chemical composition and light
extinction at Canyonlands National Park using organic paniculate material concentrations obtained with a
multisystem, multichannel diffusion denuder sampler. J. Geophys. Res. [Atmos.] 101: 19,515-19,531.
Eatough, D. J.; Obeidi, F.; Pang, Y.; Ding, Y.; Eatough, N. L.; Wilson, W. E. (1999a) Integrated and real-time
diffusion denuder sampler for PM2 5. Atmos. Environ. 33: 2835-2844.
Eatough, D. J.; Pang, Y.; Eatough, N. L. (1999b) Determination of PM25 sulfate and nitrate with a PC-BOSS
designed for routine sampling for semi-volatile paniculate matter. J. Air Waste Manage. Assoc. 49(special
issue): PM69-75.
Eatough, D. J.; Eatough, N. L.; Obeidi, F.; Pang, Y.; Modey, W.; Long, R. (2001) Continuous determination of
PM25 mass, including semi-volatile species. Aerosol Sci. Technol. 34: 1-8.
Elvers, B.; Hawkins, S.; Schulz, G., eds. (1991) Peroxo compounds. Inorganic. In: Ullmann's Encyclopedia of
Industrial Chemistry. V A19. 5th ed. New York, NY: VCH Publishers; pp. 177-197.
2-117
-------
Fairchild, C. I.; Wheat, L. D. (1984) Calibration and evaluation of a real-time cascade impactor. Am. Ind. Hyg.
Assoc. 145:205-211.
Falconer, R. L.; Bidleman, T. F.; Gotham, W. E. (1995) Preferential sorption of non- and
mono-ortho-polychlorinated biphenyls to urban aerosols. Environ. Sci. Technol. 29: 1666-1673.
Fang, C. P.; McMurry, P. H.; Marple, V. A.; Rubow, K. L. (1991) Effect of flow-induced relative humidity changes
on size cuts for sulfuric acid droplets in the microorifice uniform deposit impactor (MOUDI). Aerosol Sci.
Technol. 14: 266-277.
Farwell, S. O.; Rasmussen, R. A. (1976) Limitations of the FPD and BCD in atmospheric analysis: a review. J.
Chromatogr. Sci. 14: 224-234.
Fassel, V. A.; Kniseley, R. N. (1974) Inductively coupled plasma-optical emission spectroscopy. Anal. Chem.
46: 1110A-1120A.
Federal Register. (1987) Revisions to the national ambient air quality standards for paniculate matter. F. R. (July 1)
52: 24,634-24,669.
Federal Register. (1997) National ambient air quality standards for paniculate matter; final rule. F. R. (July 18)
62: 38,652-38,752.
Federal Register. (1998) National ambient air quality standards for paniculate matter and revised requirements for
designation of reference and equivalent methods for PM2 5 and ambient air quality surveillance for paniculate
matter; correction. F. R. (February 17) 63: 7710-7718.
Fergenson, D. P.; Song, X.-H.; Ramadan, Z.; Allen, J. O.; Hughes, L. S.; Cass, G. R.; Hopke, P. K.; Prather, K. A.
(2001) Quantification of ATOFMS data by multivariate methods. Anal. Chem. 73: 3535-3541.
Fernandez, F. J. (1989) Atomic absorption spectroscopy. In: Lodge, J. P., Jr., ed. Methods of Air Sampling and
Analysis. 3rd ed. Chelsea, MI: Lewis Publishers, Inc.; pp. 83-89.
Finn, D.; Rumburg, B.; Claiborn, C.; Bamesberger, L.; Siems, W. F.; Koenig, J.; Larson, T.; Norris, G. (2001)
Sampling artifacts from the use of denuder tubes with glycerol based coatings in the measurement of
atmospheric paniculate matter. Environ. Sci. Technol. 35: 40-44.
Fitz, D. R. (1990) Reduction of the positive organic artifact on quartz filters. Aerosol. Sci. Technol. 12: 142-148.
Flocchini, R. G.; Cahill, T. A.; Shadoan, D. J.; Lange, S. J.; Eldred, R. A.; Feeney, P. J.; Wolfe, G. W.; Simmeroth,
D. C.; Suder, J. K. (1976) Monitoring California's aerosols by size and elemental composition. Environ. Sci.
Technol. 10: 76-82.
Forney, L. J.; Ravenhall, D. G.; Winn, D. S. (1978) Aerosol impactors: a study of a fluid jet impinging upon a void.
J. Appl. Phys. 49: 2339-2345.
Forney, L. J.; Ravenhall, D. G.; Lee, S. S. (1982) Experimental and theoretical study of a two-dimensional virtual
impactor. Environ. Sci. Technol. 16: 492-497.
Friedlander, S. K. (1970) The characterization of aerosols distributed with respect to size and chemical composition.
J. Aerosol Sci. 1:295-307.
Friedlander, S. K. (2000). Smoke, dust and haze: fundamentals of aerosol dynamics. 2nd ed. New York, NY: Oxford
University Press.
Friedlander, S. K.; Yeh, E. K. (1998) The submicron atmospheric aerosol as a carrier of reactive chemical species:
case of peroxides. Appl. Occup. Environ. Hyg. 13: 416-420.
Fuchs, N. A. (1964) The mechanics of aerosols. New York, NY: Pergamon Press.
Fuchs, N. A. (1989) The mechanics of aerosols. New York, NY: Dover Publications.
Fung, K. (1990) Particulate carbon speciationby MnO2 oxidation. Aerosol Sci. Technol. 12: 122-127.
Fung, K.; Chow, J. C.; Watson, J. G. (2002) Evaluation of OC/EC speciationby thermal manganese dioxide
oxidation and the IMPROVE method. J. Air Waste Manage. Assoc. 52: 1333-1341.
Gallorini, M. (1995) Trace element monitoring in atmospheric pollution processes by neutron activation analysis.
Microchem. J. 51: 127-137.
Card, E.; Mayer, J. E.; Morrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. (1997) Real-time analysis of
individual atmospheric aerosol particles: design and performance of a portable ATOFMS. Anal. Chem.
69: 4083-4091.
Gauderman, W. J.; McConnell, R.; Gilliland, F.; London, S.; Thomas, D.; Avol, E.; Vora, H.; Berhane, K.;
Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and lung
function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
Ge, Z.; Wexler, A. S.; Johnston, M. V. (1998) Laser desorption/ionization of single ultrafine multicomponent
aerosols. Environ. Sci. Technol. 32: 3218-3223.
Geller, M. D.; Kim, S.; Misra, C.; Sioutas, C.; Olson, B. A.; Marple, V. A. (2002) Methodology for measuring
size-dependent chemical composition of ultrafine particles. Aerosol Sci. Technol. 36: 748-763.
2-118
-------
Goss, K.-U.; Eisenreich, S. J. (1997) Sorption of volatile organic compounds to particles from a combustion source
at different temperatures and relative humidities. Atmos. Environ. 31: 2827-2834.
Gray, A. L.; Williams, J. G. (1987a) Oxide and doubly charged ion response of a commercial inductively coupled
plasma mass spectrometry instrument. J. Anal. At. Spectrom. 2: 81-82.
Gray, A. L.; Williams, J. G. (1987b) System optimisation and the effect on polyatomic, oxide and doubly charged
ion response of a commercial inductively coupled plasma mass spectrometry instrument. J. Anal. At.
Spectrom. 2: 599-606.
Grosjean, D. (1983) Polycyclic aromatic hydrocarbons in Los Angeles air from samples collected on Teflon, glass
and quartz filters. Atmos. Environ. 17: 2565-2573.
Grosjean, D.; Parmar, S. S. (1990) Interferences from aldehydes and peroxyacetyl nitrate when sampling urban air
organic acids on alkaline traps. Environ. Sci. Technol. 24: 1021-1026.
Gundel, L. A.; Lane, D. A. (1999) Sorbent-coated diffusion denuders for direct measurement of gas/particle
partitioning by semi-volatile organic compounds. In: Lane, D. A., ed. Gas and particle phase measurements of
atmospheric organic compounds. Amsterdam, The Netherlands: Gordon and Breach Science Publishers.
(Advances in environmental, industrial and process control technologies: v. 2).
Gundel, L. A.; Lee, V. C.; Mahanama, K. R. R.; Stevens, R. K.; Daisey, J. M. (1995) Direct determination of the
phase distributions of semi-volatile polycyclic aromatic hydrocarbons using annular denuders. Atmos.
Environ. 29: 1719-1733.
Haller, L.; Claiborn, C.; Larson, T.; Koenig, J.; Norris, G.; Edgar, R. (1999) Airborne paniculate matter size
distributions in an arid urban area. J. Air Waste Manage. Assoc. 49: 161-168.
Hanninen, O. O.; Koistinen, K. J.; Kousa, A.; Keski-Karhu, J.; Jantunen, M. J. (2002) Quantitative analysis of
environmental factors in differential weighing of blank teflon filters. J. Air Waste Manage. Assoc.
52: 134-139.
Hansen, A. D. A.; McMurry, P. H. (1990) An intercomparison of measurements of aerosol elemental carbon during
the 1986 carbonaceous species method comparison study. J. Air Waste Manage. Assoc. 40: 894-895.
Hansen, A. D. A.; Rosen, H.; Novakov, T. (1984) The aethalometer - an instrument for the real-time measurement of
optical absorption by aerosol particles. Sci. Total Environ. 36: 191-196.
Harman, J. N. (1989) Inductively coupled plasma emission spectroscopy. In: Lodge, J. P., Jr., ed. Methods of air
sampling and analysis. 3rd ed. Chelsea, MI: Lewis Publishers, Inc.; pp. 89-92.
Haupt, O.; Klane, B.; Schaefer, C.; Dannecker, W. (1995) Preparation of quartz fibre filter standards for x-ray
fluorescence analysis of aerosol samples. X-Ray Spectrom. 24: 267-275.
He, K.; Yang, F.; Ma, Y.; Zhang, Q.; Yao, X.; Chan, C. K.; Cadle, S.; Chan, T.; Mulawa, P. (2001) The
characteristics of PM25 in Beijing, China. Atmos. Environ. 35: 4959-4970.
Hegg, D. A.; Hobbs, P. V. (1982) Measurements of sulfate production in natural clouds. Atmos. Environ.
16: 2663-2668.
Hegg, D. A.; Hobbs, P. V. (1986) Sulfate and nitrate chemistry in cumuliform clouds. Atmos. Environ. 20: 901-909.
Hegg, D. A.; Radke, L. F.; Hobbs, P. V. (1990) Particle production associated with marine clouds. J. Geophys. Res.
[Atmos.] 95: 13,917-13,926.
Hegg, D. A.; Radke, L. F.; Hobbs, P. V. (1991) Measurements of Aitken nuclei and cloud condensation nuclei in the
marine atmosphere and their relation to the DMS-cloud-climate hypothesis. J. Geophys. Res. [Atmos.]
96: 18,727-18,733.
Hering, S.; Cass, G. (1999) The magnitude of bias in the measurement of PM25 arising from volatilization of
paniculate nitrate from Teflon filters. J. Air Waste Manage. Assoc. 49: 725-733.
Hering, S. V.; Friedlander, S. K. (1982) Ortgins of aerosol sulfur size distributions in the Los Angeles basin.
Atmos. Environ. 16: 2647-2656.
Hering, S. V.; Appel, B. R.; Cheng, W.; Salaymeh, F.; Cadle, S. H.; Mulawa, P. A.; Cahill, T. A.; Eldred, R. A.;
Surovik, M.; Fitz, D.; Howes, J. E.; Knapp, K. T.; Stockburger, L.; Turpin, B. J.; Huntzicker, J. J.; Zhang,
X.-Q.; McMurry, P. H. (1990) Comparison of sampling methods for carbonaceous aerosols in ambient air.
Aerosol Sci. Technol. 12: 200-213.
Hinds, W. C. (1999) Aerosol technology: properties, behavior, and measurement of airborne particles. 2nd ed.
New York, NY: John Wiley & Sons.
Hinz, K.-P.; Kaufmann, R.; Spengler, B. (1994) Laser-induced mass analysis of single particles in the airborne state.
Anal. Chem. 66: 2071-2076.
Hitzenberger, R.; Dusek, U.; Berner, A. (1996) Black carbon measurements using an integrating sphere. J. Geophys.
Res. [Atmos.] 101: 19,601-19,606.
2-119
-------
Hitzenberger, R.; Berner, A.; Dusek, U.; Alabashi, R. (1997) Humidity-dependent growth of size-segregated aerosol
samples. Aerosol Sci. Technol. 27: 116-130.
Hitzenberger, R.; Jennings, S. G.; Larson, S. M; Dillner, A.; Cachier, H.; Galambos, Z.; Roue, A.; Spain, T. G.
(1999) Intercomparison of measurement methods for black carbon aerosols. Atmos. Environ. 33: 2823-2833.
Holynska, B.; Ptasiriski, J.; Maenhaut, W.; Annegarn, H. J. (1997) Energy-dispersive x-ray fluorescence
spectrometer with capillary optics for the chemical analysis of atmospheric aerosols with high time resolution.
J. Aerosol Sci. 28: 1455-1463.
Hoppel, W. A.; Frick, G. M.; Fitzgerald, J. W.; Larson, R. E. (1994) Marine boundary layer measurements of new
particle formation and the effects nonprecipitating clouds have on aerosol size distribution. J. Geophys. Res.
[Atmos.] 99: 14,443-14,459.
Houck, J. E.; Chow, J. C.; Watson, J. G.; Simons, C. A.; Pritchett, L. C.; Goulet, J. M.; Frazier, C. A. (1989)
Determination of particle size distribution and chemical composition of paniculate matter from selected
sources in California: volume I and executive summary (final report). Sacramento, CA: California Air
Resources Board.
Hounam, R. F.; Sherwood, R. J. (1965) The cascade centripeter: a device for determining the concentration and size
distribution of aerosols. Am. Ind. Hyg. Assoc. J. 26: 122-131.
Hung, H.-F.; Wang, C.-S. (2001) Experimental determination of reactive oxygen species in Taipei aerosols.
J. Aerosol Sci. 32: 1201-1211.
Huntzicker, J. J.; Hoffman, R. S.; Ling, C.-S. (1978) Continuous measurement and speciation of sulfur-containing
aerosols by flame photometry. Atmos. Environ. 12: 83-88.
Huntzicker, J. J.; Johnson, R. L.; Shah, J. J.; Gary, R. A. (1982) Analysis of organic and elemental carbon in ambient
aerosols by a thermal-optical method. In: Wolff, G. T.; Klimisch, R. L., eds. Paniculate carbon - atmospheric
life cycle: proceedings of an international symposium; October 1980; Warren, MI. New York, NY: Plenum
Press; pp. 79-88.
Husar, R. B.; Tratt, D. M.; Schichtel, B. A.; Falke, S. R.; Li, F.; Jaffe, D.; Gasso, S.; Gill, T.; Laulainen, N. S.;
Lu, F.; Reheis, M. C.; Chun, Y.; Westphal, D.; Holben, B. N.; Gueymard, C.; McKendry, I.; Kuring, N.;
Feldman, G. C.; McClain, C.; Frouin, R. J.; Merrill, J.; DuBois, D.; Vignola, F.; Murayama, T.; Nickovic, S.;
Wilson, W. E.; Sassen, K.; Sugimoto, N.; Malm, W. C. (2001) The Asian dust events of April 1998.
J. Geophys. Res. 106: 18,317-18,330.
Jackson, K. W.; Mahmood, T. M. (1994) Atomic absorption, atomic emission, and flame emission spectrometry.
Anal. Chem. 66: 252R-279R.
Jaklevic, J. M.; Loo, B. W.; Goulding, F. S. (1977) Photon-induced X-ray fluorescence analysis using
energy-dispersive detector and dichotomous sampler. In: Dzubay, T. G., ed. X-ray fluorescence analysis of
environmental samples. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.; pp. 3-18.
Jaklevic, J. M.; Gatti, R. C.; Goulding, F. S.; Loo, B. W. (1981a) A p-gauge method applied to aerosol samples.
Environ. Sci. Technol. 15: 680-686.
Jaklevic, J. M.; Loo, B. W.; Fujita, T. Y. (1981b) Automatic paniculate sulfur measurements with a dichotomous
sampler and on-line x-ray fluorescence analysis. Environ. Sci. Technol. 15: 687-690.
Jang, M.; Kamens, R. M. (1998) A thermodynamic approach for modeling partitioning of semivolatile organic
compounds on atmospheric paniculate matter: humidity effects. Environ. Sci. Technol. 32: 1237-1243.
Jang, M.; Kamens, R. M. (1999) Exploratory model approach for adsorption partitioning of SOCs on inorganic
particles. Environ. Sci. Technol.
Jang, M.; Kamens, R. M. (2001) Atmospheric secondary aerosol formation by heterogeneous reactions of aldehydes
in the presence of a sulfuric acid aerosol catalyst. Environ. Sci. Technol. 35: 4758-4766.
Jang, M.; Kamens, R. M.; Leach, K. B.; Strommen, M. R. (1997) A thermodynamic approach using group
contribution methods to model the partitioning of semivolatile organic compounds on atmospheric paniculate
matter. Environ. Sci. Technol. 31: 2805-2811.
Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning.
Atmos. Environ. 30: 3825-3835.
Johansson, T. B.; Askelsson, R.; Johansson, S. A. E. (1970) X-ray analysis: elemental trace analysis at the 10"12g
level. Nucl. Instrum. Methods 84: 141-143.
John, W.; Sethi, V. (1993) Threshold for resuspension by particle impaction. Aerosol Sci. Technol. 19: 69-79.
John, W.; Wall, S. M.; Ondo, J. L. (1988) A new method for nitric acid and nitrate aerosol measurement using the
dichotomous sampler. Atmos. Environ. 22: 1627-1635.
2-120
-------
John, W.; Wall, S. M; Ondo, J. L.; Winklmayr, W. (1990) Modes in the size distributions of atmospheric inorganic
aerosol. Atmos. Environ. Part A 24: 2349-2359.
Johnson, M V.; Wexler, A. S. (1995) MS of individual aerosol particles. Anal. Chem. 67: 721A-726A.
Johnson, R. L.; Shah, J. J.; Huntzicker, J. J. (1980) Analysis of organic, elemental, and carbonate carbon in ambient
aerosols. In: Verner, S. S. Sampling and analysis of toxic organics in the atmosphere: proceedings of a
symposium; August, 1979; Boulder, CO. Philadelphia, PA: American Society for Testing and Materials,
pp. 111-119. [ASTM Special Technical Publication 721].
Kamens, R. M.; Coe, D. L. (1997) A large gas-phase stripping device to investigate rates of PAH evaporation from
airborne diesel soot particles. Environ. Sci. Technol. 31: 1830-1833.
Kamens, R. M.; Jaoui, M. (2001) Modeling aerosol formation from a-pinene + NOX in the presence of natural
sunlight using gas phase kinetics and gas-particle partitioning theory. Environ. Sci. Technol. 35: 1394-1405.
Kamens, R.; Odum, J.; Fan, Z.-H. (1995) Some observations on times to equilibrium for semivolatile poly cyclic
aromatic hydrocarbons. Environ. Sci. Technol. 29: 43-50.
Kamens, R.; Jang, M.; Chien, C.-J.; Leach, K. (1999) Aerosol formation from the reaction of a-pinene and ozone
using a gas-phase kinetics-aerosol partitioning model. Environ. Sci. Technol. 33: 1430-1438.
Kao, A. S.; Friedlander, S. K. (1995) Temporal variations of paniculate air pollution: a marker for free radical
dosage and adverse health effects? Inhalation Toxicol. 7: 149-156.
Kavouras, I. G.; Koutrakis, P. (2001) Use of polyurethane foam as the impaction substrate/collection medium in
conventional inertial impactors. Aerosol Sci. Technol. 34: 46-56.
Keeler, G. J.; Pirrone, N. (1996) Atmospheric transport and deposition of trace elements to Lake Erie from urban
areas. Water Sci. Technol. 33: 259-265.
Keeler, G. J.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Pierson, W. R. (1988) Effect of ambient humidity
on dichotomous sampler coarse/fine ratios. Atmos. Environ. 22: 1715-1720.
Kegler, S. R.; Wilson, W. E.; Marcus, A. H. (2001) PM1, intermodal (PM25_l) mass, and the soil component of PM25
in Phoenix, AZ, 1995-96. Aerosol Sci. Technol. 35: 914-920.
Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
segregated mass concentration data and ultrafine particle number concentrations in urban areas. Atmos.
Environ. 33:2907-2913.
Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (2000) Size distribution and sources of
aerosol in Launceston, Australia, during winter 1997. J. Air Waste Manage. Assoc. 50: 418-427.
Khlystov, A.; Wyers, G. P.; Slanina, J. (1995) The steam-jet aerosol collector. Atmos. Environ. 29: 2229-2234.
Kim, Y. J.; Boatman, J. F.; Gunter, R. L.; Wellman, D. L.; Wilkison, S. W. (1993) Vertical distribution of
atmospheric aerosol size distribution over south-central New Mexico. Atmos. Environ. Part A 27: 1351-1362.
Kim, B. M.; Lester, J.; Tisopulos, L.; Zeldin, M. D. (1999) Nitrate artifacts during PM25 sampling in the South Coast
Air Basin of California. J. Air Waste Manage. Assoc. 49(special issue): PM142-153.
Kim, B. M.; Cassmassi, J.; Hogo, H.; Zeldin, M. D. (2001a) Positive organic carbon artifacts on filter medium
during PM2 5 sampling in the South Coast Air Basin. Aerosol Sci. Technol. 34: 35-41.
Kim, S.; Jaques, P. A.; Chang, M. C.; Froines, J. R.; Sioutas, C. (200Ib) Versatile aerosol concentration enrichment
system (VACES) for simultaneous in vivo and in vitro evaluation of toxic effects of ultrafine, fine and coarse
ambient particles Part I: development and laboratory characterization. J. Aerosol Sci. 32: 1281-1297.
Kim, S.; Jaques, P. A.; Chang, M. C.; Barone, T.; Xiong, C.; Friedlander, S. K.; Sioutas, C. (2001c) Versatile
aerosol concentration enrichment system (VACES) for simultaneous in vivo and in vitro evaluation of toxic
effects of ultrafine, fine and coarse ambient particles Part II: field evaluation. J. Aerosol Sci. 32: 1299-1314.
Kirchstetter, T. W.; Corrigan, C. E.; Novakov, T. (2001) Laboratory and field investigation of the adsorption of
gaseous organic compounds onto quartz filters. Atmos. Environ. 35: 1663-1671.
Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
Kittelson, D. B.; McKenzie, R.; Vermeersch, M.; Dorman, F.; Pui, D.; Linne, M.; Liu, B.; Whitby, K. (1978) Total
sulfur aerosol concentration with an electrostatically pulsed flame photometric detector system. Atmos.
Environ. 12: 105-111.
Klouda, G. A.; Klinedinst, D. B.; Steel, E. B.; Benner, B. A., Jr.; Parish, H. J. (1996) Exploring a method to produce
an urban dust particle filter standard. J. Aerosol Sci. 27(suppl. 1): S351-S352.
Koistinen, K. J.; Kousa, A.; Tenhola, V.; Hanninen, O.; Jantunen, M. J.; Oglesby, L.; Kuenzli, N.; Georgoulis, L.
(1999) Fine particle (PM2 5) measurement methodology, quality assurance procedures and pilot results of the
EXPOLIS study. J. Air Waste Manage. Assoc. 49: 1212-1220.
Korhonen, P.; Kulmala, M.; Laaksonen, A.; Viisanen, Y.; McGraw, R.; Seinfeld, J. H. (1999) Ternary nucleation of
H2SO4, NH3, andH2O in the atmosphere. J. Geophys. Res. 104: 26,349-26,353.
2-121
-------
Koutrakis, P.; Kelly, B. P. (1993) Equilibrium size of atmospheric aerosol sulfates as a function of particle acidity
and ambient relative humidity. J. Geophys. Res. [Atmos.] 98: 7141-7147.
Koutrakis, P.; Wolfson, J. M; Slater, J. L.; Brauer, M; Spengler, J. D.; Stevens, R. K.; Stone, C. L. (1988a)
Evaluation of an annular denuder/filter pack system to collect acidic aerosols and gases. Environ. Sci.
Technol. 22: 1463-1468.
Koutrakis, P.; Wolfson, J. M.; Spengler, J. D. (1988b) An improved method for measuring aerosol strong acidity:
results from a nine-month study in St. Louis, Missouri and Kingston, Tennessee. Atmos. Environ.
22: 157-162.
Koutrakis, P.; Wolfson, J. M.; Spengler, J. D.; Stern, B.; Franklin, C. A. (1989) Equilibrium size of atmospheric
aerosol sulfates as a function of the relative humidity. J. Geophys. Res. [Atmos.] 94: 6442-6448.
Koutrakis, P.; Thompson, K. M.; Wolfson, J. M.; Spengler, J. D.; Keeler, G. J.; Slater, J. L. (1992) Determination of
aerosol strong acidity losses due to interactions of collected particles: results from laboratory and field
studies. Atmos. Environ. Part A 26: 987-995.
Koutrakis, P.; Sioutas, C.; Ferguson, S. T.; Wolfson, J. M.; Mulik, J. D.; Burton, R. M. (1993) Development and
evaluation of a glass honeycomb denuder/filter pack system to collect atmospheric gases and particles.
Environ. Sci. Technol. 27: 2497-2501.
Koutrakis, P.; Wang, P.-Y.; Wolfson, J. M.; Sioutas, C., inventors. (1996) Method and apparatus to measure
paniculate matter in gas. U.S. patent 5,571,945. November 5. Available: www.patents.ibm.com
[1999, November 24].
Krieger, M. S.; Hites, R. A. (1992) Diffusion denuder for the collection of semivolatile organic compounds. Environ.
Sci. Technol. 26: 1551-1555.
Kulmala, M.; Toivonen, A.; Makela, J. M.; Laaksonen, A. (1998) Analysis of the growth of nucleation mode
particles observed in Boreal forest. Tellus 50B: 449-462.
Kulmala, M.; Pirjola, L.; Makela, J. M. (2000) Stable sulphate clusters as a source of new atmospheric particles.
Nature (London, U. K.) 404: 66-69.
Kusko, B. H.; Cahill, T. A.; Eldred, R. A.; Matsuda, Y.; Miyake, H. (1989) Nondestructive analysis of total
nonvolatile carbon by Forward Alpha Scattering Technique (FAST). Aerosol Sci. Technol. 10: 390-396.
Lammel, G.; Novakov, T. (1995) Water nucleation properties of carbon black and diesel soot particles. Atmos.
Environ. 29: 813-823.
Landsberger, S. (1988) Improved methodology for the determination of the seven elemental tracer long-distance
pollution signatures using thermal and epithermal neutron activation analysis. Anal. Chern. 60: 1842-1845.
Landsberger, S.; Wu, D. (1993) Improvement of analytical sensitivities for the determination of antimony, arsenic,
cadmium, indium, iodine, molybdenum, silicon and uranium in airborne paniculate matter by epithermal
neutron activation analysis. J. Radioanal. Nucl. Chem. 167: 219-225.
Landsberger, S.; Wu, D.; Vermette, S. J.; Cizek, W. (1997) Intercomparison of IAEA airborne paniculate matter
reference material. J. Radioanal. Nucl. Chem. 215: 117-127.
Lane, D. A.; Johnson, N. D.; Barton, S. C.; Thomas, G. H. S.; Schroeder, W. H. (1988) Development and evaluation
of a novel gas and particle sampler for semivolatile chlorinated organic compounds in ambient air. Environ.
Sci. Technol. 22: 941-947.
Lavanchy, V. M. H.; Gaggeler, H. W.; Nyeki, S.; Baltensperger, U. (1999) Elemental carbon (EC) and black carbon
(BC) measurements with a thermal method and an aethalometer at the high-alpine research station
Jungfraujoch. Atmos. Environ. 33: 2759-2769.
Lawless, P. A.; Rodes, C. E.; Evans, G.; Sheldon, L.; Creason, J. (2001) Aerosol concentrations during the 1999
Fresno Exposure Studies as functions of size, season and meteorology. Aerosol Sci. Technol. 34: 66-74.
Lawrence, J. E.; Koutrakis, P. (1994) Measurement of atmospheric formic and acetic acids: methods evaluation and
results from field studies. Environ. Sci. Technol. 28: 957-964.
Lawrence, J.; Koutrakis, P. (1996a) Measurement and speciation of gas and paniculate phase organic acidity in an
urban environment: 1. analytical. J. Geophys. Res. [Atmos.] 101: 9159-9169.
Lawrence, J.; Koutrakis, P. (1996b) Measurement and speciation of gas and paniculate phase organic acidity in an
urban environment: 2. speciation. J. Geophys. Res. [Atmos.] 101: 9171-9184.
Lawson, D. R. (1990) The Southern California Air Quality Study. J. Air Waste Manage. Assoc. 40: 156-165.
Lazrus, A. L.; Haagenson, P. L.; Kok, G. L.; Huebert, B. J.; Kreitzberg, C. W.; Likens, G. E.; Mohnen, V. A.;
Wilson, W. E.; Winchester, J. W. (1983) Acidity in air and water in a case of warm frontal precipitation.
Atmos. Environ. 17: 581-591.
Leach, K. B.; Kamens, R. M.; Strommen, M. R.; Jang, M. (1999) Partitioning of semivolatile organic compounds in
the presence of a secondary organic aerosol in a controlled atmosphere. J. Atmos. Chem. 33: 241-264.
2-122
-------
Lee, W.-M. G.; Shih, P.-M.; Wang, C.-S. (1997) The influence of relative humidity on the size of atmospheric
aerosol. J. Environ. Sci. Health Part A: Environ. Sci. Eng. Toxic Hazard. Subst. Control 32: 1085-1097.
Lewis, C. W.; Macias, E. S. (1980) Composition of size-fractionated aerosol in Charleston, West Virginia. Atmos.
Environ. 14: 185-194.
Lewtas, J.; Pang, Y.; Booth, D.; Reimer, S.; Eatough, D. J.; Gundel, L. A. (2001) Comparison of sampling methods
for semi-volatile organic carbon associated with PM25. Aerosol Sci. Technol. 34: 9-22.
Li, W.; Montassier, N.; Hopke, P. K. (1992) A system to measure the hygroscopicity of aerosol particles. Aerosol
Sci. Technol. 17: 25-35.
Liang, Z.; Chan, C. K. (1997) A fast technique for measuring water activity of atmospheric aerosols. Aerosol Sci.
Technol. 26: 255-268.
Liang, C.; Pankow, J. F.; Odum, J. R.; Seinfeld, J. H. (1997) Gas/particle partitioning of semivolatile organic
compounds to model inorganic, organic, and ambient smog aerosols. Environ. Sci. Technol. 31: 3086-3092.
Lin, J. J.; Noll, K. E.; Holsen, T. M. (1994) Dry deposition velocities as a function of particle size in the ambient
atmosphere. Aerosol Sci. Technol. 20: 239-252.
Lindskog, A.; Brorstrom-Lunden, E.; Alfheim, I; Hagen, I. (1987) Chemical transformation of PAH on airborne
particles by exposure to NO2 during sampling: a comparison between two filter media. Sci. Total Environ.
61:51-57.
Liousse, C.; Cachier, H.; Jennings, S. G. (1993) Optical and thermal measurements of black carbon aerosol content
in different environments: variation of the specific attenuation cross-section, sigma a. Atmos. Environ. Part A
27: 1203-1211.
Long, R. W.; Smith, R.; Smith, S.; Eatough, N. L.; Mangelson, N. F.; Eatough, D. J.; Pope, C. A., Ill; Wilson, W. E.
(2002) Sources of fine paniculate material along the Wasatch Front. Ener. Fuels 16: 282-293.
Loo, B. W.; Cork, C. P. (1988) Development of high efficiency virtual impactors. Aerosol Sci. Technol. 9: 167-176.
Loo, B. W.; Jaklevic, J. M.; Goulding, F. S. (1976) Dichotomous virtual impactors for large scale monitoring of
airborne paniculate matter. In: Liu, B. Y. H., ed. Fine particles: aerosol generation, measurement, sampling,
and analysis. New York, NY: Academic Press; pp. 311-350.
Loo, B. W.; French, W. R.; Gatti, R. C.; Goulding, F. S.; Jaklevic, J. M.; Llacer, J.; Thompson, A. C. (1978)
Large-scale measurement of airborne paniculate sulfur. Atmos. Environ. 12: 759-771.
Lundgren, D. A.; Burton, R. M. (1995) Effect of particle size distribution on the cut point between fine and coarse
ambient mass fractions. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate
air pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 131-148.
Lundgren, D. A.; Hausknecht, B. J. (1982) Ambient aerosol size distribution determination using a mobile wide
range aerosol classifier. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Environmental Monitoring Systems Laboratory; EPA contract no. CR-808606-01-1.
Lundgren, D. A.; Hausknecht, B. J.; Burton, R. M. (1984) Large particle size distribution in five U.S. cities and the
effect on a new ambient paniculate matter standard (PM10). Aerosol Sci. Technol. 7: 467-473.
Lynch, A. J.; McQuaker, N. R.; Brown, D. F. (1980) ICP/AES analysis and the composition of airborne and soil
materials in the vicinity of a lead/zinc smelter complex. J. Air Pollut. Control Assoc. 30: 257-260.
Mader, B. T.; Flagan, R. C.; Seinfeld, J. H. (2001) Sampling atmospheric carbonaceous aerosols using a particle trap
impactor/denuder sampler. Environ. Sci. Technol. 35: 4857-4867.
Mahanama, K. R. R.; Gundel, L. A.; Daisey, J. M. (1994) Selective fluorescence detection of poly cyclic aromatic
hydrocarbons in environmental tobacco smoke and other airborne particles. Int. J. Environ. Anal. Chem.
56: 289-309.
Makela, J. M.; Aalto, P.; Jokinen, V.; Pohja, T.; Nissinen, A.; Palmroth, S.; Markkanen, T.; Seitsonen, K.;
Lihavainen, H.; Kulmala, M. (1997) Observations of ultrafme aerosol particle formation and growth in
boreal forest. Geophys. Res. Lett. 24: 1219-1222.
Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
Mamane, Y.; Willis, R.; Conner, T. (2001) Evaluation of computer-controlled scanning electron microscopy applied
to an ambient urban aerosol sample. Aerosol Sci. Technol. 34: 97-107.
Marple, V. A.; Chien, C. M. (1980) Virtual impactors: a theoretical study. Environ. Sci. Technol. 14: 976-985.
Marple, V. A.; Olson, B. A. (1995) A high volume PM10/PM2J/PMLO trichotomous sampler. In: Paniculate matter:
health and regulatory issues: proceedings of an international specialty conference; April; Pittsburgh, PA.
Pittsburgh, PA: Air & Waste Management Association; pp. 237-261. (A&WMA publication VTP-49).
Marple, V. A.; Liu, B. Y. H.; Burton, R. M. (1990) High-volume impactorfor sampling fine and coarse particles.
J. Air Waste Manage. Assoc. 40: 762-767.
2-123
-------
Marple, V. A.; Rubow, K. L.; Behm, S. M. (1991) A microorifice uniform deposit impactor (MOUDI): description,
calibration, and use. Aerosol Sci. Technol. 14: 434-446.
Martin, L. R. (1994) Aqueous sulfur(IV) oxidation revisited. In: Nriagu, J. O.; Simmons, M. S., eds. Environmental
oxidants. New York, NY: John Wiley & Sons, Inc.; pp. 221-268. (Nriagu, J. O., ed. Advances in
environmental science and technology: v. 28).
Masuda, H.; Nakasita, S. (1988) Classification performance of a rectangular jet virtual impactor-effects of nozzle
width ratio of collection nozzle to acceleration nozzle. J. Aerosol Sci. 19: 243-252.
Masuda, H.; Hochrainer, D.; Stober, W. (1979) An improved virtual impactor for particle classification and
generation of test aerosols with narrow size distributions. J. Aerosol Sci. 10: 275-287.
McGovern, F. M. (1999) An analysis of condensation nuclei levels at Mace Head, Ireland. Atmos. Environ.
33: 1711-1723.
McGovern, F. M.; Jennings, S. G.; O'Connor, T. C.; Simmonds, P. G. (1996) Aerosol and trace gas measurements
during the Mace Head experiment. Atmos. Environ. 30: 3891-3902.
Mclnnes, L. M.; Quinn, P. K.; Covert, D. S.; Anderson, T. L. (1996) Gravimetric analysis, ionic composition, and
associated water mass of the marine aerosol. Atmos. Environ. 30: 869-884.
McMurry, P. H. (2000) A review of atmospheric aerosol measurements. Atmos. Environ. 34: 1959-1999.
McMurry, P. H.; Friedlander, S. K. (1979) New particle formation in the presence of an aerosol. Atmos. Environ.
13: 1635-1651.
McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
aerosols. Atmos. Environ. 23: 497-507.
McMurry, P. H.; Wilson, J. C. (1982) Growth laws for the formation of secondary ambient aerosols: implications for
chemical conversion mechanisms. Atmos. Environ. 16: 121-134.
McMurry, P. H.; Wilson, J. C. (1983) Droplet phase (heterogeneous) and gas phase (homogeneous) contributions to
secondary ambient aerosol formation as functions of relative humidity. J. Geophys. Res. 88: 5101-5108.
McMurry, P. H.; Rader, D. J.; Stith, J. L. (1981) Studies of aerosol formation in power plant plumes--!, growth laws
for secondary aerosols in power plant plumes: implications for chemical conversion mechanisms. In: White,
W. H.; Moore, D. J.; Lodge, J. P., Jr., eds. Plumes and visibility: measurements and model components:
proceedings of the symposium; November 1980; Grand Canyon National Park, AZ. Atmos. Environ.
15:2315-2327.
McMurry, P. H.; Woo, K. S.; Weber, R.; Chen, D.-R.; Pui, D. Y. H. (2000) Size distributions of 3 to 10 nm
atmospheric particles: implications for nucleation mechanisms. Philos. Trans. R. Soc. London Ser. A
358: 2625-2642.
McQuaker, N. R.; Kluckner, P. D.; Chang, G. N. (1979) Calibration of an inductively coupled plasma-atomic
emission spectrometer for the analysis of environmental materials. Anal. Chem. 51: 888-895.
Meyer, A. S., Jr.; Boyd, C. M. (1959) Determination of water by titration with coulometrically generated
Karl Fischer reagent. Anal. Chem. 31: 215-219.
Meyer, M. B.; Rupprecht, E. G. (1996) Paniculate matter sampling methods: the importance of standardization.
J. Aerosol. Sci. 27(suppl. 1): S349-S350.
Meyer, M. B.; Rupprecht, E.; Patashnick, H. (1995) Considerations for the sampling and measurement of ambient
paniculate matter. In: Paniculate matter: health and regulatory issues: proceedings of an international
specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association;
pp. 204-210. (A&WMA publication VIP-49).
Meyer, M. B.; Patashnick, H.; Ambs, J. L.; Rupprecht, E. (2000) Development of a sample equilibration system for
the TEOM continuous PM monitor. J. Air Waste Manage. Assoc. 50: 1345-1349.
Middlebrook, A. M.; Murphy, D. M.; Thomson, D. S. (1998) Observations of organic material in individual marine
particles at Cape Grim during the first aerosol characterization experiment (ACE 1). J. Geophys. Res.
103: 16475-16483.
Middlebrook, A. M.; Murphy, D. M.; Lee, S.-H.; Thomson, D. S.; Prather, K. A.; Wenzel, R. J.; Liu, D.-Y.; Phares,
D. J.; Rhoads, K. P.; Wexler, A. S.; Johnston, M. V.; Jimenez, J. L.; Jayne, J. T.; Worsnop, D. R.;
Yourshaw, I.; Seinfeld, J. H.; Flagan, R. C. (2003) A comparison of particle mass spectrometers during the
1999 Atlanta Supersites project. J. Geophys. Res. [Atmos.] 108(D7): 10.1029/2001JD000660.
Mignacca, D.; Stubbs, K. (1999) Effects of equilibration temperature onPM10 concentrations from the TEOM
method in the Lower Fraser Valley. J. Air Waste Manage. Assoc. 49: 1250-1254.
Miller, F. J.; Gardner, D. E.; Graham, J. A.; Lee, R. E., Jr.; Wilson, W. E.; Bachmann, J. D. (1979) Size
considerations for establishing a standard for inhalable particles. J. Air Pollut. Control Assoc. 29: 610-615.
2-124
-------
Misra, C.; Geller, M. D.; Shah, P.; Sioutas, C.; Solomon, P. A. (2001) Development and evaluation of a continuous
coarse (PM10-PM25) particle monitor. J. Air Waste Manage. Assoc. 51: 1309-1317.
Modey, W. K.; Pang, Y.; Eatough, N. L.; Eatough, D. J. (2001) Fine paniculate (PM25) composition in Atlanta,
USA: assessment of the particle concentrator-Brigham Young University organic sampling system,
PC-BOSS, during the EPA supersite study. Atmos. Environ. 35: 6493-6502.
Moosmtiller, H.; Arnott, W. P.; Rogers, C. F. (1998) Photoacoustic and filter measurements related to aerosol light
absorption during the Northern Front Range Air Quality Study (Colorado 1996/1997). J. Geophys. Res.
103: 28,149-28,157.
Morawska, L.; Bofinger, N. D.; Kocis, L.; Nwankwoala, A. (1998) Submicrometer and supermicrometer particles
from diesel vehicle emissions. Environ. Sci. Technol. 32: 2033-2042.
Morio, L. A.; Hooper, K. A.; Brittingham, J.; Li, T.-H.; Gordon, R. E.; Turpin, B. J.; Laskin, D. L. (2001) Tissue
injury following inhalation of fine paniculate matter and hydrogen perioxide is associated with altered
production of inflammatory mediators and antioxidants by alveolar macrophages. Toxicol. Appl.
Pharmacol. 177: 188-199.
Mueller, P. K.; Collins, J. F. (1980) Development of a paniculate sulfate analyzer. Palo Alto, CA: Electric Power
Research Institute; report no. P-1382F.
Mueller, P. K.; Fung, K. K.; Heisler, S. L.; Grosjean, D.; Hidy, G. M. (1982) Atmospheric paniculate carbon
observations in urban and rural areas of the United States. In: Wolff, G. T.; Klimisch, R. L., eds. Paniculate
carbon: atmospheric life cycle: proceedings of an international symposium; October 1980; Warren, MI.
New York, NY: Plenum Press; pp. 343-370.
Mueller, P. K.; Hansen, D. A.; Watson, J. G., Jr. (1986) The subregional cooperative electric utility, Department of
Defense, National Park Service, and EPA study (SCENES) on visibility: an overview. Palo Alto, CA:
Electric Power Research Institute; report no. EA-4664-SR.
Mukerjee, S.; Ellenson, W. D.; Lewis, R. G.; Stevens, R. K.; Somerville, M. C.; Shadwick, D. S.; Willis, R. D.
(1997) An environmental scoping study in the lower Rio Grande Valley of Texas-Ill, residential
microenvironmental monitoring for air, house dust, and soil. Environ. Int. 23: 657-673.
Munksgaard, N. C.; Parry, D. L. (1998) Lead isotope ratios determined by ICP-MS: monitoring of mining-derived
metal particulates in atmospheric fallout, Northern Territory, Australia. Sci. Total Environ. 217: 113-126.
Murphy, D. M.; Thomson, D. S. (1997) Chemical composition of single aerosol particles at Idaho Hill: negative ion
measurements. J. Geophys. Res. 102: 6353-6368.
Nadkarni, R. A. (1975) Multielement analysis of coal and coal fly ash standards by instrumental neutron activation
analysis. Radiochem. Radioanal. Lett. 21: 161-176.
Nam, S.-H.; Masamba, W. R. L.; Montaser, A. (1993) Investigation of helium inductively coupled plasma-mass
spectrometry for the detection of metals and nonmetals in aqueous solutions. Anal. Chem. 65: 2784-2790.
NARSTO. (2003) Fine particle assessment (review draft). Oak Ridge, TN: U.S. Department of Energy, Oak Ridge
National Laboratory, North American Research Strategy for Tropospheric Ozone, Quality Systems Science
Center. Available: http://www.cgenv.com/Narsto/ (22 April 2003).
National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
National Academy Press. 3v.
Nelson, M. P.; Zugates, C. T.; Treado, P. J.; Casucio, G. S.; Exline, D. L.; Schlaegle, S. F. (2000) Combining Raman
chemical imaging scanning electron microscopy (SEM) to characterize ambient fine paniculate. Aerosol
Sci. Technol. 34: 108-117.
Neubauer, K. R.; Johnston, M. V.; Wexler, A. S. (1998) Humidity effects on the mass spectra of single aerosol
particles. Atmos. Environ. 32: 2521-2529.
Nham, T. T. (1996) Cool plasma technique for the ultra-trace level determination of 56Fe, 40Ca, 39K, 7Li, 24Mg
and 23Na for semi-conductor applications. Mulgrave, Victoria, Australia: Varian Australia Pty Ltd;
ICP-MS instruments at work no. ICP-MS-7. Available:
http://www.varianinc.com/cgi-bin/nav?varinc/docs/osi/icpms/atwork/icpms7&cid=680052
[18 January, 2002].
Noble, C. A.; Prather, K. A. (1996) Real-time measurement of correlated size and composition profiles of individual
atmospheric aerosol particles. Environ. Sci. Technol. 30: 2667-2680.
Noble, C. A.; Prather, K. A. (1998) Single particle characterization of albuterol metered dose inhaler aerosol in near
real-time. Aerosol Sci. Technol. 29: 294-306.
Noble, C. A.; Vanderpool, R. W.; Peters, T. M.; McElroy, F. F.; Gemmill, D. B.; Wiener, R. W. (2001) Federal
reference and equivalent methods for measuring fine paniculate matter. Aerosol Sci. Technol. 34: 457-464.
2-125
-------
Noel, M. A.; Topart, P. A. (1994) High-frequency impedance analysis of quartz crystal microbalances. 1. General
considerations. Anal. Chem. 66: 484-491.
Novakov, T.; Corrigan, C. E.; Penner, J. E.; Chuang, C. C.; Rosario, O.; Mayol Bracero, O. L. (1997) Organic
aerosols in the Caribbean trade winds: a natural source? J. Geophys. Res. [Atmos.] 102: 21307-21313.
O'Dowd, C. D.; Aalto, P.; Hameri, K.; Kulmala, M.; Hoffmann, T. (2002) Atmospheric particles from organic
vapours. Nature (London, U. K.) 416: 497-498.
Obeidi, F. D.; Eatough, D. J. (2002) Continuous measurement of semi-volatile fine paniculate mass inProvo, UT.
Aerosol Sci. Technol. 36: 191-203.
Obeidi, F.; Eatough, N. L.; Eatough, D. J. (2002) Use of the RAMS to measure semivolatile fine paniculate matter at
Riverside and Bakersfield, California. Aerosol Sci. Technol. 36: 204-216.
Odum, J. R.; Yu, J.; Kamens, R. M. (1994) Modeling the mass transfer of semivolatile organics in combustion
aerosols. Environ. Sci. Technol. 28: 2278-2285.
Ohta, S.; Okita, T. (1990) A chemical characterization of atmospheric aerosol in Sapporo. Atmos. Environ. Part A
24A: 815-822.
Ohta, S.; Hori, M.; Yamagata, S.; Murao, N. (1998) Chemical characterization of atmospheric fine particles in
Sapporo with determination of water content. Atmos. Environ. 32: 1021-1025.
Olin, J. G.; Sem, G. J. (1971) Piezoelectric microbalance for monitoring the mass concentration of suspended
particles. Atmos. Environ. 5: 653-668.
Olmez, I. (1989) Trace element signatures in groundwater pollution. In: Watson, J. G., ed. Receptor models in air
resources management: an international specialty conference; February 1988; San Francisco, CA.
Pittsburgh, PA: Air & Waste Management Association; pp. 3-11. (APCA transactions series: no. 14).
Ondov, J. M. (1996) Paniculate tracers for source attribution: potential for application to California's San Joaquin
Valley. J. Aerosol Sci. 27(suppl. 1): S687-S688.
Ondov, J. M.; Divita, F., Jr. (1993) Size spectra for trace elements in urban aerosol particles by instrumental neutron
activation analysis. J. Radioanal. Nucl. Chem. 167: 247-258.
Ondov, J. M.; Kelly, W. R.; Holland, J. Z.; Lin, Z. C.; Wight, S. A. (1992) Tracing fly ash emitted from a coal-fired
power plant with enriched rare-earth isotopes: an urban scale test. Atmos. Environ. Part B 26B: 453-462.
Ono, D. M.; Hardebeck, E.; Parker, J.; Cox, B. G. (2000) Systematic biases in measured PM10 values with U.S.
Environmental Protection Agency-approved samplers at Owens Lake, California. J. Air Waste Manage.
Assoc. 50: 1144-1156.
Pak, S. S.; Liu, B. Y. H.; Rubow, K. L. (1992) Effect of coating thickness on particle bounce in inertia! impactors.
Aerosol Sci. Technol. 16: 141-150.
Pang, Y.; Ren, Y.; Obeidi, F.; Hastings, R.; Eatough, D. J.; Wilson, W. E. (2001) Semi-volatile species in PM2s:
comparison of integrated and continuous samplers for PM2 5 research or monitoring. J. Air Waste Manage.
Assoc. 51:25-36.
Pang, Y.; Eatough, N. L.; Wilson, J.; Eatough, D. J. (2002a) Effect of semivolatile material on PM2 5 measurement
by the PM2 5 federal reference method sampler at Bakersfield, California. Aerosol Sci. Technol.
36: 289-299.
Pang, Y.; Eatough, N. L.; Eatough, D. J. (2002b) PM2 5 semivolatile organic material at Riverside, CA: implications
for the PM25 federal reference method sampler. Aerosol Sci. Technol. 36: 277-288.
Pankow, J. F. (1987) Review and comparative analysis of the theories on partitioning between the gas and aerosol
paniculate phases in the atmosphere. Atmos. Environ. 21: 2275-2283.
Pankow, J. F. (1994a) An absorption model of the gas/aerosol partitioning involved in the formation of secondary
organic aerosol. Atmos. Environ. 28: 189-193.
Pankow, J. F. (1994b) An absorption model of gas/particle partitioning of organic compounds in the atmosphere.
Atmos. Environ. 28: 185-188.
Pankow, J. F.; Bidleman, T. F. (1991) Effects of temperature, TSP and per cent non-exchangeable material in
determining the gas-particle partitioning of organic compounds. Atmos. Environ. Part A 25: 2241-2249.
Pankow, J. F.; Storey, J. M. E.; Yamasaki, H. (1993) Effects of relative humidity on gas/particle partitioning of
semivolatile organic compounds to urban paniculate matter. Environ. Sci. Technol. 27: 2220-2226.
Parmar, S. S.; Grosjean, D. (1990) Laboratory tests of KI and alkaline annular denuders. Atmos. Environ. Part A
24: 2695-2698.
Patashnick, H.; Rupprecht, G.; Ambs, J. L.; Meyer, M. B. (2001) Development of a reference standard for
paniculate matter mass in ambient air. Aerosol Sci. Technol. 34: 42-45.
Pellizzari, E. D.; Krost, K. J. (1984) Chemical transformations during ambient air sampling for organic vapors.
Anal. Chem. 56: 1813-1819.
2-126
-------
Peng, C.; Chan, C. K. (2001) The water cycles of water-soluble organic salts of atmospheric importance.
Atmos. Environ. 35: 1183-1192.
Penner, J. E.; Novakov, T. (1996) Carbonaceous particles in the atmosphere: a historical perspective to the fifth
international conference on carbonaceous particles in the atmosphere. J. Geophys. Res. [Atmos.]
101: 19,373-19,378.
Perry, K. D.; Hobbs, P. V. (1994) Further evidence for particle nucleation in clear air adjacent to marine cumulus
clouds. J. Geophys. Res. [Atmos.] 99: 22,803-22,818.
Perry, K. D.; Cahill, T. A.; Schnell, R. C.; Harris, J. M. (1999) Long-range transport of anthropogenic aerosols to the
National Oceanic and Atmospheric Administration baseline station at Mauna Loa Observatory, Hawaii.
J. Geophys. Res. Atmos. 104: 18,521-18,533.
Peters, A.; Liu, E.; Verrier, R. L.; Schwartz, J.; Gold, D. R.; Mittleman, M.; Baliff, J.; Oh, J. A.; Allen, G.;
Monahan, K.; Dockery, D. W. (2000) Air pollution and incidence of cardiac arrhythmia. Epidemiology
11: 11-17.
Peters, T. M.; Vanderpool, R. W.; Wiener, R. W. (2001a) Design and calibration of the EPA PM2 5 well impactor
ninety-six (WINS). Aerosol Sci. Technol. 34: 389-397.
Peters, T. M.; Norris, G. A.; Vanderpool, R. W.; Gemmill, D. B.; Wiener, R. W.; Murdoch, R. W.; McElroy, F. F.;
Pitchford, M. (200 Ib) Field performance of PM25 federal reference method samplers. Aerosol Sci. Technol.
34: 433-443.
Peters, T. M.; Gussman, R. A.; Kenny, L. C.; Vanderpool, R. W. (2001c) Evaluation of PM2 5 size selectors used in
speciation samplers. Aerosol Sci. Technol. 34: 422-429.
Petzold, A.; Niessner, R. (1996) Photoacoustic soot sensor for in-situ black carbon monitoring. Appl. Phys. B:
Lasers Opt. 63: 191-197.
Petzold, A.; Kopp, C.; Niessner, R. (1997) The dependence of the specific attenuation cross-section on black carbon
mass fraction and particle size. Atmos. Environ. 31: 661-672.
Pilinis, C.; Seinfeld, J. H.; Grosjean, D. (1989) Water content of atmospheric aerosols. Atmos. Environ.
23: 1601-1606.
Piorek, S. (1994) Modern, PC based, high resolution portable EDXRF analyzer offers laboratory performance for
field, in-situ analysis of environmental contaminants. Nucl. Instr. Meth. Phys. Res. A 353: 528-533.
Pirrone, N.; Keeler, G. J. (1996) The Rouge River watershed pollution by trace elements: atmospheric depositions
and emission sources. Water Sci. Technol. 33: 267-275.
Plantz, M. (1996) Common molecular ion interferences in ICP-MS. Mulgrave, Victoria, Australia: Varian Australia
Pty Ltd; ICP-MS instruments at work no. ICP-MS-6. Available:
http://www.varianinc.com/cgi-bin/nav?varinc/docs/osi/icpms/atwork/index&cid=680052
[18 January, 2002].
Poor, N.; Clark, T.; Nye, L.; Tamanini, T.; Tate, K.; Stevens, R.; Atkeson, T. (2002) Field performance of
dichotomous sequential PM air samplers. Atmos. Environ. 36: 3289-3298.
Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
(1995) Paniculate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
Respir. Crit. Care Med. 151: 669-674.
Pope, C. A., Ill; Hill, R. W.; Villegas, G. M. (1999) Paniculate air pollution and daily mortality on Utah's Wasatch
Front. Environ. Health Perspect: 107: 567-573.
Potts, P. J.; Webb, P. C. (1992) X-ray fluorescence spectrometry. J. Geochem. Explor. 44: 251-296.
Radke, L. F.; Hobbs, P. V. (1991) Humidity and particle fields around some small cumulus clouds. J. Atmos. Sci.
48: 1190-1193.
Raes, F.; VanDingenen, R.; Cuevas, E.; Van Velthoven, P. F. J.; Prospero, J. M. (1997) Observations of aerosols in
the free troposphere and marine boundary layer of the subtropical Northeast Atlantic: discussion of
processes determining their size distribution. J. Geophys. Res. [Atmos.] 102: 21315-21328.
Ranweiler, L. E.; Moyers, J. L. (1974) Atomic absorption procedure for analysis of metals in atmospheric paniculate
matter. Environ. Sci. Technol. 8: 152-156.
Rao, A. K.; Whitby, K. T. (1977) Nonideal collection characteristics of single stage and cascade impactors. Am. Ind.
Hyg. Assoc. J. 38: 174-177.
Rao, A. K.; Whitby, K. T. (1978) Non-ideal collection characteristics of inertial impactors-I. single-stage impactors
and solid particles. J. Aerosol Sci. 9: 77-86.
Ravenhall, D. G.; Forney, L. J.; Jazayeri, M. (1978) Aerosol sizing with a slotted virtual impactor. J. Colloid
Interface Sci. 65: 108-117.
2-127
-------
Reid, J. S.; Cahill, T. A.; Wakabayashi, P. H.; Dunlap, M. R. (1994) Geometric/aerodynamic equivalent diameter
ratios of ash aggregate aerosols collected inhuming Kuwaiti well fields. Atmos. Environ. 28: 2227-2234.
Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. (1998) Real-time characterization of the organic
composition and size of individual diesel engine smoke particles. Environ. Sci. Technol. 32: 2672-2679.
Reist, P. C. (1984) Introduction to aerosol science. New York, NY: Macmillan Publishing Company.
Reist, P. C. (1993) Introduction to aerosol science. New York, NY: McGraw-Hill.
Risse, U.; Flammenkamp, E.; Kettrup, A. (1996) Determination of aromatic hydrocarbons in air using diffusion
denuders. Fresenius' J. Anal. Chem. 356: 390-395.
Roberts, R. A.; Corkill, J. (1998) Grass seed field smoke and its impact on respiratory health. J. Environ. Health
60: 10-16.
Rounds, S. A.; Pankow, J. F. (1990) Application of a radial diffusion model to describe gas/particle sorption
kinetics. Environ. Sci. Technol. 24: 1378-1386.
Rounds, S. A.; Tiffany, B. A.; Pankow, J. F. (1993) Description of gas/particle sorption kinetics with an intraparticle
diffusion model: desorption experiments. Environ. Sci. Technol. 27: 366-377.
Rumburg, B.; Alldredge, R.; Claiborn, C. (2001) Statistical distributions of paniculate matter and the error
associated with sampling frequency. Atmos. Environ. 35: 2907-2920.
Rupprecht, G.; Patashnick, H.; Beeson, D. E.; Green, R. N.; Meyer, M. B. (1995) A new automated monitor for the
measurement of paniculate carbon in the atmosphere. In: Paniculate matter: health and regulatory issues:
proceedings of an international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
Management Association; pp. 309-319. (A&WMA publication VIP-49).
Sakata, K.; Kawabata, K. (1994) Reduction of fundamental polyatomic ions in inductively coupled plasma mass
spectrometry. Spectrochim. Acta49B: 1027-1038.
Sarangapani, R.; Wexler, A. S. (1996) Growth and neutralization of sulfate aerosols in human airways. J. Appl.
Physiol. 81:480-490.
Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem.
24: 57-109.
Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. H. (1995) Organics alter hygroscopic behavior of
atmospheric particles. J. Geophys. Res. [Atmos.] 100: 18,755-18,770.
Schmid, H.; Laskus, L.; Abraham, J. H.; Baltensperger, U.; Lavanchy, V.; Bizjak, M.; Burba, P.; Cachier, H.;
Crow, D.; Chow, J.; Gnauk, T.; Even, A.; Ten Brink, H. M.; Giesen, K.-P.; Hitzenberger, R.; Hueglin, C.;
Maenhaut, W.; Pio, C.; Carvalho, A.; Putaud, J.-P.; Toom-Sauntry, D.; Puxbaum, H. (2001) Results of the
"carbon conference" international aerosol carbon round robin test stage I. Atmos. Environ. 35: 2111-2121.
Schroder, F.; Strom, J. (1997) Aircraft measurements of submicrometer aerosol particles (> 7 nm) in the midaltitude
free troposphere and tropopause region. Atmos. Res. 44: 333-356.
Schwartz, S. E. (1984) Gas-aqueous reactions of sulfur and nitrogen oxides in liquid-water clouds. In: Calvert, J. G.,
ed. SO2, NO and NO2 oxidation mechanisms: atmospheric considerations. Boston, MA: Butterworth
Publishers; pp. 173-208. (Teasley, J. L, ed. Acid precipitation series: v. 3).
Schwartz, S. E. (1986) Chemical conversions in clouds. In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P., eds.
Aerosols: research, risk assessment and control strategies, proceedings of the second U.S.-Dutch
international symposium; May 1985; Williamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 349-375.
Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles? J. Air
Waste Manage. Assoc. 46: 927-939.
Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change.
New York, NY: John Wiley & Sons, Inc.
Sem, G. J.; Tsurubayashi, K.; Homma, K. (1977) Performance of the piezoelectric microbalance respirable aerosol
sensor. Am. Ind. Hyg. Assoc. J. 38: 580-588.
Shah, J. J.; Rau, J. A. (1990) Carbonaceous methods comparison study: interlaboratory round robin interpretation
results. Sacramento, CA: California Air Resources Board. Available:
http://www.arb.ca.gov/research/apr/past/pm.htm (21 March 2002).
Silva, P. J.; Prather, K. A. (1997) On-line characterization of individual particles from automobile emissions.
Environ. Sci. Technol. 31: 3074-3080.
Simon, P. K.; Dasgupta, P. K. (1995) Continuous automated measurement of the soluble fraction of atmospheric
paniculate matter. Anal. Chem. 67: 71-78.
Sioutas, C.; Koutrakis, P. (1996) Inertial separation of ultrafine particles using a condensational growth/virtual
impaction system. Aerosol Sci. Technol. 25: 424-436.
2-128
-------
Sioutas, C.; Koutrakis, P.; Burton, R. M. (1994a) Development of a low outpoint slit virtual impactor for sampling
ambient fine particles. J. Aerosol Sci. 25: 1321-1330.
Sioutas, C.; Koutrakis, P.; Burton, R. M. (1994b) A high-volume small outpoint virtual impactor for separation of
atmospheric paniculate from gaseous pollutants. Part. Sci. Technol. 12: 207-221.
Sioutas, C.; Koutrakis, P.; Olson, B. A. (1994c) Development and evaluation of a low cutpoint virtual impactor.
Aerosol Sci. Technol. 21: 223-235.
Sioutas, C.; Koutrakis, P.; Wolfson, J. M. (1994d) Particle losses in glass honeycomb denuder samplers. Aerosol
Sci. Technol. 21: 137-148.
Sioutas, C.; Koutrakis, P.; Burton, R. M. (1995a) A technique to expose animals to concentrated fine ambient
aerosols. Environ. Health Perspect. 103: 172-177.
Sioutas, D.; Koutrakis, P.; Ferguson, S. T.; Burton, R. M. (1995b) Development and evaluation of a prototype
ambient particle concentrator for inhalation exposure studies. Inhalation Toxicol. 7: 633-644.
Sioutas, C.; Koutrakis, P.; Wolfson, J. M.; Azaroff, L. S.; Mulik, J. D. (1996a) Field evaluation of a glass
honeycomb denuder/filter pack system to collect atmospheric gases and particles. In: Gay, B. W.; Fuerst,
R. G.; Jayanty, R. K. M., eds. Measurement of toxic and related air pollutants: proceedings of a conference;
May, 1994; Durham, NC. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
Research and Development; report no. EPA/600/R-94/136; pp. 426-435.
Sioutas, C.; Wang, P. Y.; Ferguson, S. T.; Koutrakis, P.; Mulik, J. D. (1996b) Laboratory and field evaluation of an
improved glass honeycomb denuder/filter pack sampler. Atmos. Environ. 30: 885-895.
Sioutas, C.; Koutrakis, P.; Wang, P.-Y.; Babich, P.; Wolfson, J. M. (1999) Experimental investigation of pressure
drop with particle loading inNuclepore filters. Aerosol Sci. Technol. 30: 71-83.
Sioutas, C.; Kim, S.; Chang, M.; Terrell, L. L.; Gong, H., Jr. (2000) Field evaluation of a modified DataRAM MIE
scattering monitor for real-time PM2 5 mass concentration measurements. Atmos. Environ. 34: 4829-4838.
Solomon, P. A.; Moyers, J. L.; Fletcher, R. A. (1983) High-volume dichotomous virtual impactor for the
fractionation and collection of particles according to aerodynamic size. Aerosol Sci. Technol. 2: 455-464.
Solomon, P. A.; Salmon, L. G.; Fall, T.; Cass, G. R. (1992) Spatial and temporal distribution of atmospheric nitric
acid and paniculate nitrate concentrations in the Los Angeles area. Environ. Sci. Technol. 26: 1596-1601.
Song, X.-H.; Faber, N. M.; Hopke, P. K.; Suess, D. T.; Prather, K. A.; Schauer, J. J.; Cass, G. R. (2001) Source
apportionment of gasoline and diesel by multivariate calibration based on single particle mass spectral data.
Anal. Chim. Acta446: 329-343.
Speer, R. E.; Barnes, H. M.; Brown, R. (1997) An instrument for measuring the liquid water content of aerosols.
Aerosol Sci. Technol. 27: 50-61.
Spengler, J. D.; Thurston, G. D. (1983) Mass and elemental composition of fine and coarse particles in six U.S.
cities. J. AirPollut. Control Assoc. 33: 1162-1171.
Stahlschmidt, T.; Schulz, M.; Dannecker, W. (1997) Application of total-reflection X-ray fluorescence for the
determination of lead, calcium and zinc in size-fractionated marine aerosols. Spectrochim. Acta B
52: 995-1001.
Stein, S. W.; Turpin, B. J.; Cai, X.; Huang, P.-F.; McMurry, P. H. (1994) Measurements of relative
humidity-dependent bounce and density for atmospheric particles using the DMA-impactor technique.
Atmos. Environ. 28: 1739-1746.
Stevens, R. K.; O'Keeffe, A. E.; Ortman, G. C. (1969) Absolute calibration of a flame photometric detector to
volatile sulfur compounds at sub-part-per-million levels. Environ. Sci. Technol. 3: 652-655.
Stevens, R. K.; Mulik, J. D.; O'Keefe, A. E.; Krost, K. J. (1971) Gas chromatography of reactive sulfur gases in air
at the parts-per-billion level. Anal. Chem. 43: 827-831.
Stolzenburg, M. R.; Hering, S. V. (2000) Method for the automated measurement of fine particle nitrate in the
atmosphere. Environ. Sci. Technol. 34: 907-914.
Streit, N.; Weingartner, E.; Zellweger, C.; Schwikowski, M.; Gaggeler, H. W.; Baltensperger, U. (2000)
Characterization of size-fractionated aerosol from the Jungfraujoch (3580 m asl) using total reflection x-ray
fluorescence (TXRF). Int. J. Environ. Anal. Chem. 76: 1-16.
Strobel, H. A.; Heineman, W. R. (1989) Chemical instrumentation: a systematic approach. 3rd ed. New York, NY:
Wiley and Sons, Inc.
Strommen, M. R.; Kamens, R. M. (1997) Development and application of a dual-impedance radial diffusion model
to simulate the partitioning of semivolatile organic compounds in combustion aerosols. Environ. Sci.
Technol. 31:2983-2990.
2-129
-------
Strothmann, J. A.; Schiermeier, F. A. (1979) Documentation of the regional air pollution study (RAPS) and related
investigations in the St. Louis air quality control region. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Environmental Science Research Laboratory; report no. EPA-600/4-79-076.
Suarez, A. E.; Caffrey, P. F.; Ondov, J. M; Thaung, K. C. (1996) Size distribution of Ir-tagged soot from diesel
sanitation trucks. J. Aerosol Sci. 27(suppl. 1): S697-S698.
Suggs, J. C.; Burton, R. M. (1983) Spatial characteristics of inhalable particles in the Philadelphia metropolitan area.
J. AirPollut. Control Assoc. 33: 688-691.
Suh, H. H.; Spengler, J. D.; Koutrakis, P. (1992) Personal exposures to acid aerosols and ammonia. Environ. Sci.
Technol. 26:2507-2517.
Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1994) The relationship between airborne acidity and ammonia in indoor
environments. J. Exposure Anal. Environ. Epidemiol. 4: 1-23.
Sutherland, J. L.; Bhardwaja, P. S. (1987) Composition of the aerosol in northern Arizona and southern Utah. In:
Bhardwaja, P. S., ed. Transactions, Visibility Protection: Research and Policy Aspects. Pittsburgh, PA:
Air Pollution Control Association; pp. 373-385.
Tan, S. H.; Horlick, G. (1986) Background spectral features in inductively coupled plasma/mass spectrometry.
Appl. Spectrosc. 40: 445-460.
Tang, I. N. (1980) Deliquescence properties and particle size change of hygroscopic aerosols. In: Willeke, K., ed.
Generation of aerosols and facilities for exposure experiments. Ann Arbor, MI: Ann Arbor Science
Publishers, Inc.; pp. 153-167.
Tang, H.; Lewis, E. A.; Eatough, D. J.; Burton, R. M.; Farber, R. J. (1994) Determination of the particle size
distribution and chemical composition of semi-volatile organic compounds in atmospheric fine particles
with a diffusion denuder sampling system. Atmos. Environ. 28: 939-947.
Tanner, R. L.; D'Ottavio, T.; Garber, R.; Newman, L. (1980) Determination of ambient aerosol sulfur using a
continuous flame photometric detection system. I. Sampling system for aerosol sulfate and sulfuric acid.
Atmos. Environ. 14: 121-127.
Thomson, D. S.; Murphy, D. M. (1994) Analyzing single aerosol particles in real time. CHEMTECH 24: 30-35.
Tobias, H. J.; Beving, D. E.; Ziemann, P. J.; Sakurai, H.; Zuk, M.; McMurry, P. H.; Zarling, D.; Wayrulonis, R.;
Kittelson, D. B. (2001) Chemical analysis of diesel engine nanoparticles using a nano-DMA/thermal
desorption particle beam mass spectrometer. Environ. Sci. Technol. 35: 2233-2243.
Tolocka, M. P.; Peters, T. M.; Vanderpool, R. W.; Chen, F. L.; Wiener, R. W. (2001a) On the modification of the
low flow-rate PM10 dichotomous sampler inlet. Aerosol Sci. Technol. 34: 407-415.
Tolocka, M. P.; Solomon, P. A.; Mitchell, W.; Norris, G. A.; Gemmill, D. B.; Wiener, R. W.; Vanderpool, R. W.;
Homolya, J. B.; Rice, J. (200Ib) East vs. west in the US: chemical characteristics of PM2 5 during the winter
of 1999. Aerosol Sci. Technol. 34: 88-96.
Torok, S. B.; Labar, J.; Schmeling, M.; Van Grieken, R. E. (1998) X-ray spectrometry. Anal. Chem. 70: 495R-517R.
Tsai, C.-J.; Cheng, Y.-H. (1996) Comparison of two ambient beta gauge PM10 samplers. J. Air Waste Manage.
Assoc. 46: 142-147.
Tsai, C.-J.; Huang, H.-Y. (1995) Atmospheric aerosol sampling by an annular denuder system and a high-volume
PM10 sampler. Environ. Int. 21: 283-291.
Turner, D. B. (1994) Workbook of atmospheric dispersion estimates: an introduction to dispersion modeling. 2nd ed.
Boca Raton, FL: Lewis Publishers.
Turner, J. R.; Hering, S. V. (1987) Greased and oiled substrates as bounce-free impaction surfaces. J. Aerosol Sci.
18: 215-224.
Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin: a descriptive
analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
Turpin, B. J.; Gary, R. A.; Huntzicker, J. J. (1990) An in situ, time-resolved analyzer for aerosol organic and
elemental carbon. Aerosol Sci. Technol. 12: 161-171.
Turpin, B. J.; Liu, S.-P.; Podolske, K. S.; Gomes, M. S. P.; Eisenreich, S. J.; McMurry, P. H. (1993) Design and
evaluation of a novel diffusion separator for measuring gas/particle distributions of semivolatile organic
compounds. Environ. Sci. Technol. 27: 2441-2449.
Turpin, B. J.; Huntzicker, J. J.; Hering, S. V. (1994) Investigation of organic aerosol sampling artifacts in the
Los Angeles basin. Atmos. Environ. 28: 3061-3071.
Turpin, B. J.; Saxena, P.; Andrews, E. (2000) Measuring and simulating paniculate organics in the atmosphere:
problems and prospects. Atmos. Environ. 34: 2983-3013.
2-130
-------
Tyler, G. (1992) ICP vs. AA: how do you choose? Environ. Lab. (October/November): 31-34.
U.S. Court of Appeals for the District of Columbia. (1999) American Trucking Associations, Inc. vs. U.S.
Environmental Protection Agency. 175 F.3d 1027 (D.C. Cir. 1999).
U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA;
PB84-156777.
U.S. Environmental Protection Agency. (1996a) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
U.S. Environmental Protection Agency. (1996b) Review of the national ambient air quality standards for paniculate
matter: policy assessment of scientific and technical information. OAQPS staff paper. Research Triangle
Park, NC: Office of Air Quality Planning and Standards; report no. EPA/452/R-96-013. Available from:
NTIS, Springfield, VA; PB97-115406REB.
U.S. Environmental Protection Agency. (1998) Monitoring PM2 5 in ambient air using designated reference of class I
equivalent methods. In: Quality Assurance Handbook for Air Pollution Measurement Systems, V. II, Part
II, Section 2.12. Research Triangle Park, NC: Research and Development, Environmental Monitoring
Systems Laboratory.
U.S. Environmental Protection Agency. (1999) Paniculate matter (PM2 5) speciation guidance. Final draft. Edition 1.
Research Triangle Park, NC: Office of Air Quality Planning and Standards. Available:
www.epa.gov/ttn/amtic/files/ambient/pm25/spec/specfinl.pdf [2000, October 11].
U.S. Environmental Protection Agency. (2001) List of designated reference and equivalent methods. Available:
http://www.epa.gov/ttn/amtic/files/ambient/critena/repmlist.pdf [05 March 02].
U.S. Environmental Protection Agency. (2002a) Ambient monitoring technology information center (AMTIC).
Washington, DC: Office of Air Quality Planning and Standards; Monitoring and Quality Assurance Group.
Available: http://www.epa.gov/ttn/amtic/ (20 March 2002).
U.S. Environmental Protection Agency. (2002b) Support center for regulatory air models. Washington, DC: Office
of Air Quality Planning and Standards. Available: http://www.epa.gov/scram001 (21 March 2002).
Van Dingenen, R.; Raes, F.; Jensen, N. R. (1995) Evidence for anthropogenic impact on number concentration and
sulfate content of cloud-processed aerosol particles over the North Atlantic. J. Geophys. Res. [Atmos.]
100: 21,057-21,067.
Van Loy, M.; Bahadori, T.; Wyzga, R.; Hartsell, B.; Edgerton, E. (2000) The Aerosol Research and Inhalation
Epidemiology Study (ARIES): PM25 mass and aerosol component concentrations and sampler
intercomparisons. J. Air Waste Manage. Assoc. 50: 1446-1458.
Vanderpool, R. W.; Peters, T. M.; Natarajan, S.; Gemmill, D. B.; Wiener, R. W. (2001) Evaluation of the loading
characteristics of the EPA WINS PM25 separator. Aerosol Sci. Technol. 34: 444-456.
Vasiliou, J. G.; Sorenson, D.; McMurry, P. H. (1999) Sampling at controlled relative humidity with a cascade
impactor. Atmos. Environ. 30: 1049-1056.
Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci.
Technol. 28: 555-562.
Vincent, J. H. (1989) Aerosol sampling: science and practice. New York, NY: John Wiley & Sons.
Vincent, J. H. (1995) Aerosol science for industrial hygienists. Oxford, United Kingdom: Pergamon.
Wallace, J. C.; Kites, R. A. (1995) Computer-controlled low-volume air sampler for the measurement of semivolatile
organic compounds. Environ. Sci. Technol. 29: 2099-2106.
Wallace, L.; Howard-Reed, C. (2002) Continuous monitoring of ultrafine, fine, and coarse particles in a residence
for 18 months in 1999-2000. J. Air Waste Manage. Assoc. 52: 828-844.
Wang, P. Y. (1997) Continuous aerosol mass measurement by flow obstruction. Boston, MA: Harvard University,
School of Public Health.
Wang, H.-C.; John, W. (1987) Comparative bounce properties of particle materials. Aerosol Sci. Technol.
7: 285-299.
Wang, H.-C.; John, W. (1988) Characteristics of the Berner impactor for sampling inorganic ions. Aerosol Sci.
Technol. 8: 157-172.
Ward, M. D.; Buttry, D. A. (1990) In situ interfacial mass detection with piezoelectric transducers. Science
(Washington, DC) 249: 1000-1007.
2-131
-------
Watson, J. G.; Chow, J. C; Richards, L. W.; Neff, W. D.; Andersen, S. R.; Dietrich, D. L.; Houck, J. E.; Olmez, I.
(1988a) The 1987-88 metro Denver brown cloud study volume II: measurements. Reno, NV: Desert
Research Institute; final report no. 8810.1F2 (2 of 3).
Watson, J. G.; Chow, J. C.; Richards, L. W.; Neff, W. D.; Andersen, S. R.; Dietrich, D. L.; Houck, J. E.; Olmez, I.
(1988b) The 1987-88 metro Denver brown cloud study volume III: data interpretation. Reno, NV: Desert
Research Institute; final report no. 8810.1F3 (3 of 3).
Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
(1990a) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B.; Nguyen, Q. T.; Meyer, E. L.; Pace, T. G.
(1990b) Receptor model technical series, v. Ill (1989 revision): CMB7 user's manual. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; report no.
EPA-450/4-90-004. Available: http://www.epa.gov/scram001/ [19 February 2002].
Watson, J. G.; Chow, J. C.; Lu, Z.; Gertler, A. W. (1993) Paniculate and gaseous organic receptor modeling for the
Southern California Air Quality Study. Sacramento, CA: California Air Resource Board; report no.
8733.IF. Available: http://www.arb.ca.gov/research/apr/past/atmospheric.htmMir%20Quality%20Study
(5 June 2003).
Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M.; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994a)
Chemical mass balance source apportionment of PM10 during the Southern California Air Quality Study.
Aerosol Sci. Technol. 21: 1-36.
Watson, J. G.; Chow, J. C.; Lowenthal, D. H.; Pritchett, L. C.; Frazier, C. A.; Neuroth, G. R.; Robbins, R. (1994b)
Differences in the carbon composition of source profiles for diesel- and gasoline-powered vehicles. Atmos.
Environ. 28: 2493-2505.
Watson, J. G.; Fujita, E. M.; Chow, J. C.; Zielinska, B.; Richards, L. W.; Neff, W.; Dietrich, D. (1998) Northern
front range air quality study. Final report. Fort Collins, CO: Colorado State University, Cooperative
Institute for Research in the Atmosphere. Available: http://www.nfraqs.colostate.edu/index2.html
(16 Jan 2002).
Watson, J. G.; Chow, J. C.; Frazier, C. A. (1999) X-ray fluorescence analysis of ambient air samples. In:
Landsberger, S.; Creatchman, M., eds. Elemental Analysis of Airborne Particles. Newark, NJ: Gordon and
Breach Science Publishers; pp. 67-96.
Weber, R. J.; McMurry, P. H.; Eisele, F. L.; Tanner, D. J. (1995) Measurement of expected nucleation precursor
species and 3-500-nm diameter particles at Mauna Loa Observatory, Hawaii. J. Atmos. Sci. 52: 2242-2257.
Weber, R. J.; Marti, J. J.; McMurry, P. H.; Mauldin, L.; Tanner, D.; Eisele, F.; Brechtel, F.; Kreidenweis, S.; Kok,
G.; Schillawski, R.; Baumgardner, D. (1997) Measurements of new particle formation and ultrafine particle
growth rates at a clean continental site. J. Geophys. Res. 102: 4375-4385.
Weber, R. J.; McMurry, P. H.; Mauldin, L.; Tanner, D. J.; Eisele, F. L.; Brechtel, F. J.; Kreidenweis, S. M.; Kok,
G. L.; Schillawski, R. D.; Baumgardner, D. (1998) A study of new particle formation and growth involving
biogenic and trace gas species measured during ACE 1. J. Geophys. Res. A 103: 16,385-16,396.
Weber, R. J.; McMurry, P. H.; Mauldin, R. L., Ill; Tanner, D. J.; Eisele, F. L.; Clarke, A. D.; Kapustin, V. N. (1999)
New particle formation in the remote troposphere: a comparison of observations at various sites. Geophys.
Res. Lett. 26:307-310.
Weber, R. J.; Orsini, D.; Daun, Y.; Lee, Y.-N.; Kotz, P. J.; Brechtel, F. (2001) A particle-into-liquid collector for
rapid measurement of aerosol bulk chemical composition. Aerosol Sci. Technol. 35: 718-727.
Wedding, J. B.; Weigand, M. A. (1993) An automatic particle sampler with beta gauging. J. Air Waste Manage.
Assoc. 43:475-479.
Weingartner, E.; Burtscher, H.; Baltensperger, U. (1997) Hygroscopic properties of carbon and diesel soot particles.
Atmos. Environ. 31: 2311-2327.
Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
Environ. 33:2301-2312.
Wesolowski, J. J.; John, W.; Devor, W.; Cahill, T. A.; Feeney, P. J.; Wolfe, G.; Flocchini, R. (1977) Collection
surfaces of cascade impactors. In: Dzubay, T. G., ed. X-ray fluorescence analysis of environmental
samples. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.; pp. 121-131.
Wexler, A. S.; Sarangapani, R. (1998) Particles do not increase vapor deposition in human airways. J. Aerosol Sci.
29: 197-204.
Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
2-132
-------
Whitby, K. T.; Sverdrap, G. M. (1980) California aerosols: their physical and chemical characteristics. In: Hidy,
G. M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. I, eds. The character and origins of
smog aerosols: a digest of results from the California Aerosol Characterization Experiment (ACHEX).
New York, NY: John Wiley & Sons, Inc.; pp. 477-517. (Advances in environmental science and
technology: v. 9).
Whitby, K. T.; Charlson, R. E.; Wilson, W. E.; Stevens, R. K. (1974) The size of suspended particle matter in air.
Science (Washington, DC) 183: 1098-1099.
White, W. H. (1998) Statistical considerations in the interpretation of size-resolved paniculate mass data. J. Air
Waste Manage. Assoc. 48: 454-458.
Wiedensohler, A.; Hansson, H.-C.; Orsini, D.; Wendisch, M.; Wagner, F.; Bower, K. N; Chourlarton, T. W.;
Wells, M.; Parkin, M.; Acker, K.; Wieprecht, W.; Facchini, M. C.; Lind, J. A.; Fuzzi, S.; Arends, B. G.;
Kulmala, M. (1997) Night-time formation and occurrence of new particles associated with orographic
clouds. Atmos. Environ. 31: 2545-2559.
Willeke, K.; Baron, P. A., eds. (1993) Aerosol measurement: principles, techniques and applications. New York,
NY: Van Nostrand Reinhold Publishers.
Willeke, K.; Whitby, K. T. (1975) Atmosperic aerosols: size distribution interpretation. J. AirPollut. Control Assoc.
25: 529-534.
Williams, E. L., II; Grosjean, D. (1990) Removal of atmospheric oxidants with annular denuders. Environ. Sci.
Technol. 24: 811-814.
Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
(2000) The 1998 Baltimore Paniculate Matter Epidemiology-Exposure Study: Part 1. Comparison of
ambient, residential outdoor, indoor and apartment paniculate matter monitoring. J. Exposure Anal.
Environ. Epidemiol. 10: 518-532.
Wilson, W. E. (1995) Aerosol exposure, physics, and chemistry. Inhalation Toxicol. 7: 769-772.
Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
Wilson, W. E.; Spiller, L. L.; Ellestad, T. G.; Lamothe, P. J.; Dzubay, T. G.; Stevens, R. K.; Macias, E. S.; Fletcher,
R. A.; Husar, J. D.; Husar, R. B.; Whitby, K. T.; Kittelson, D. B.; Cantrell, B. K. (1977) General Motors
sulfate dispersion experiment: summary of EPA measurements. J. AirPollut. Control Assoc. 27: 46-51.
Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; McMurry, P. H. (2001a) Measurement of Atlanta aerosol size distributions:
observations of ultrafine particle events. Aerosol Sci. Technol. 34: 75-87.
Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; Wilson, W. E. (2001b) Use of continuous measurements of integral aerosol
parameters to estimate particle surface area. Aerosol Sci. Technol. 34: 57-65.
Wu, C. C.; Suarez, A. E.; Lin, Z.; Kidwell, C. B.; Borgoul, P. V.; Caffrey, P. F.; Ondov, J. M.; Sattler, B. (1998)
Application of an Ir tracer to determine soot exposure to students commuting to school on Baltimore public
buses. Atmos. Environ. 32: 1911-1919.
Yamasaki, H.; Kuwata, K.; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne poly cyclic
aromatic hydrocarbons. Environ. Sci. Technol. 16: 189-194.
Yatin, M.; Tuncel, S. G.; Tuncel, G.; Aras, N. K. (1994) Trace element composition of atmospheric aerosols in
Ankara, Turkey, determined by instrumental neutron activation analysis. J. Radioanal. Nucl. Chem.
181:401-411.
Zeng, X.; Wu, X.; Yao, H.; Yang, F.; Cahill, T. A. (1993) PIXE-induced XRF with transmission geometry. Nucl.
Instr. Meth. Phys. Res. B 75: 99-104.
Zhang, X. Q.; McMurry, P. H. (1987) Theoretical analysis of evaporative losses from impactor and filter deposits.
Atmos. Environ. 21: 1779-1789.
Zhang, X.; McMurry, P. H. (1992) Evaporative losses of fine paniculate nitrates during sampling. Atmos. Environ.
Part A 26: 3305-3312.
Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
27: 1593-1607.
Zhu, Y.; Hinds, W. C.; Kim, S.; Shen, S.; Sioutas, C. (2002a) Study of ultrafine particles near a major highway with
heavy-duty diesel traffic. Atmos. Environ. 36: 4323-4335.
Zhu, Y.; Hinds, W. C.; Kim, S.; Sioutas, C. (2002b) Concentration and size distribution of ultrafine particles near a
major highway. J. Air Waste Manage. Assoc. 52: 1032-1042.
Zoller, W. H.; Gordon, G. E. (1970) Instrumental neutron activation analysis of atmospheric pollutants utilizing
Ge(Li) y-ray detectors. Anal. Chem. 42: 257-265.
2-133
-------
APPENDIX 2A. TECHNIQUES FOR MEASURING OF SEMIVOLATILE
ORGANIC COMPOUNDS
Use ofDenuder Systems To Measure Semivolatile Organic Compounds
Phase distribution of semivolatile organic species has been the subject of several studies
that have employed denuder technology (see Gundel et al., 1995; Gundel and Lane, 1999) to
directly determine the phase distributions while avoiding some of the positive and negative
sampling artifacts associated with back-up quartz filters. In an ideal system with a denuder that
is 100% efficient, the gas phase would be collected in the denuder and the particle phase would
be the sum of the material collected on the filter and the adsorbent downstream. Denuder
collection efficiency depends on the denuder surface area (+), the diffusivity (+) and vapor
pressure (-) of the compound, the temperature (-) and flow rate (-) of the air stream, and the
presence of competing species (-), including water vapor (Cui et al., 1998; Kamens and Coe,
1997; Lane et al., 1988). (The + and - symbols in parentheses indicate qualitatively the effect
increasing each parameter would have on efficiency). In a system with a denuder collection
efficiency < 100%, the collection efficiency must be known to accurately attribute adsorbed
organics from denuder breakthrough to the gas phase and adsorbed organics volatilized from
collected particles to the particle phase. In calculating the overall phase distributions of SVOC
PAH from a denuder system, the collection efficiency for each compound is needed.
The efficiency of silicone-grease-coated denuders for the collection of polynuclear
aromatic hydrocarbons was examined by Coutant et al. (1992), who examined the effects of
uncertainties in the diffusion coefficients and in the collisional reaction efficiencies on the
overall phase distributions of SVOC PAH calculated using denuder technology. In their study,
they used a single stage, silicone-grease-coated aluminum annular denuder with a filter holder
mounted ahead of the denuder and an XAD trap deployed downstream of the denuder. In a
series of laboratory experiments, they spiked the filter with a mixture of perdeuterated PAH,
swept the system with ultra-high purity air for several hours, and then analyzed the filter and the
XAD. They found that the effects of these uncertainties, introduced by using a single compound
as a surrogate PAH (in their case, naphthalene) for validation of the denuder collection
2A-1
-------
efficiency, are less significant than normal variations because of sampling and analytical effects.
Field study results using their sampling system have not been published.
For measuring particulate phase organic compounds, the denuder-based sampling system
represents an improvement over the filter/adsorbent collection method (Turpin et al., 1993).
Some researchers, however, have reported that denuder coatings themselves can introduce
contamination (Mukerjee et al., 1997) and that the adsorbed species may be difficult to remove
from the coating (Eatough et al., 1993).
In a study conducted in southern California (Eatough et al., 1995), the Brigham Young
University Organic Sampling System (BOSS; Eatough et al., 1993) was used to determine POM
composition, and a high-volume version (BIG BOSS; flow rate 200 L/min) was used to
determine the particulate size distribution and the chemical composition of SVOC in fine
particles. The BOSS, a multi-channel diffusion denuder sampling system, consists of two
separate samplers (each operating at 35 L/min). The first sampler consists of a multi-parallel
plate diffusion denuder with charcoal-impregnated filter papers as the collection surfaces
followed by a two-stage quartz filter pack and a two-stage charcoal-impregnated filter pack. The
second sampler operating in parallel with the first consists of a two-stage quartz filter pack,
followed by the parallel plate denuder, followed by the two-stage charcoal-impregnated filter
pack. The filter samples collected by the BOSS sampler were analyzed by temperature-
programmed volatilization analysis. The second channel allows calculations of the efficiency of
the denuder in removing gas-phase specifics that would be absorbed by the charcoal impregnated
filter. Eatough et al. (1995) also operated a two-stage quartz filter pack alongside the BOSS
sampler. The BIG BOSS system (Tang et al., 1994) consists of four systems (each with a flow
rate of 200 L/min). Particle size cuts of 2.5, 0.8, and 0.4 jim are achieved by virtual impaction,
and the sample subsequently flows through a denuder, then is split, with the major flow
(150 L/min) flowing through a quartz filter followed by an XAD-II bed. The minor flow is
sampled through a quartz filter backed by a charcoal-impregnated filter paper. The samples
derived from the major flow (quartz filters and XAD-II traps) were extracted with organic
solvents and analyzed by gas chromatography (GC) and GC-mass spectroscopy. The organic
material lost from the particles was found to represent all classes of organic compounds.
Eatough et al. (1996) operated the BOSS sampler for a year at the IMPROVE site at
Canyonlands National Park, UT alongside the IMPROVE monitor and a separate sampler
2A-2
-------
consisting of a two-stage quartz filter pack. They found that concentrations of particulate carbon
determined from the quartz filter pack sampling system were lower by an average 39%, which
was attributed to volatilization losses of SVOC from the quartz filters. In another study
conducted with the BOSS in southern California, losses of 35% of the POM, on average, were
found and attributed to losses of the SVOC during sampling (Eatough et al., 1995).
The denuder used in the various BOSS samplers consists of charcoal-impregnated cellulose
fiber filter material. Denuder collection efficiencies of greater than 95% have been reported for
organic gases that adsorb on quartz and charcoal-impregnated filters (Eatough et al., 1999a; Ding
et al., 2002; Lewtas et al., 2001). However, because the mass concentration of gas phase species
that adsorb on quartz and charcoal-impregnated filters is so much greater than the mass of
semivolatile organic material in the particulate phase, it is necessary to measure and account for
the inefficiency of the denuder in the BOSS samplers. To address this problem, Brigham Young
University (BYU) developed a particle-concentrator (PC)-BOSS system (Ding et al., 2002;
Eatough et al., 1999b; Lewtas et al., 2001; Modey et al., 2001; Pang et al., 2001, 2002a,b). The
PC-BOSS includes a virtual impactor upstream of the denuder to improve the denuder collection
efficiency by removing a majority of the gases from the aerosol flow. With this system, denuder
collection efficiencies of > 99% have been reported for organic gases and other species, such as
SO2(g), HNO3(g), etc., that adsorb on quartz and charcoal-impregnated filters (Pang et al., 2001).
Since the concentrations of semivolatile organic and other gases are not altered by the virtual-
impaction concentration of accumulation mode particles (except possibly by slight changes in
pressure), it is anticipated that the gas-particle distribution will not be significantly altered by the
concentration process. The virtual impactor has a 50% cut point at 0.1 jim Da. As a result, some
particles in the 0.05 to 0.2 |im diameter size range will be removed in the major flow along with
the majority of the gases. Therefore, the mass collection efficiency of the virtual impactor
concentrator will be a function of the particle size distribution in the 0.05 to 0.1 |im size range.
This collection efficiency is measured by comparing the concentration of nonvolatile
components measured in the concentrated sample with that measured in an unconcentrated
sample. The concentration efficiency varies from 50 to 75%. It is relatively constant over
periods of weeks, but varies by season and by site, presumably as the particle size distribution
changes. Previous studies at Harvard (Sioutas et al., 1995a,b) have shown that the composition
of the sampled aerosol is little changed by the concentration process. The BYU studies listed
2A-2
-------
above have shown that the concentration efficiencies for sulfate, OC, and EC are comparable for
given sampling locations. Furthermore, the concentrations of these species and of fine
particulate nitrate determined using the PC-BOSS have been shown to be comparable to those
determined using more conventional samplers for sulfate or EC or using simpler denuder
systems for OC and nitrate.
Ding et al. (1998a) developed a method for the determination of total n-nitroso compounds
in air samples and used the method to examine organic compounds formed from NOX chemistry
in Provo, UT (Ding et al., 1998b). In their method, n-nitroso compounds are selectively
decomposed to yield nitric oxide, which is then detected using chemiluminescence. Using the
Provo samples, they found that the majority of the n-nitroso and nitrite organic compounds that
were present in fine particulate matter were SVOCs that could be evaporated from the particles
during sampling. They found particulate n-nitroso compound concentrations ranging between
< 1 and 3 nmoles/m3 and gas-phase n-nitroso compound concentrations in the same range.
Particulate organic nitrite concentrations were found in the range of <1 to «5 nmoles/m3, and
gas-phase concentrations as high as 10 nmoles/m3 were found.
Turpin et al. (1993) developed a sampling system that corrects for the loss of SVOCs
during sampling by removing most of the gas phase material from the particles in a diffusion
separator sampling system. Unlike the previously mentioned systems, wherein the particulate
phase is measured directly, in the system of Turpin et al. (1993), the gas-phase is measured
directly. In the laminar flow system, ambient, particle-laden air enters the sampler as an annular
flow. Clean, particle-free air is pushed through the core inlet of the separator. The clean air and
ambient aerosol join downstream of the core inlet section, and flow parallel to each other
through the diffusion zone. Because of the much higher diffusivities for gases compared to
particles, the SVOC in the ambient air diffuses to the clean, core flow. The aerosol exits the
separator in the annular flow, and the core flow exiting the separator now contains a known
fraction of the ambient SVOC. Downstream of the diffusion separator, the core exit flow goes
into a polyurethane foam (PUF) plug where the SVOC is collected. The adsorbed gas phase on
the PUF plug is extracted with supercritical fluid CO2 and then analyzed by gas
chromatography/mass-selective detection (GC/MSD). The gas-phase SVOC is thus determined.
Ultimately, to determine particulate phase SVOC concentrations, the total compound
concentration also has to be measured and the particulate phase obtained by difference.
2A-4
-------
The system was tested for the collection of PAHs. The diffusional transport of gas-phase PAHs
and particle concentrations agreed well with theory. Breakthrough was problematic for low
molecular weight PAHs (mol. wt. < 160). Detection limits ranged from 20 to 50 pg of injected
mass for all PAHs.
Gundel et al. (1995) recently developed a technique for the direct determination of phase
distributions of semivolatile polycyclic aromatic hydrocarbons using annular denuder
technology. The method, called the integrated organic vapor/particle sampler (IOVPS), uses a
cyclone inlet with a 50% cut point of 2.5 jim at a sampling rate of 10 L/min. The airstream then
goes through two or three sandblasted glass annular denuders that are coated with ground
adsorbent resin material (XAD-4 was initially examined) that traps vapor-phase organics. The
airstream subsequently passes through a filter, followed by a backup denuder. The denuder
collection efficiency is high and compares well with predictions based on the diffusivity of the
compounds. The denuder can also be extracted to obtain gas-phase concentrations directly
(Gundel and Lane, 1999). Particle-phase PAHs are taken to be the sum of material on the filter
and XAD adsorbent downstream after correction for denuder collection efficiency. The IOVPS
was tested for sampling semivolatile PAH in laboratory indoor air and in environmental tobacco
smoke (ETS). After exposure, the denuders, filters, and sorbent traps were extracted with
cyclohexane (Gundel et al., 1995) and analyzed for PAHs from naphthalene to chrysene using
dual-fluorescence detection (Mahanama et al., 1994). Recoveries from both denuders and filters
were -70% for 30 samples. Detection limits (defined as three times the standard deviation of the
blanks) for gas-phase SVOC PAHs ranged from 0.06 ng for anthracene to 19 ng for
2-methylnaphthalene. The 95% confidence interval (CI) for reproduction of an internal standard
concentration was 6.5% of the mean value. Relative precision, from a propagation of errors
analysis or from the 95% CI from replicate analyses of standard reference material SRM 1649
(urban dust/organics), was 12% on average (8% for naphthalene to 22% for fluorene). Sources
of error included sampling flow rate, internal standard concentration, and coeluting peaks.
Gundel and Lane (1999) reported that roughly two-thirds of particulate PAH fluoranthene,
pyrene, benz[a]anthracene, and chrysene were found on the postfilter denuders, so that it is
likely that considerable desorption from the collected particles took place.
2A-5
-------
Solid adsorbent-based denuder systems have been investigated by other researchers as
well. Bertoni et al. (1984) described the development of a charcoal-based denuder system for
the collection of organic vapors. Risse et al. (1996) developed a diffusion denuder system to
sample aromatic hydrocarbons. In their system, denuder tubes with charcoal coating and
charcoal paper precede a filter pack for particulate collection and an adsorption tube to capture
particle blow-off from the filter sample. Breakthrough curves for benzene, toluene, ortho-
xylene, and meta-xylene were developed for 60-, 90-, and 120-cm denuder tubes. The effects of
relative humidity on the adsorption capacities of the denuder system were examined, and it was
found that the capacity of the charcoal was not affected significantly by increases in relative
humidity. The feasibility of outdoor air sampling with the system was demonstrated.
Krieger and Kites (1992) designed a diffusion denuder system that uses capillary gas
chromatographic columns as the tubes for SVOC collection. The denuder was followed by a
filter to collect particles, which in turn was followed by a PUF plug to collect organic material
volatilizing off the filter. Denuder samples were analyzed by liquid solvent extraction (CH2C12)
followed by GC-mass spectrometric analysis. The PUF plugs and filters were fluid extracted
with supercritical N2O. Using this system, an indoor air sample was found to contain primarily
chlorinated biphenyls, ranging from trichlorobiphenyls (vapor pressures 10~3 to 10~4 Torr at
25 °C) to octachlorobiphenyls (10~6 - 10"7 Torr). This demonstrated that the sampler collects
compounds with a wide range of volatility. They also found that on-line desorption is successful
in maintaining good chromatographic peak shape and resolution. The entire method, from
sample collection to the end of the chromatographic separation, took 2 hours.
Organic acids in both the vapor and particulate phases may be important contributors to
ambient acidity, as well as representing an important fraction of organic PM. Lawrence and
Koutrakis (1996a,b) used a modified Harvard/EPA annular denuder system (HEADS) to sample
both gas and particulate phase organic acids in Philadelphia, PA in the summer of 1992. The
HEADS sampler inlet had a 2. l-|im cut point impactor (at 10 L/min), followed by two denuder
tubes, and finally a Teflon filter. The first denuder tube was coated with potassium hydroxide
(KOH) to trap gas phase organic acids. The second denuder tube was coated with citric acid to
remove ammonia to avoid neutralizing particle phase acids collected on the filter. The KOH-
coated denuder tube was reported to collect gas phase formic and acetic acids at better than
98.5% efficiency and with precisions of 5% or better (Lawrence and Koutrakis, 1994). It was
2A-6
-------
noted that for future field measurements of particulate organic acids, a Na2CO3-coated filter
should be deployed downstream of the Teflon filter to trap organic acids that may evaporate
from the Teflon filter during sampling.
Role of the Collection Media
The role of the collection media was recently examined in a study conducted in Seattle
(Lewtas et al., 2001). In that study, the influence of denuder sampling methods and filter
collection media on the measurement of SVOC associated with PM2 5 was evaluated. Activated
carbon and XAD collection media were used in diffusion denuders and impregnated back-up
filters in two different samplers, the Versatile Air Pollution Sampler (VAPS) and the PC-BOSS.
XAD-coated glass annular denuders and charcoal-impregnated cellulose fiber (GIF) filter
denuders also were used. GIF filters were also compared to XAD-coated quartz filters as backup
filter collection media. Lewtas et al. (2001) found that the two denuder types resulted in an
equivalent measurement of parti culate OC and particle mass. The carbon-coated denuders in the
BOSS sampler were more efficient than the XAD-coated denuders for the collection of the more
highly volatile organic compounds (MHVOCs). Lewtas et al. (2001) concluded that this
MHVOC fraction that is collected in the carbon-coated BOSS denuder does not contribute
substantially to the particle mass or to the SVOC measured as OC on quartz filters. However,
this MHVOC fraction would be captured in the carbon-impregnated filters placed behind the
quartz filters, so that, in the XAD denuder configuration, the captured MHVOC would cause a
higher OC concentration and an overestimation of the SVOC.
Some of the recent research in denuder technology has also focused on reducing the size of
the denuder, optimizing the residence time in the denuder, understanding the effect of diffusion
denuders on the positive quartz filter artifact, identifying changes in chemical composition that
occur during sampling, determining the effects of changes in temperature and relative humidity,
and identifying possible losses by absorption in coatings.
Reducing the Size of Denuders
The typical denuder configuration is an annular diffusion denuder tube of significant length
(e.g., 26.5 cm for 10 L/min [Koutrakis et al., 1988a,b]). A more compact design based on a
honeycomb configuration was shown to significantly increase the capacity (Koutrakis et al.,
2A-7
-------
1993). However, in intercomparisons with an annular denuder/filter pack system (Koutrakis
et al., 1993), significant losses of ammonia and nitric acid were observed for the honeycomb
configuration and were attributed to the large inlet surface area and long sample residence time
of the honeycomb design relative to the annular denuder system. Sioutas et al. (1996a)
subsequently designed a modified glass honeycomb denuder/filter pack sampler (HDS) with an
inlet that minimizes vapor losses on the inlet surfaces. The modified HDS has reduced inlet
surfaces and decreased residence time for sampled gases (NH3 and HNO3) compared to its
predecessor (Sioutas et al., 1994d). Sioutas et al. (1996b) also tested various inlet materials
(glass, PFA, and polytetrafluoroethylene [PTFE]) in laboratory tests and found that a PTFE
Teflon coated inlet minimized loss of sampled gases (1 to 8% loss of HNO3 observed, and -4 to
2% loss of NH3 observed). The highest inlet losses were observed for HNO3 lost to PFA
surfaces (14 to 25%). The modified HDS was tested in laboratory and field tests and found to
agree within 10% with the annular denuder system.
Residence Time in the Denuder
The efficiency of a diffusion denuder sampler for the removal of gas phase material can be
improved by increasing the residence time of the sampled aerosol in the denuder. However, the
residence time can only be increased within certain limits. Because the diffusion denuder
reduces the concentration of gas-phase semivolatile organic material, semivolatile organic matter
present in the particles passing through the denuder will be in a thermodynamically unstable
environment and will tend to outgas SVOC during passage through the denuder. The residence
time of the aerosol in the denuder, therefore, should be short enough to prevent significant loss
of particulate-phase SVOC to the denuder. Various studies have suggested that the residence
time in the denuder should be less than about 2 s (Gundel and Lane, 1999; Kamens and Coe,
1997; Kamens et al., 1995). The residence times in the various denuder designs described by
Gundel and Lane (1999) are from 1.5 to 0.2 s. The equilibria and evaporation rates are not as
well understood for organic components as they are for NH4NO3 (Zhang and McMurry, 1987,
1992; Hering and Cass, 1999).
2A-8
-------
Effect of Diffusion Denuders on the Positive Quartz Filter Artifact
The adsorption of organic compounds by a second quartz filter has been shown to be
reduced, but not eliminated, in samples collected in the Los Angeles Basin using a multi-channel
diffusion denuder with quartz filter material as the denuder collection surface preceding the
quartz filters (Fitz, 1990). This artifact can be further reduced by using activated charcoal as the
denuder surface and using a particle concentrator to reduce the amount of gas phase organic
compounds relative to condensed phase organic compounds (Cui et al., 1997, 1998; Eatough,
1999). Other experiments (Gotham and Bidleman, 1992; Cui et al., 1998; Eatough et al., 1995,
1996) have shown that the quartz filter artifact can result both from the collection of gas phase
organic compounds and from the collection of SVOCs lost from particles during sampling.
Thus, available results suggest that both a "positive" and a "negative" artifact can occur with the
determination of particulate phase organic compounds using two tandem quartz filters.
The importance of the adsorption of organic vapors on PM or filters relative to the
volatilization of organic compounds from PM collected on a filter continues to be a topic of
active debate. The relative importance of positive and negative artifacts differ for denuded and
undenuded filters; depend on face velocity, sample loading, and the vapor pressures of the
compounds of interest; and may vary with season and location because of variations in the
composition of volatile and semivolatile organic material. Evidence exists for substantial
positive and negative artifacts in the collection of organic PM.
Changes in Chemical Composition During Sampling
The use of sampling systems designed to correctly identify the atmospheric gas and
particulate phase distributions of collected organic material has been outlined above.
An additional sampling artifact that has received little consideration in the collection of
atmospheric samples is the potential alteration of organic compounds by the sampling process
itself. These alterations appear to result from the movement of ambient air containing oxidants
and other reactive compounds past the collected particles. The addition of NO2 (< 1 ppm) or O3
(< 200 ppb) to the sampled air stream (at 0 to 5 °C) for a high-volume sampler reduced the
concentrations of benzo[a]pyrene and benzo[a]anthracene from a few percent to 38%, with the
observed reduction increasing with increased concentration of the added gases (Brorstrom et al.,
1983). Spiking a filter with an amine increased the measured concentrations of nitrosamines in
2A-9
-------
both the filter and following XAD sorbent bed for a mid-volume sampler (Ding et al., 1998a,b).
Similar results have been obtained for the exposure of a deuterated amine on a filter to NOX
(Pellizzari and Krost, 1984). When Tenax columns spiked with deuterated styrene and
cyclohexene were exposed to ppm concentrations of ozone or halogens, oxygenated and
halogenated compounds were formed (Pellizzari and Krost, 1984). Similar oxidation of
aldehydes and peroxyacetylnitrate (PAN) during sampling has been observed (Grosjean and
Parmar, 1990). Collected PAH compounds can be oxygenated or nitrated on a filter (Davis
et al., 1987; Lindskog et al., 1987), but 1-nitropyrene has been shown to be resistant to additional
nitration (Grosjean, 1983). These various chemical transformations of collected organic
compounds can be eliminated by removing the gas phase oxidants, NOX, HNO3, etc., by reaction
or adsorption prior to particle collection (Ding, 1998a,b; Grosjean and Parmar, 1990; Parmar and
Grosjean, 1990; Pellizzari and Krost, 1984; Williams and Grosjean, 1990). The BOSS denuder
should be effective in eliminating most of the chemical transformation artifacts because reactive
gases are removed by the charcoal denuder that precedes the particle collection filter.
Temperature and Relative Humidity Effects
The problems of sampling artifacts associated with SVOC adsorption and evaporation are
compounded by temperature and relative humidity (RH) effects (Pankow and Bidleman, 1991;
Pankow et al., 1993; Falconer et al., 1995; Goss and Eisenreich, 1997). The effects of
temperature on the partitioning of PAHs were examined by Yamasaki et al. (1982), who found
that the partition coefficient (PAI-^/PAHp.^) was inversely related to temperature and could be
described using the Langmuir adsorption concept. The dissociation of ammonium nitrate aerosol
is also a function of temperature. Bunz et al. (1996) examined the dissociation and subsequent
redistribution of NH4NO3 within a bimodal distribution using a nine-stage low-pressure Berner
impactor followed by ion chromatography analysis and found a strong temperature dependency
on the redistribution. Bunz et al. (1996) found that at lower temperatures (below 10 °C) there
was little change in the aerosol size distribution. At temperatures between 25 and 45 °C,
however, the lifetime of NH4NO3 particles decreased by more than a factor of 10, and size
redistribution, as measured by average ending particle diameter, increased more for higher
temperatures than for lower temperatures.
2 A-10
-------
The effects of relative humidity on the sorption of S VOC on particles are not well
understood. In a series of laboratory experiments, Goss and Eisenreich (1997) examined the
sorption of both nonpolar (hydrocarbons and chlorinated hydrocarbons) and polar (ethyl ether
and acetone) VOCs onto combustion soot particles as a function of temperature and relative
humidity. The soot particles used in their experiments were collected from oil furnaces and
contained 60% (w/w) iron sulfate (water-soluble fraction) and 9% (w/w) EC and OC. They
found that, for all compounds, the sorption of VOCs onto soot particles decreased with
increasing relative humidity over the range of 10 to 95%. They also observed hysteresis in the
relative humidity dependency, with sorption coefficients at a given relative humidity higher
when the RH is being increased than when the RH is being decreased. The sorption coefficients
were fit with an exponential function to the RH so that the slope of the regression line would
provide a measure of the influence of relative humidity. Based on the magnitude of the slope,
they concluded that the RH-dependency of sorption was stronger for water-soluble organic
compounds.
In another study by Jang and Kamens (1998), humidity effects on SVOC gas-particle
partitioning were examined using outdoor environmental chambers and the experimentally
determined partitioning coefficients were compared to theoretical values. They examined the
partitioning of SVOC onto wood soot, diesel soot, and secondary aerosols and concluded that
"the humidity effect on partitioning was most significant for hydrophobic compounds adsorbing
onto polar aerosols." Although these two studies seem to be contradictory, on closer
examination, it is difficult to compare the two studies for several reasons. The experiments
conducted by Jang and Kamens (1998) were conducted in outdoor chambers at ambient
temperatures and humidities. Their model was for absorptive partitioning of SVOC on
liquid-like atmospheric PM. In contrast, the results of Goss and Eisenreich (1997) were obtained
using a GC system operated at 70 °C higher than ambient conditions. Goss and Eisenreich
(1997) modeled adsorptive partitioning of VOC on solid-like atmospheric PM. In the study of
Jang and Kamens (1998), calculated theoretical values for water activity coefficients for diesel
soot were based on an inorganic salt content of 1 to 2%; whereas, the combustion particles
studied by Goss and Eisenreich (1997) contained 60% water-soluble, inorganic salt content.
Jang and Kamens (1998) obtained their diesel soot from their outdoor chamber, extracted it with
organic solvent (mixtures of hexane and methylene chloride), and measured the organic fraction.
2 A-11
-------
The resulting salt content of 2% of the PM studied in Jang and Kamens (1998) is enough to
affect water uptake but presumably not enough to affect the sorption partitioning of organics.
Impactor Coatings
Impactors are used to achieve a size cut point and to collect particles on surfaces. Particles
collected on impactors are exposed to smaller pressure drops than filter-collected particles,
making them less susceptible to volatile losses (Zhang and McMurry, 1987). However, size
resolution can be affected by bounce when samples are collected at low humidities (Stein et al.,
1994). Other sources of error inherent in some currently acceptable practices could potentially
affect PM concentration measurements and will likely become even more important as more
emphasis is placed on chemical speciation. Allen et al. (1999a) reported that the practice of
greasing impaction substrates may introduce an artifact from the absorption of semivolatile
species from the gas phase by the grease, because the grease could artificially increase the
amount of PAHs and other organic compounds attributed to the aerosol. Allen et al. (1999b)
offered several criteria to ensure that this absorption artifact is negligible, including selecting
impaction oils in which analytes of interest are negligibly soluble and ensuring that species do
not have time to equilibrate between the vapor and oil phases (this criterion is met for
nonvolatile species). They recommend using oiled impaction substrates only if the absorption
artifact is negligible as determined from these criteria. The application of greases and impaction
oils for preventing or reducing bounce when sampling with impactors is not suitable for carbon
analysis, because the greases contain carbon (Vasilou et al., 1999).
Kavouras and Koutrakis (2001) investigated the use of polyurethane foam (PUF) as a
substrate for conventional inertial impactors. The PUF impactor substrate is not rigid like the
traditional impactor substrate, so particle bounce and reentrainment artifacts are reduced
significantly. Kavouras and Koutrakis (2001) found that the PUF impaction substrate resulted in
a much smaller 50% cut point at the same flow rate and Reynolds number. Moreover, the lower
50% cut point was obtained at a lower pressure drop than with the conventional substrate, which
could reduce artifact vaporization of semivolatile components.
2 A-12
-------
APPENDIX 2B. ANALYTICAL TECHNIQUES
2B.1 INORGANIC ELEMENTS
2B.1.1 Energy Dispersive X-Ray Fluorescence (EDXRF)
EDXRF has usually been the method of choice for analysis of trace elements on filters.
EDXRF is preferred for aerosol analysis over wavelength dispersive X-ray fluorescence (XRF)
because its fast analysis over the total spectrum allows simultaneously analysis of numerous
elements. EDXRF can accommodate small sample sizes and requires little sample preparation
or operator time after the samples are placed in the analyzer. It also leaves the sample intact
after analysis, making further analysis possible. XRF irradiates a uniform particle deposit on the
surface of a membrane filter with 1 to 50 kev X-rays that eject inner shell electrons from the
atoms of each element in the sample (Dzubay and Stevens, 1975; Jaklevic et al., 1977; Billiet
et al., 1980; Potts and Webb, 1992; Piorek, 1994; Bacon et al., 1995; deBoer et al., 1995;
Holyiiska et al., 1997; Torok et al., 1998; Watson et al., 1999). When a higher energy electron
drops into the vacant lower energy orbital, a fluorescent x-ray photon is released. The energy of
this photon is unique to each element, and the number of photons is proportional to the
concentration of the element. Concentrations are quantified by comparing photon counts for a
sample with those obtained from thin-film standards of known concentration (Dane et al., 1996).
The previous 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a) included a
detailed discussion of EDXRF.
Emitted x-rays with energies less than ~4 kev (affecting the elements sodium, magnesium,
aluminum, silicon, phosphorus, sulfur, chlorine, and potassium) are absorbed by the filter, by a
thick particle deposit, or even by large particles in which these elements are contained. Very
thick filters also scatter much of the excitation radiation or protons, thereby lowering the
signal-to-noise ratio for XRF and proton (or particle) induced X-ray emission (PIXE). For this
reason, thin membrane filters with deposits in the range of 10 to 50 jig/cm2 provide the best
accuracy and precision for XRF and PIXE analysis (Davis et al., 1977; Haupt et al., 1995).
2B-1
-------
2B.1.2 Synchrotron Induced X-ray Fluorescence (S-XRF)
S-XRF is a form of EDXRF in which the exciting X-rays are generated in a synchrotron.
Bremmstrahlung X-rays are generated when energetic electrons (generally several GeV in
energy) are forced by a magnetic field to make a bend in their path. The advantages of the
technique are that an extremely high flux of X-rays may be obtained and that the X-rays are
100% polarized in the plane of the electron beam. The former allows for X-ray beams generally
of 50 to 500 |im in diameter. However, the beams can be focused into X-ray microprobes, with
spot sizes on the order of 1 |im diameter. The X-ray polarization allows for removal of most of
the background normally found under characteristic X-ray peaks, greatly improving sensitivity
compared to other XRF techniques. The primary disadvantages are the limited number of
synchrotrons, and that few synchrotrons have S-XRF capabilities. Thus, the technique has been
rarely used for PM, and then generally only for special problems, such as for analysis of the
smoke from the Kuwaiti oil fires (Cahill et al., 1992; Reid et al., 1994). However, with the
increasing availability of S-XRF facilities dedicated to PM analysis (the first of which was the
Advanced Light Source opened at Lawrence Berkeley National Laboratory last year), utilization
of S-XRF for PM analysis is increasing.
2B.1.3 Proton (or Particle) Induced X-ray Emission (PIXE)
PIXE differs from XRF analysis in the excitation source for producing fluorescence. The
filter deposit is bombarded with high-energy protons to remove inner shell electrons, and the
resulting characteristic X-rays are analyzed as in XRF (Johansson, 1970; Cahill, 1981, 1985,
1990; Zeng et al., 1993). Small accelerators, generally Van de Graaffs, generate intense beams
of low energy protons, usually of a few MeV in energy. These have the ability to remove
electrons from the inner shells of atoms of any element. Thus, PIXE can measure a very wide
range of elements in a single analysis. The cross section for producing X-rays using protons of a
few MeV in energy tends to favor lighter elements, Na through Ca, but sensitivities for
equivalent PIXE and multi-wavelength XRF analysis are roughly comparable. The technique
has been widely used in the United States (Flocchini et al., 1976; Malm et al., 1994) and around
the world, as many universities have the small accelerators needed for the method. Like S-XRF,
the proton beams can be focused into |im size beams, but these have not been widely used for
PM. However, the mm-size beams used in both S-XRF and PIXE are well suited to analyzing
2B-2
-------
the limited mass and small deposits that result from detailed particle size measurements by
impactors (Perry et al., 1999).
2B.1.4 Proton Elastic Scattering Analysis (PESA)
With the routine availability of elemental analyses for all elements sodium and heavier,
organic components remain the major unmeasured species for mass balance. For programs like
IMPROVE (Malm et al., 1994), parallel filters are collected for separate OC and EC
determinations. Aerosol programs that use PIXE can directly measure hydrogen simultaneously
by scattering protons from Teflon filters, which lack hydrogen (Cahill et al., 1989, 1992). After
correction for hydrogen in ammonium sulfate and nitrate, the remaining hydrogen can be used to
infer the OC mass. Generally, analyses of organic matter by carbon combustion from quartz
filters and analysis of organic matter by hydrogen in particles on Teflon filters are in agreement,
assuming certain assumptions about the chemical states of sulfates and nitrates are met (Malm
etal., 1994; Cahill etal., 1996).
2B.1.5 Total Reflection X-Ray Fluorescence (TRXRF)
One of the limitations of the EDXRF method is its minimum detection limit, which may be
high due to high background values (Streit et al., 2000). However, by using implementation of
X-ray optical geometries that use the total reflection of the primary radiation on flat surfaces,
scattering on the substrate is lessened, thus reducing detection limits. This is the basis for the
total reflection X-ray fluorescence (TRXRF) method (Aiginger and Streli, 1997). This
modification to the EDXRF technique improves detection limits and eliminates the need to
correct for matrix effects. Despite its apparent advantages, TRXRF has not yet become widely
used for atmospheric aerosol analysis, although it has been used in the analysis of marine aerosol
(Stahlschmidt et al., 1997) and at a high elevation site (Streit et al., 2000). Streit et al. (2000)
sampled ambient air at the High Alpine Research Station (3,580 m above sealevel) in the
Bernese Alps, Switzerland, using a nine-stage, single-jet, low-pressure, cascade impactor
equipped with quartz impactor plates that were coated with silicon oil diluted in 2-propanol. The
typical sample volume for a weekly sample was 10 m3. The quartz plates were analyzed directly
by TRXRF. Streit et al. reported that the minimum detection limits, defined by the 3a values of
the blanks, ranged from 25 ng for S, decreased monotonically with increasing atomic number
2B-2
-------
down to 5 pg for Rb, and decreased after that. The use of TRXRF is expected to increase as
EDXRF users become more aware of the method. A relatively low-cost, add-on unit has been
developed that would allow EDXRF users to test the TRXRF technique (Aiginger, 1991).
2B.1.6 Instrumental Neutron Activation Analysis (INAA)
INAA irradiates a sample in the core of a nuclear reactor for few minutes to several hours,
depending on the elements being quantified (Dams et al., 1970; Zoller and Gordon, 1970;
Nadkarni, 1975; Landsberger, 1988; Olmez, 1989; Ondov andDivita, 1993). The neutron
bombardment transforms many elements into radioactive isotopes. The energies of the gamma
rays emitted by these isotopes identify them and, therefore, their parent elements. Furthermore,
the intensity of these gamma rays is proportional to the amount of the parent element present in
the sample. Different irradiation times and cooling periods are used before counting with a
germanium detector. In source apportionment studies, it is possible to use a combination of XRF
and INAA to develop a relatively complete set of elemental measurements. Between these two
analytical techniques, good sensitivity is possible for many elements, including most of the toxic
metals of interest. In general, XRF provides better sensitivity for some metals (Ni, Pb, Cu, and
Fe); whereas INAA provides better sensitivity for others (Sb, As, Cr, Co, Se, and Cd). Both
methods provide similar detection limits for still other elements (V, Zn, and Mn). INAA does
not quantify some of the abundant species in ambient PM such as silicon, nickel, tin, and lead.
While INAA is technically nondestructive, sample preparation involves folding the sample
tightly and sealing it in plastic, and the irradiation process makes the filter membrane brittle and
radioactive. These factors limit the use of the sample for subsequent analyses.
INAA has been used to examine the chemical composition of atmospheric aerosols in
several studies either as the only method of analysis or in addition to XRF (e.g., Yatin et al.,
1994; Gallorini, 1995). INAA has higher sensitivity for many trace species, and it is particularly
useful in analyzing for many trace metals. Landsberger and Wu (1993) used INAA to analyze
air samples collected near Lake Ontario for Sb, As, Cd, In, I, Mo, Si, and V using INAA. They
and found that using INAA in conjunction with epithermal neutrons and Compton suppression
produces very precise values with relatively low detection limits.
Enriched rare-earth isotopes have been analyzed via INAA and used to trace sources of PM
from a coal-fired power plant (Ondov et al., 1992), from various sources in the San Joaquin
2B-4
-------
Valley (Ondov, 1996), from intentially tagged (indium) diesel emissions from sanitation trucks
(Suarez et al., 1996; Wu et al., 1998), and from iridium-tagged emissions from school buses
(Wuetal., 1998).
An intercomparison was conducted in which 18 pairs of filters were sent to participants in
the Coordinated Research Program (CRP) on Applied Research on Waste Using Nuclear Related
Analytical Techniques (Landsberger et al., 1997). As part of that study, participants used PIXE,
INAA, XRF, or AAS to analyze the samples. Many of the results for XRF and PIXE in the
coarse fraction were observed to be biased low compared to INAA. The authors speculated that
self-attenuation of the X-rays resulting from the particle size effect could create a systematic
error.
2B.1.7 Atomic Absorption Spectrophotometry (AAS)
AAS is applied to the residue of a filter extracted in a strong solvent to dissolve the solid
material; the filter or a portion of it is also dissolved during this process (Ranweiler and Moyers,
1974; Fernandez, 1989; Jackson and Mahmood, 1994; Chow et al., 2000). A few milliliters of
the extract are injected into a flame where the ions are reduced to elements and vaporized.
Elements absorb light at certain wavelengths in the visible spectrum, and a light beam with
wavelengths specific to the elements being measured is directed through the flame to be detected
by a monochrometer. The light absorbed by the flame containing the extract is compared with
the absorption from known standards to quantify the elemental concentrations. AAS requires an
individual analysis for each element, and a large filter or several filters are needed to obtain
concentrations for a large variety of elements. AAS is a useful complement to other methods
(such as XRF and PIXE) for species such as beryllium, sodium, and magnesium, which are not
well quantified by these methods. Airborne particles are chemically complex and do not
dissolve easily into complete solution, regardless of the strength of the solvent. Thus, insoluble
residues may be left behind, and soluble species may co-precipitate on them or on container
walls.
AAS was used to characterize the atmospheric deposition of trace elements Zn, Ni, Cr, Cd,
Pb, and Hg to the Rouge River watershed by particulate deposition (Pirrone and Keeler, 1996).
The modeled deposition rates were compared to annual emissions of trace elements that were
estimated from the emissions inventory for coal and oil combustion utilities, iron and steel
2B-5
-------
manufacturing, metal production, cement manufacturing, and solid waste and sewage sludge
incinerators. They found generally good agreement between the trend observed in atmospheric
inputs to the river (dry + wet deposition) and annual emissions of trace elements, with R2s
varying from «0.84 to 0.98. Both atmospheric inputs and emissions for Pb showed downward
trends. For the period of 1987 to 1992, steady increases were observed for Cd (major sources
are municipal solid waste incineration, coal combustion, sludge incineration, and iron and steel
manufacturing), Cr and Ni (major sources are iron and steel production and coal combustion),
and Hg (major sources are coal, the contribution from which had decreased from 53 to 45%, as
well as municipal, solid, and medical waste incineration, the contribution from which has
increased).
2B.1.8 Inductively Coupled Plasma with Atomic Emission Spectroscopy
(ICP-AES)
ICP-AES introduces an extracted sample into an atmosphere of argon gas seeded with free
electrons induced by high voltage from a surrounding Tesla coil (Fassel and Kniseley, 1974;
McQuaker et al., 1979; Lynch et al., 1980; Harman, 1989; Tyler, 1992; Baldwin et al., 1994).
The high temperatures in the induced plasma raise valence electrons above their normally stable
states. When these electrons return to their stable states, a photon of light is emitted that is
unique to the element that was excited. This light is detected at specified wavelengths to identify
the elements in the sample. ICP-AES can determine a large number of elemental concentrations
using small sample volumes with acceptable detection limits for atmospheric samples. As with
AAS, this method requires complete extraction and destruction of the sample.
2B.1.9 Inductively Coupled Plasma with Mass Spectroscopy (ICP-MS)
ICP-MS has been applied in the analysis of personal exposure samples (Tan and Horlick,
1986; Gray and Williams, 1987a,b; Nam et al., 1993; Munksgaard and Parry, 1998; Campbell
and Humayun, 1999). Ion species generated from ICP and from the sample matrix can produce a
significant background at certain masses resulting in the formation of polyatomic ions that can
limit the ability of ICP-MS to determine some elements of interest. Cool plasma techniques
have demonstrated the potential to detect elements at the ultra-trace level (Nham et al., 1996)
and to minimize common molecular ion interferences (Sakata and Kawabata, 1994; Turner,
2B-6
-------
1994; Plantz, 1996). Detection limits of TCP-MS using a one-second scan are typically in the
range of 10"3 ng/m3, which is an order of magnitude lower than other elemental analysis methods.
The instrument can also be set up to analyze a wide dynamic range of aerosol concentrations.
Isotope analysis can also be performed with ICP-MS. Intercomparison studies are needed to
establish the comparability of ICP-MS with other nondestructive filter analysis methods.
Keeler and Pirrone (1996) used ICP-MS to determine trace elements Cd, Mn, V, As, Se,
and Pb in atmospheric fine particulate (PM2 5) and total suspended particulate samples collected
in two Detroit sites. The results were used in a deposition model with additional measurements
using AAS to estimate the dry deposition flux of trace elements to Lake Erie.
2B.1.10 Scanning Electron Microscopy (SEM)
Mamane et al. (2001) investigated the use of computer-controlled scanning electron
microscopy (CCSEM) as a way of supplementing XRF analysis and providing automated
analysis of particle size, chemistry, and particle classification. An ambient coarse particulate
sample from Baltimore was collected on a polycarbonate filter for this analysis. CCSEM
analyses were conducted for 2,819 particles in 78 randomly selected fields of view during an
unattended 8-h run. Mamane et al. (2001) confirmed the stability of the CCSEM instrument
over several hours of operation. The physical properties of the sample such as particle diameter,
mass loading per field, and particle number per field were well represented by analyzing
approximately 360 particles, and little additional information was gained by analyzing more
particles. Teflon filters are not well suited for SEM analyses. Unfortunately, analysis of fine
PM is expected to pose analytical challenges not addressed in the present study (Mamane et al.,
2001).
Nelson et al. (2000) applied Raman chemical imaging and SEM (Raman/SEM) to study the
size, morphology, elemental and molecular composition, and molecular structure of fine PM.
In their study, filter compatibility was examined, and Raman/SEM chemical imaging was
conducted for several standard materials as well as for ambient PM2 5 samples. Polycarbonate
was determined to be a suitable substrate for both SEM and Raman chemical imaging analysis.
Conner et al. (2001) used CCSEM with individual X-ray analysis to study the chemical and
physical attributes of indoor and outdoor aerosols collected around a retirement home in
Baltimore. The CCSEM technique was demonstrated to be capable of identifying spherical
2B-7
-------
particles typical of combustion or other high temperature (presumably industrial) processes as
well as pollens and spores. Indoor particles originating from cosmetics were also identified.
2B.2 ORGANIC AND INORGANIC CARBON
Large scale efforts to characterize carbonaceous aerosol require cost effective methods that
can analyze samples rapidly. Commercial thermal-optical (TO) instruments were developed to
serve this need. The IMPROVE and STN networks have employed these instruments to
accumulate large data sets, including measurements taken over the past 18 years by the
IMPROVE network and over the past 3 years by STN. In addition to the protocols developed for
the IMPROVE and STN networks, a number of alternative TO-based protocols and techniques
have been employed by the academic research community. Protocols vary in temperature range
and step size, in the duration of heating at each step, in the timing for introduction of oxygen for
the conversion of black carbon into CO2, in the use of a catalyst to reduce pyrolysis, and in the
assignment of OC and EC fractions. These operational differences have complicated efforts to
compare and combine data sets from studies using different TO protocols.
Thermal-Optical Reflectance
The thermal-optical reflectance (TOR) method of carbon analysis developed by Huntzicker
et al. (1982) has been adapted by several laboratories for the quantification of OC and EC in PM
collected on quartz-fiber filters. Although the principle applied by these laboratories is identical
to that of Huntzicker et al. (1982), the details differed with respect to calibration standards,
analysis time, temperature ramping, and volatilization/combustion temperature. The IMPROVE
network employs a version of the TOR method for its OC/EC analyses.
In the most commonly used version of the TOR method (Chow et al., 1993), a punch from
the filter sample is heated to temperatures ranging from ambient levels to 550 °C in a pure
helium atmosphere. In principle, the OC fraction of the PM contained in the filter punch will
vaporize, leaving behind only refractory EC. The organic carbon that evolves at each
temperature step is first oxidized to CO2, then converted to methane and finally quantified with a
flame ionization detector (FID). The filter punch is incubated at 550 °C for a period sufficient to
allow the flame ionization signal to return to its baseline value. The punch is then exposed to a
2B-8
-------
2% oxygen and 98% helium atmosphere and heated from 550 °C to 800 °C with several
temperature ramping steps. The reflectance from the deposit side of the filter punch is
monitored throughout the analysis. This reflectance decreases during OC volatilization in the
helium atmosphere owing to the pyrolysis of nonvolatile organic material present in the filter
punch. When oxygen is added, the reflectance increases as light-absorbing carbon is combusted
and removed. It is assumed that the first light-absorbing carbon compounds to combust are
pyrolysis artifacts from the first phase of the analysis. Therefore, the OC mass is defined as that
fraction that evolves up to the introduction of oxygen plus the oxidized carbon that evolves up to
the point when the filter reflectance signal is restored to its preanalysis value. Once the original
reflectance level is re-attained, all further carbon evolving from the sample punch is assigned to
the EC fraction. Accordingly, "organic carbon" (OC) is carbon that does not absorb light at the
laser wavelength (632.8 nm) typically used by TOR instruments, and all other carbon is defined
as "elemental carbon" (EC).
Thermal-Optical Transmission
The primary difference between TOR and thermal-optical transmission (TOT) methods is
in the choice of absorption detection: light transmission through the filter punch, rather than its
reflectance, is monitored throughout the analysis. The TOT method of Birch and Gary (1996)
also uses a pure helium atmosphere for volatilizing OC, but the second stage involves a higher
oxygen/helium (10%) gas mixture to oxidize the black carbon remaining on the filter punch.
The temperature is raised to approximately 820 °C in the helium phase, during which both
organic and carbonate carbon are volatilized from the filter. In the second stage, the oven
temperature is reduced, then raised to about 860 °C. During this stage, pyrolysis correction and
the EC measurement is made. Figure 2B-1 is an example of a TOT thermogram, showing
temperature, transmittance, and FID response traces. The peaks that correspond to the
concentrations of CO2 that evolve from the filter punch during the course of the analysis are
assigned to OC, carbonate carbon (CC), pyrolitic carbon (PC), and EC. The high temperature in
the first stage of the TOT thermal profile is included in order to decompose CC and to volatilize
very high-boiling organic compounds. However, many organic carbon compounds will pyrolyze
at this temperature to generate PC. The ability to quantify PC is particularly important in high
OC/EC regions such as wood smoke-impacted air sheds. Wood smoke aerosol contains many
2B-9
-------
OC - EC split
Time (min)
Figure 2B-1. Thermogram for a sample containing rock dust (a carbonate source) and
diesel exhaust, showing three traces that correspond to temperature, filter
transmittance, and Flame lonization Detector (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).
complex compounds that generate substantial PC. Significant error in the EC fraction can result
in the absence of a careful PC correction.
Informal intercomparisons among different filter transmission methods have shown high
correlations of absorption, but with differences of up to a factor of two in absolute values
(Watson et al., 1988a,b). These differences are functions of the type of filter, filter loading, the
chemical and physical nature of the deposit, the wavelengths of light used, calibration standards,
and light diffusing methods. Currently, there is no agreement on which combination most
accurately represents light absorption in the atmosphere.
The National Institute for Occupational Safety and Health (NIOSH) Method 5040 is based
on the TOT method (Birch and Gary, 1996). The NIOSH protocol consists of a two-stage
process with the first stage being conducted in a pure helium atmosphere at temperatures of 250,
2B-10
-------
500, 650, and 850 °C for a total of 4.5 min and the second stage conducted in a 2% oxygen/98%
helium mix at temperatures of 650, 750, 850, and 940 °C for 4 min. A pyrolysis base correction
is made based on the transmittance measurement. The EPA adopted a modified version of the
NIOSH Method 5040 protocol for use in its STN.
IMPROVE versus NIOSH protocols
Although the IMPROVE protocol and the NIOSH protocol in use by the STN network
yield closely comparable TC values, the differences in thermal profiles and operational
definitions of organic versus black carbon result in very different mass assignments to these
fractions. Other methodological differences may make data comparisons difficult, including the
different choice of absorption detection, e.g. reflectance versus transmission, the different
temperature ranges and incubation periods, and the different approaches used to account for
background OC. Examples of thermograms obtained with the IMPROVE and NIOSH protocols
are shown in Figure 2B-2.
Using both the IMPROVE and NIOSH protocols, Chow et al. (2000) analyzed 60 quartz
filter samples on a prototype reflectance/transmittance analyzer that represented a wide variety
of aerosol compositions and concentrations. The two TC data sets possessed no statistically
significant differences. However, marked differences were found in the fraction of TC that was
attributed to EC as determined by the IMPROVE versus NIOSH thermal evolution protocols.
The IMPROVE EC measurements were typically higher than the NIOSH EC measurements.
When the NIOSH protocol was modified to exclude the helium-only 850 °C temperature step,
however, the OC/EC ratios came into agreement between the two methods. Because OC and EC
are operationally defined parameters, Chow et al. (2000) pointed out that it is important to retain
ancillary information when reporting EC and OC by these analytical methods, so that
comparisons can be made among measurements taken at different sites using these two methods.
The NIOSH and IMPROVE protocols both require correction for positive organic artifacts
resulting from the absorption of background organic vapor by the heat-treated quartz filters used
for OC/EC measurements. Both the IMPROVE and STN science teams have evaluated the
presence of carbon artifacts in their measurements. The IMPROVE team has established that
heat-treated quartz filters adsorb OC vapors up to a saturation threshold over the course of a few
days in the field. The STN science team observed that positive carbon artifacts can vary with
2B-11
-------
1000
900
800
IMPROVE Method
100% He *-4
- 2% O,/98% He •
FID Temp
»«» Laser Reflect •••• Laser Trans
Reflect-Split Trans-Split
*»< _1
11*11 IT. 1
00
1
•s
o:
0 120 240 360 480 600 720 840 960108012001320144015601680
Time (sec)
1800
NIOSH Method
•
900-
•
800-
• — •
0 700-
Q)
2- 600-
3 500-
2
Q. 400-
E
|2 300-
•
200-
•
100-
0-
^—
i
...
**!
••<
j
'— r
.-"
1
L
i A*
I :
\.:
Si"
\
C1 | (
I
i
I
i
/ i
1 *
:Ai^ "I7*1
*"*/*•*•" "'"t—r
/ i 1
...,; i 1
i 1
• i
i i
A i 1
l\ I
1---JV---I-X---J- J:
^C/'^<7**N"i /!
VJ™/ I
! i
DC2| OC3| OC4 JEC1
_ ._
/ :
A^XI^'AAAAAA'
4 — ECR — »•
4— ECT — ^
^* , ,u ,^ JM ,„
EC2 | ECS |
^
1
\
cal
2
A*
'.
\
V
b 1
FID
Temp
"A Laser Reflect
• ••• Laser Trans
Reflect-Split
Trans-Split
i/r
'c
>
^
g
a>
OJ
H
S
03
00
"o
a>
5=
o
Q£
b
u>
ro
I
Q"
u_
0 120 240 360 480 600 720 840 9601080120013201440156016801800
Time (sec)
Figure 2B-2. Examples of thermograms obtained by (a) the IMPROVE protocol, and by
(b) the NIOSH protocol. Flame ionization detection is used in both examples.
FID denotes the observed signal intensity. The intervals denoted ECR/ECT,
i.e., elemental carbon reflectance/transmittance, refer to the time and
temperature intervals during which evolved CO2 is assumed to come from
oxidized elemental carbon (EC). Laser Reflect and Laser Trans refer to
laser reflectance or transmission measurement of filter "blackness" at the
He-Ne wavelength, 633 nm.
Source: Birch and Gary (1996).
2B-12
-------
sampler type. Total carbon artifacts for the samplers used by the STN range from 9.5 to 33.9%
of the carbon collected during a typical ambient measurement. Documents describing the issues
regarding blank subtraction and the current protocols used by the IMPROVE and STN networks
are available on the network websites:
IMPROVE - http://vista.cira.colostate.edu/improve
STN - http://www.epa.gov/ttn/amtic/files/ambient/pm25/spec
Pyrolitic Carbon (PC) and Other Confounders
In a methods comparison study, Countess (1990) determined that it is necessary to
minimize or correct for the PC and OC found in wood smoke and automobile exhaust samples
that pyrolyze to create interferences during the course of some thermal-optical protocols.
Pyrolysis of organic compounds occurs during analysis by both TO methods, although the
greater temperature to which samples are exposed in the NIOSH protocol is likely to produce
larger quantities of pyrolysis-derived EC. During the initial heating phase, pyrolysis is indicated
by an increase in optical density (blackness) in the filter sample. Both methods distinguish
artifact pyrolysis-derived EC from ambient EC at the point at which the transmittance or
reflectance signal is restored to the pre-heating level. The assumption made is that heating does
not alter the absorption properties of the material collected on the filter. This assumption is
reasonable if the only light absorbing species of carbon is strictly a graphite-like EC that is
unlikely to undergo a change in its absorption properties over the temperature ranges used by TO
methods. The effects of heat-induced changes to the light absorption and chemical properties of
atmospheric organic compounds undergoing TO analysis are being evaluated by NIST. NIST
has identified three assumptions that must be met in order for TO methods to reliably measure
EC: (1) absorptivity of carbonaceous PM remains constant up to the point of pyrolysis; (2) once
formed, pyrolyzed carbon (char) absorbs at the analytic wavelength and its absorptivity remains
constant within the high temperature step; (3) pyrolyzed OC has the same absorptivity as the EC
that is native to the sample. Using urban dust, forest fire emissions, and ambient laboratory
aerosol, they observed changes in the absorptivities of these materials during heating before and
during the formation of pyrolysis artifacts, up to the OC/EC split point. NIST, therefore,
recommends that standard TOR/T protocols be developed that account for these changes.
2B-13
-------
Light-absorbing carbon or black carbon (BC) includes a large number of compounds that
may be altered, not only in their light-absorbing properties but also in their oxidation chemistry.
The materials responsible for defining the original blackness of the filter sample may be altered
during the initial heating phase of TO analysis, so that the mix present during the oxidation cycle
may not be representative of the original, atmospherically derived BC. An error in determining
the mass of EC would arise, for example, if a heavy, but weakly light-absorbing, compound were
transformed into a material that possessed a higher absorption coefficient and higher resistance
to oxidation than absorbing compounds collected from the atmosphere. The pyrolized form of
this compound would resist oxidation past the OC-EC split point in the analysis, leading to a
falsely high EC fraction.
Chow et al. (2000) noted that neither the IMPROVE nor NIOSH methods were able to
accurately detect further blackening on optically dense, i.e., very black, filters that are typical of
source profile samples. Predictions of artifact pyrolysis carbon by TOR and TOT differed
widely. Note that both transmittance and reflectance methods rely on derivations of optics laws
(i.e., the Beer-Lambert and Lambert Laws) that predict a linear transmittance/reflectance signal
response with species concentration, but only for optically thin samples (Strobel and Heineman,
1989). Very black filters exceed this limitation; thus, the signal response of these methods may
not be a linear or otherwise predictable function of BC concentration.
Another important source of error in any TO measurement of aerosol OC/EC arises when
samples contain transition metal oxides, such as iron oxide. Many transition metal oxides are
found in crustal material. Fung et al. (2002) reported that such oxides can serve as oxidizing
agents for BC at high temperatures. The consequence of such an effect is an elevation of the
signal usually assigned to OC and corresponding reduction in apparent BC.
Thermal Manganese Oxidation
The thermal manganese oxidation (TMO) method (Mueller et al., 1982; Fung, 1990) uses
manganese dioxide in contact with the sample throughout the analysis as the oxidizing agent.
Temperature is relied upon to distinguish between OC and EC. Carbon evolving at 525 °C is
classified as OC, and carbon evolving at 850 °C is classified as EC. TMO does not correct for
pyrolized OC, which may lead to the overestimation of EC. This method has been used in the
SCENES (Subregional Cooperative Electric Utility, Department of Defense, National Park
2B-14
-------
Services, and Environmental Protection Agency Study); (Sutherland and Bhardwaja, 1987;
Mueller et al., 1986) visibility network, as well as in the SCAQS (Southern California Air
Quality Study) (Chow et al., 1994a,b; Watson et al., 1993, 1994a,b).
OC/ECMethods Intercomparisons
Chow et al. (1993) documented several variations of the thermal (T), thermal-optical
reflectance (TOR), thermal-optical transmission (TOT), and thermal manganese oxidation
(TMO) methods for OC and EC. Comparisons among the results of the majority of these
methods show that they yield comparable quantities of TC in aerosol samples, but the
distinctions between OC and EC are quite different (Cadle and Groblicki, 1982; Cadle and
Mulawa, 1990; Countess, 1990; Hering et al., 1990; Birch, 1998; Schmid et al., 2001). For the
sum of OC and EC, these methods reported agreement within 5 to 15% for ambient and source
samples (Houck et al., 1989; Kusko et al., 1989; Countess, 1990; Shah and Rau, 1990) and
within 3% on carefully prepared standards. Evaluation of these methods is thus a matter of
assessing how they differentiate between OC and EC. The EC:TC ratio for samples is method
dependent.
An international methods intercomparison study on the analysis of carbonaceous aerosols
on quartz fiber filters was organized by the Vienna University of Technology and involved 17
laboratories and 9 different thermal and optical methods (Schmid et al., 2001). All of the
participating laboratories were sent punches from three 150-mm quartz-fiber filters that had been
exposed for 24 h near a high-traffic street in Berlin. Five laboratories employed the German
official standard VDI2465 methods. Two of these laboratories used the VDI 2465/1 method
that determines extractable OC, non-extractable OC, and EC using a solvent-based extraction
protocol. Other laboratories participating in the intercomparison used variations of the VDI
2465 standard that rely upon differences in thermal stability to accomplish the separation of
carbonaceous aerosol fractions. A range of thermal protocols, TC determination techniques
and CO2 detection schemes were employed by the participating laboratories.
Good agreement of the TC results was obtained by all laboratories, with only two outliers
in the complete data set. The relative standard deviations between laboratories for the TC
results, were 6.7, 10.6, and 8.8% for the three samples. In contrast, the EC results were much
more variable. The relative standard deviations between laboratories for the EC results, were
2B-15
-------
36.6, 24.4, and 45.5% for the three samples. The VDI methods, especially the VDI 2465/2, were
found to give generally higher amounts of EC than the thermal-optical methods. This trend was
detected for all samples. The authors recognized that uncorrected thermal methods are prone to
positive artifacts by charring during pyrolysis. They also noted that when using solvent
extraction methods, the dissolution of polymeric aerosol constituents may not be successful.
Both of these effects would lead to an overestimation of the EC fraction. When the laboratories
were grouped according to their methods, the relative standard deviation between laboratories
was much smaller. This study demonstrated that the TC measurement can yield similar results
from a variety of methods, but the EC measurement is highly dependent upon the method used.
The problems associated with the determination of EC are exacerbated by the lack of a standard
reference material and consistent definitions of EC.
Measuring Black Carbon (BC) Instead of EC
Light-absorbing or black carbon (BC) can be measured by optical techniques (Penner and
Novakov, 1996). Both EC and BC define a similar fraction of aerosol; but EC is defined in
terms of both the thermal and light-absorption properties of the sample, whereas BC is based on
solely on its light-absorption properties. The aethalometer, the integrating sphere
sunphotometer, and photoacoustic spectroscopy (described in Section 2B.3) are example
techniques for determining BC.
Hitzenberger et al. (1996) investigated the feasibility of using an integrating sphere
photometer as an adequate measurement system for the BC content and the absorption
coefficient. In another study (Hitzenberger et al., 1999), the integrating sphere method was
compared to an aethalometer (Hansen et al., 1984), the thermal method of Cachier et al. (1989),
and the thermal-optical method of Birch and Gary (1996). The absorption coefficients that were
obtained from both the integrating sphere and the aethalometer were comparable. The BC mass
concentrations obtained from the aethalometer were 23% of those obtained from the integrating
sphere. Compared to the thermal method, the integrating sphere method overestimated the BC
mass concentrations by 21%. Compared to the thermal-optical method, the integrating sphere
was within 5% of the 1:1 line. However, the data were not well correlated. Thermal EC and
aethalometer BC measurements have also been compared by Lavanchy et al. (1999).
2B-16
-------
The Carbonaceous Species Methods Comparison Study (CSMCS), as mentioned in the
1996 PM AQCD (U.S. Environmental Protection Agency, 1996a) was conducted in Los Angeles
in 1986. Hansen and McMurry (1990) compared the thermal manganese method with the
aethalometer for aerosol EC. The first method involved collection of impactor samples backed
by a quartz-fiber after-filter followed by EC analysis by oxidation in helium over a MnO2
catalyst; the other method involved conducting real-time measurements using an aethalometer
(an optical absorption technique). They found good agreement between these two very different
methods. The CSMCS interlaboratory precision for TC was 4.2% (Turpin et al., 2000).
However, because the split between OC and EC is operationally defined, there was substantial
interlaboratory variability in OC and EC (e.g., 34% for EC [Turpin et al., 1990]).
EC/OC Summary. With the limitations and precautions described above, laboratory
analyses for the carbonaceous properties of collected particles have matured to the point at
which they can be performed with commercially available instruments following established
standard operating procedures. However, carbon analysis continues to be a subject of active
research within the atmospheric sciences community and EPA, and carbon speciation methods
comparisons are being undertaken in such studies as the Atlanta Supersite. The state of the art
for carbonaceous PM measurements continues to advance; and, although progress is being made,
the definitions of OC, EC, and BC continue to be operationally defined in reference to the
method employed. Reports of EC/OC measurements should therefore include mention of the
method with which the species were determined. Finally, if possible, all ancillary data should be
retained, to allow later comparison to other methods.
2B.3 CONTINUOUS METHODS
2B.3.1 Continuous Measurement of Mass
Tapered Element Oscillating Microbalance (TEOM)
The advantages of continuous PM monitoring and the designation of the TEOM as an
equivalent method for PM10, have led to the deployment of the TEOM at a number of air
monitoring sites. The TEOM also is being used to measure PM2 5. The TEOM differs from the
federal reference methods for PM in that it does not require equilibration of the samples at a
2B-17
-------
specified temperature and RH. The TEOM samples at a constant temperature and is typically
heated to some temperature higher than the ambient temperature (Meyer et al., 1995; Meyer and
Rupprecht, 1996), whereas the FRM samples at the ambient temperature. Thus, the TEOM does
not provide data equivalent to the FRM because of losses of volatile species. Volatilization
losses in the TEOM sampler can be reduced by operating the instrument at 30 °C, rather than the
typical 50 °C specified, or by using a Nafion diffusion dryer instead of heating to dehumidify the
particles (with a 30 °C temperature).
This difference in operation and implications for fine particle measurements was examined
by researchers at Commonwealth Scientific and Industrial Research Organization (CSIRO)
Atmospheric Research in Australia (Ayers et al., 1999). That group compared 24-h mean PM2 5
mass concentrations as determined by a TEOM and by two manual, gravimetric samplers (a low-
volume filter sampler and a MOUDI sampler) in four Australian cities, on 15 days in the
wintertime. The TEOM was operated at 50 °C at one location and at 35 °C at the other three
locations. A systematically lower TEOM response in comparison to the integrated gravimetric
methods was observed. In a comprehensive study, Allen et al. (1997) reported results in which
TEOM data collected at 10 urban sites in the United States and Mexico were compared with
24-h integrated mass concentrations for both PM10 and PM2 5. They collected a large data set that
included both winter and summer seasons. Allen et al. (1997) concluded that, especially for
urban areas, a significant fraction of PM10 may be semivolatile compounds that could be lost
from the heated filter in the TEOM leading to a systematic difference between the TEOM and
the EPA FRM for PM10. They suggested that this difference is likely to be larger for PM2 5 than
it is for PM10 (Allen et al., 1997).
In a similar study conducted in Vancouver, British Columbia the effect of equilibration
temperature on PM10 concentrations from the TEOM was examined. Two collocated TEOM
monitors, operated at 30 and 50 °C, respectively, collected data for a period of-17 months in the
Lower Fraser Valley in British Columbia (Mignacca and Stubbs, 1999). A third TEOM
operating at 40 °C was operated for 2 months during this period. They found that, on average,
the 1-h average PM10 from the TEOM operating at 30 °C was consistently greater than that from
the TEOM operated at 50 °C. For the period during which the third TEOM was operated (at
40 °C), the PM10 from that instrument was between the values for the other two instruments.
They also found that the differences in masses were proportional to the PM10 loading, and more
2B-18
-------
strongly correlated to the PM10 from the TEOM operated at the lower temperature. They
recommended that the TEOM monitors be operated at 40 °C in summer and 30 °C in winter,
in order to avoid introducing a methodological seasonal bias.
A new sample equilibration system (SES) was developed to reduce losses of semivolatile
species from the PM2 5 TEOM by conditioning the sample stream to a lower RH and temperature
(Meyer et al., 2000). The SES utilizes humidity sensors and a Nafion dryer designed for low
particle loss. The dryer fits between the flow splitter that follows the size-selective inlet and the
sensor unit. A dry purge gas flows over the exterior of the Nafion tubing and allows for self-
regeneration. A TEOM with PM2 5 inlet and equipped with an SES was operated at 30 °C
alongside another TEOM operating at 50 °C without the SES in Albany, NY, over a 6-day period
during a summertime high-temperature, high-relative-humidity episode. The SES maintained
the sample air relative humidity under 30%, and the TEOM with the SES generally measured
more mass than the other TEOM. The TEOM with SES also was operated alongside an FRM-
type sampler for the period of June 6 through September 25, 1999. The correlation between the
FRM and TEOM/SES showed a slope of 1.0293 and R2 of 0.9352; whereas the correlation
between the FRM and the TEOM without SES and operating at 50 °C showed a slope of 0.8612
and R2 of 0.8209. The SES can be installed on existing TEOM monitors.
Beta-Gauge Techniques
Beta radiation absorption, as a indicator of particle mass, has been used effectively to
measure the mass of equilibrated PM collected on Teflon filters (Jaklevic et al., 198 la; Courtney
et al., 1982). The technique also has been used to provide near real-time measurements with
time intervals on the order of an hour (Wedding and Weigand, 1993). However, real-time beta-
gauge monitors experience the same problems as other continuous or near real-time PM mass
monitoring techniques. Particle-bound water (PBW) must be removed to reduce the sensitivity
of the indicated mass to relative humidity. However, the simplest technique, mild heating, will
remove a portion of the ammonium nitrate and the semivolatile organic compounds as well as
the PBW.
An intercomparison study of two beta gauges at three sites indicated that the Wedding beta
gauge and the Sierra Anderson (SA) 1200 PM10 samplers were highly correlated, r > 0.97 (Tsai
and Cheng, 1996). The Wedding beta gauge was not sensitive to relative humidity but yielded
2B-19
-------
results approximately 7% lower. This suggests that the mild heating in the beta gauge causes
losses comparable to those caused by equilibration, although the differences could result from
slight differences in the upper cut points. The Kimoto beta gauge that was operated at ambient
temperature was sensitive to relative humidity yielding significantly higher mass concentrations
relative to the SA 1200 for RH > 80% than for RH < 80% even though the correlation with the
SA 1200 was reasonable (r = 0.94 for RH > 80% and 0.83 for RH < 80%).
Piezoelectric Microbalance
Piezoelectric crystals have mechanical resonances that can be excited by applying an
alternating electrical voltage to the crystal. As the resonance frequencies are well defined, such
crystals (quartz in particular) have found applications as secondary time and frequency standards
in clocks and watches. As for all mechanical resonators, the resonance frequency is a function
of mass. Therefore, by monitoring the resonance frequency in comparison with a second crystal,
one can continuously measure the mass deposited on the crystal (Sem et al., 1977; Bowers and
Chuan, 1989; Ward and Buttry, 1990; Noel and Topart, 1994). Comparison with a second
crystal largely compensates for the effect of temperature changes on the resonance frequency.
The piezoelectric principle has been used to measure particle mass by depositing the
particles on the crystal surface either by electrostatic precipitation or by impaction (Olin and
Sem, 1971). The collection efficiency of either mechanism has to be determined as function of
particle size to achieve quantitative measurements. In addition, the mechanical coupling of large
particles to the crystal is uncertain. Both single and multi-stage impactors have been used (Olin
and Sem, 1971; Fairchild and Wheat, 1984). Quartz crystals have sensitivities of several
hundred hertz per microgram. This sensitivity results in the ability to measure the mass
concentration of a typical 100 |ig/m3 aerosol to within a few percent in less than one minute
(Olin and Sem, 1971).
Coarse Particle Mass
The RAMS and CAMM are only appropriate for fine particle measurements (e.g., PM2 5
or PMj). However, the TEOM, beta gauge, and piezoelectric microbalance may be used to
measure either PM2 5 or PM10 (or a sample with any specified upper 50% size cut). A pair of
such samplers may be used to measure thoracic coarse PM mass concentration (PM10_2 5) by
2B-20
-------
difference between the PM10 and PM2 5 concentrations. However, concerns have been raised
concerning the quality of the data from such difference calculations and the resulting potential
biases in exposure assessment and risk determinations (Wilson and Suh, 1997; White, 1998).
Misra et al. (2001) describe the development and evaluation of a continuous coarse particle
monitor (CCPM) that may provide direct measurements of coarse mode PM mass concentrations
at short time intervals (on the order of 5-10 min). The basis of the CCPM is enrichment of the
coarse particle concentrations through use of virtual impaction while maintaining fine particle
concentrations at ambient levels. The resulting aerosol mixture is analyzed using a standard
TEOM for which the response is now dominated by the enriched coarse PM mass. The coarse
PM concentrations determined from the CCPM were compared to those obtained with a
MOUDI, operating with only the 10- and 2.5-|im cut point stages, and a Partisol dichotomous
sampler. The CCPM coarse particulate concentrations were highly correlated with both the
MOUDI (R2= 0.88) and the Partisol (R2 = 0.88) coarse PM concentrations. By operating the
CCPM at a coarse particle enrichment factor of 25, the coarse PM concentration can be
determined a priori without determination of the fine particle concentration, so long as the fine-
to-coarse particle concentration ratios are not unusually high (i.e., 4 to 6). Misra et al. (2001)
also concluded from field experiments that the coarse particulate concentrations determined from
the CCPM were independent of the ambient fine-to-coarse particulate concentration ratio due to
the decrease in particle mass median diameter that should accompany fine-to-coarse particulate
concentration ratios during stagnation conditions.
2B.3.2 Continuous Measurement of Organic and/or Elemental Carbon
Testing and refinement of models that simulate aerosol concentrations from gas and
particle emissions require air quality measurements of approximately 1-h time resolution to
reflect the dynamics of atmospheric transport, dispersion, transformation, and removal. Below
instruments are described that have been used to collect and analyze atmospheric organic PM
with better than 2-h resolution. These instruments were all present at the Atlanta Supersite
experiment during the summer of 1999, and an intercomparison of results is underway.
2B-21
-------
Automated Carbon Analyses
Turpin et al. (1990) describe an in situ, time-resolved analyzer for particulate organic and
EC that can operate on a time cycle as short as 90 min. This analyzer collects PM on a quartz-
fiber filter mounted in a thermal-optical transmittance carbon analyzer (Turpin et al., 1990).
A second quartz-fiber filter behind a Teflon filter in a second sampling port may also be
analyzed to provide an estimate of the positive sampling artifact (i.e., gas adsorption on the
quartz sampling filter). The organic material in the collected PM is thermally desorbed from the
filter at 650 °C and oxidized at 1000 °C over a MnO2 catalyst bed. The evolved CO2 is
converted to methane over a nickel catalyst, and the methane is measured in a flame ionization
detector. Then the EC is oxidized at 350 °C in a 98% He-2% O2 atmosphere. Correction is
made for pyrolytic conversion of some of the organic PM. The instrument was operated with a
2-h resolution during SCAQS in 1987 (Turpin and Huntzicker, 1991, 1995), as well as during
CSMCS in 1986 (Turpin et al., 1990). By using EC as a tracer for primary, combustion-
generated OC, these authors estimated the contributions of primary sources (i.e., material
emitted in particulate form) and secondary sources (i.e., PM formed in the atmosphere) to the
total atmospheric particulate OC concentrations in these locations.
An automated carbon analyzer with 15-min to 1-h resolution is now commercially
available (Rupprecht et al., 1995) and has been operated in several locations including the
Atlanta Supersite. It collects samples on a 0.1 jim impactor downstream of an inlet with a
2.5 |im cut point. Use of an impactor eliminates the gas adsorption that must be addressed when
filter collection is used. However, this collection system may experience substantial particle
bounce and loss of a sizable fraction of EC since some EC is in particles < 0.2 jim. It is possible
that ongoing research, in which particle size is increased by humidification prior to impaction,
may result in an improvement in collection efficiency. In the analysis step, carbonaceous
compounds are removed by heating in filtered ambient air. Carbonaceous material removed
below 340 °C is reported as OC, and material removed between 340 and 750 °C is reported as
EC. Turpin et al. (2000) commented that it would be more appropriate to report carbon values
obtained by this method as "low-" and "high-temperature" carbon, because some organics are
known to evolve at temperatures greater than 340 °C (e.g., organics from woodsmoke).
2B-22
-------
Aethalometer for Elemental Carbon
As discussed earlier, black carbon (BC), a carbon fraction very similar to EC, is most
commonly measured using an aethalometer, a commercially available, automated, time-resolved
instrument (i.e., 5- to 15-min sample duration) that measures the light attenuation of aerosol
particles collected on a filter tape (Hansen et al., 1984). The concentration of EC is derived from
the light absorption measured on a filter using an estimate of the specific absorption (m2/g) of
EC on the filter; the specific absorption value is derived from laboratory and atmospheric tests
and is specified by the manufacturer. The specific absorption value could be expected to vary
with location, season, and source mix. Comparisons in atmospheric experiments at some
locations with EC values measured by thermal methods confirm that the aethalometer provides a
statistically meaningful estimate of EC concentration (Allen et al., 1999c; Liousse et al., 1993).
For instance, Allen et al. (1999c) found the following statistical relationship for Uniontown, PA
during summer 1990: black carbon (aethalometer) = 0.95*EC (thermal) - 0.2 (r2 = 0.925, n not
specified but appears to be > 50, EC range from 0 to 9 |ig/m3). Another source of error in
aethalometer measurements arises from the sampling procedure. Particles are trapped within a
three-dimensional filter matrix. Therefore, scattering of transmitted and reflected light may
erroneously be attributed to absorption, thus causing errors in the BC calculation. Ballach et al.
(2001) investigated immersing the filter in oil of a similar refractive index as a means to
minimize the interferences due to light scattering effects from the filter, similar to a procedure
common in microscopy. Black carbon measurements determined using the oil immersion
technique were compared to those from an integrating sphere, a polar photometer, and Mie
calculations. Aerosol tests included several pure carbon blacks from different generating
procedures that were used to calibrate the immersion technique, pure ammonium sulfate aerosol,
and external and internal mixtures of ammonium sulfate with varying amounts of carbon blacks.
The oil immersion technique was also tested on ambient air samples collected at two different
sites in the cities of Frankfurt am Main and Freiburg, Germany. Optical measurements, both of
blank and loaded filters, showed that the oil immersion technique minimizes scattering losses.
Ballach et al. (2001) found that site-related effects were reduced and that there was reasonably
good agreement with the other optical techniques as well as with the Mie calculations.
2B-23
-------
Photoacoustic Measurement of Elemental Carbon
Measurement of aerosol light absorption utilizing photoacoustic spectroscopy has been
examined as a continuous method for measuring EC mass concentrations (Petzold and Niessner,
1996; Petzold et al., 1997; Arnott et al., 1999, 2000). Like the aethalometer, this method
measures light absorption; however, unlike most other light absorption methods, the
photoacoustic technique does not require a filter. The photoacoustic spectrometer of Arnott and
coworkers was demonstrated during the Northern Front Range Air Quality Study and compared
to an aethalometer (Moosmiiller et al., 1998). Neither the aethalometer nor the photoacoustic
spectrometer measure EC mass directly. Because the photoacoustic spectrometer measures the
absorption coefficient directly, the specific absorption efficiency must be known or assumed in
order to determine EC mass. Assuming a light absorption efficiency of 10 m2/g, Arnott et al.
(1999) reported a lower detection limit for light absorption of 0.4 M/m corresponding to a mass
concentration of EC of approximately 40 ng/m3.
2B.3.3 Continuous Measurements of Nitrate and Sulfate
Nitrate
An integrated collection and vaporization cell was developed by Stolzenburg and Hering
(2000) that provides automated, 10-min resolution monitoring of fine-particulate nitrate. In this
system, particles are collected by a humidified impaction process and analyzed in place by flash
vaporization and chemiluminescent detection of the evolved nitrogen oxides. In field tests in
which the system was collocated with two FRM samplers, the automated nitrate sampler results
followed the results from the FRM, but were offset lower. The system also was collocated with
a HEADS and a SASS speciation sampler (MetOne Instruments). In all these tests, the
automated sampler was well correlated to other samplers with slopes near 1 (ranging from 0.95
for the FRM to 1.06 for the HEADS) and correlation coefficients ranging from 0.94 to 0.996.
During the Northern Front Range Air Quality Study in Colorado (Watson et al., 1998), the
automated nitrate monitor captured the 12-min variability in fine-particle nitrate concentrations
with a precision of approximately ±0.5 |ig/m3 (Chow et al., 1998). A comparison with denuded
filter measurements followed by ion chromatographic (1C) analysis (Chow and Watson, 1999)
showed agreement within ±0.6 |ig/m3 for most of the measurements, but exhibited a discrepancy
of a factor of two for the elevated nitrate periods. More recent intercomparisons took place
2B-24
-------
during the 1997 Southern California Ozone Study (SCOS97) in Riverside, CA. Comparisons
with 14 days of 24-h denuder-filter sampling gave a correlation coefficient of R2 = 0.87 and
showed no significant bias (i.e., the regression slope is not significantly different from 1). As
currently configured, the system has a detection limit of 0.7 |ig/m3 and a precision of 0.2 |ig/m3.
Sulfate
Continuous methods for the quantification of aerosol sulfur compounds first remove
gaseous sulfur (e.g., SO2, H2S) from the sample stream by a diffusion tube denuder followed by
the analysis of particulate sulfur (Cobourn et al., 1978; Durham et al., 1978; Huntzicker et al.,
1978; Mueller and Collins, 1980; Tanner et al., 1980). Another approach is to measure total
sulfur and gaseous sulfur separately by alternately removing particles from the sample stream.
Particulate sulfur is obtained as the difference between the total and gaseous sulfur (Kittelson
et al., 1978). The total sulfur content is measured by a flame photometric detector (FPD) by
introducing the sampling stream into a fuel-rich, hydrogen-air flame (e.g., Stevens et al., 1969;
Farwell and Rasmussen, 1976) that reduces sulfur compounds and measures the intensity of the
chemiluminescence from electronically excited sulfur molecules (S2*).
Because the formation of S2* requires two sulfur atoms, the intensity of the
chemiluminescence is theoretically proportional to the square of the concentration of molecules
that contain a single sulfur atom. In practice, the exponent is between 1 and 2 and depends on
the sulfur compound being analyzed (Dagnall et al., 1967; Stevens et al., 1971). Calibrations are
performed using both particles and gases as standards. The FPD can also be replaced by a
chemiluminescent reaction with ozone that minimizes the potential for interference and provides
a faster response time (Benner and Stedman, 1989, 1990).
Capabilities added to the basic system include in situ thermal analysis and sulfuric acid
speciation (Cobourn et al., 1978; Huntzicker et al., 1978; Tanner et al., 1980; Cobourn and
Husar, 1982). Sensitivities for particulate sulfur as low as 0.1 |ig/m3, with time resolution
ranging from 1 to 30 min, have been reported. Continuous measurements of particulate sulfur
content have also been obtained by on-line XRF analysis with resolution of 30 min or less
(Jaklevic et al., 1981b). During a field-intercomparison study of five different sulfur
instruments, Camp et al. (1982) reported four out of five FPD systems agreed to within
± 5% during a 1-week sampling period.
2B-25
-------
2B.4 OTHER CONTINUOUS MEASUREMENTS
Continuous Ion Chromatography of Water-Soluble Ions
Two particle collection systems that grow particles by increasing the relative humidity and
collect the particles in an aqueous solution suitable for injection into an ion chromatography
have been developed independently (Simon and Dasgupta, 1995; Khlystov et al., 1995).
Automation of these systems yield semi-continuous monitors for those ions that can be
determined using ion chromatography. A similar system using a particle size magnifier has been
reported by Weber et al. (2001).
Determination of Aerosol Surf ace Area in Real Time
Aerosol surface area is an important aerosol property for health effects research. However,
methods for on-line measurement of surface area are not widely available. Woo et al. (2001b)
used three continuous aerosol sensors to determine aerosol surface area. They used a
condensation particle counter (CPC, TSI, Inc., Model 3020), an aerosol mass concentration
monitor (MCM, TSI, Inc., Model 8520), and an electrical aerosol detector (BAD, TSI, Inc.,
Model 3070) to measure integral moments of the particle size distribution. The three sensor
signals were inverted to obtain the aerosol size distribution, using a lognormal size distribution
model (by minimizing the difference between the measured signals and the theoretical values
based upon a size distribution model, the instrument calibration, and its theoretical responses).
The lognormal function was then integrated to calculate the total surface area concentration.
Woo et al. (2001b) demonstrated that this method can give near real-time measurements of
aerosol surface area.
Light Scattering
A variety of types of nephelometers that integrate aerosol light scattering over various solid
angles are available (McMurry, 2000). When used to measure visibility, e.g., to provide pilots
with real-time data on visual range, it is desirable to include the light scattering due to PBW.
However, when used as an indicator of fine particle mass, it is desirable to exclude PBW. This
is frequently done by heating the ambient aerosol to a low reference relative humidity of 40%.
However, this heating has the potential of also causing the loss of semivolatile components of
the aerosol. The evaporation of ammonium nitrate aerosol in a heated nephelometer has been
2B-26
-------
examined. Bergin et al. (1997) conducted laboratory experiments at low relative humidity
(-10%) and as a function of temperature (27 to 47 °C), mean residence time in the nephelometer,
and initial particle-size distribution. The evaporation of ammonium nitrate aerosol was also
modeled for comparison and was found to describe accurately the decrease in aerosol scattering
coefficient as a function of aerosol physical properties and nephelometer operating conditions.
Bergin et al. (1997) determined an upper limit estimate of the decrease in the aerosol light
scattering coefficient at 450 nm due to evaporation for typical field conditions. The model
estimates for their worst-case scenario suggest that the decrease in the aerosol scattering
coefficient could be roughly 40%. Under most conditions, however, they estimate that the
decrease in aerosol scattering coefficient is generally expected to be < 20%.
2B-27
-------
3. CONCENTRATIONS, SOURCES, AND EMISSIONS
OF ATMOSPHERIC PARTICULATE MATTER
3.1 INTRODUCTION
This chapter discusses topics that were covered in Chapter 5 (Sources and Emissions of
Atmospheric Particles) and Chapter 6 (Environmental Concentrations) of the previous document,
Air Quality Criteria for Paniculate Matter or "1996 PM AQCD" (U.S. Environmental
Protection, 1996) and updates the subject information based on new data, as available.
Information about concentrations, the composition, and the spatial and temporal variability
of ambient particulate matter (PM) across the United States is presented in Section 3.2. Ambient
concentration data obtained during the first 3 years of operation of the recently deployed
nationwide network of Federal Reference Method (FRM) PM25 monitors in 27 metropolitan
statistical areas (MSAs) are presented and analyzed in Appendix 3A. Initial data from the pilot
method evaluation study for the national speciation network are presented in Appendix 3B.
Results of field studies that help to characterize the composition of organic compounds in
ambient air particles are summarized in Appendix 3C, to complement data on the inorganic
composition of ambient particles presented in Appendix 6 A of the 1996 PM AQCD and
Appendix 3B of this document. Data useful for characterizing the daily and seasonal variability
of PM25 concentrations are discussed in Section 3.2.1, and the intraday variability of PM25
concentrations is discussed in Section 3.2.2. Relationships among different size fractions are
discussed in Section 3.2.3; interrelationships and correlations among PM components are
discussed in Section 3.2.4; and the spatial variability of various PM components is discussed in
Section 3.2.5.
Atmospheric PM, unlike gaseous criteria pollutants (SO2, NO2, CO, O3) which are
well-defined chemical entities, is composed of a variety of particles differing in size and
chemical composition. Therefore, sources of each component of the atmospheric aerosol must
be considered in turn. Differences in the composition of particles emitted by different sources
also lead to spatial and temporal heterogeneity in the composition of the atmospheric aerosol.
The nature of the sources and the composition of the emissions from these sources are discussed
in Section 3.3. The chemistry of formation of secondary PM from gaseous precursors is
3-1
-------
discussed in Section 3.3.1. Estimates of contributions of various sources to ambient PM levels
given by source apportionment studies also are presented in Section 3.3.2. More detailed
information about the composition of emissions from various sources is given in Appendix 3D.
The determination of "policy-relevant background" (PRB) concentrations, including
contributions from natural local sources and the long-range transport of PM from sources outside
the United States, is discussed in Section 3.3.3. Reviews of transport of PM and its precursors
within the United States can be found in the North American Research Strategy for Tropospheric
Ozone (NARSTO) Fine Particle Assessment (NARSTO, 2002). More detailed information
regarding sulfur and nitrogen species can be found in Hidy (1994). Further information about
PM concentrations observed at relatively remote monitoring sites (RRMS), i.e., monitoring sites
located in relatively remote areas that are not obviously influenced by local anthropogenic
sources, is given in Appendix 3E. Because PM is composed of both primary and secondary
constituents, emissions of both the primary components and the gaseous precursors of secondary
PM are considered here. Nationwide emissions estimates of primary PM and precursors to
secondary PM are discussed in Section 3.3.4, and uncertainties in emissions estimates are
discussed in Section 3.3.5.
The organization of topics in this chapter (ambient measurements, source characterization
and apportionment, and emissions inventories) reflects, in a broad sense, the order in which these
topics are addressed in scientific studies and, arguably, the increasing levels of uncertainty that
are associated with these different topics.
3.2 PATTERNS AND TRENDS IN AMBIENT PM CONCENTRATIONS
Considerable data exists for characterizing PM10 mass concentrations and trends, and those
available at the time were presented in the 1996 PM AQCD. In contrast, less extensive data sets
were available for characterizing PM2 5 and PM10_2 5 mass or trends. Sources of data for PM2 5
(fine) and PM10_25 (coarse) discussed in the 1996 PM AQCD include: EPA's Aerometric
Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2000a);
Interagency Monitoring of Protected Visual Environments (IMPROVE; Eldred and Cahill, 1994;
Cahill, 1996); the California Air Resources Board (CARB) Data Base (California Air Resources
Board, 1995); the Harvard Six-Cities Data Base (Spengler et al., 1986; Neas, 1996); and the
5-2
-------
Harvard-Philadelphia Data Base (Koutrakis, 1995). The Inhalable Particulate Network (IPN)
(Inhalable Paniculate Network, 1985; Rodes and Evans, 1985) provided TSP, PM15, and PM25
data but only a small amount of PM10 data.
New sources of PM data include the recently deployed nationwide PM2 5 compliance-
monitoring network, which provides mass measurements using a FRM. This section
summarizes data obtained by this network from 1999 to 2001 and provides an approximate
characterization of nationwide PM10_25 concentrations by comparing PM10 to PM25 data from
sites where both types of compliance monitors are located. Various aspects of these data are
presented in greater detail in Appendix 3 A. In addition, a small number of recent studies in
which daily mass and composition measurements are available for extended periods are
discussed in this section. The results of quality-assured aerosol composition data obtained by
X-ray fluorescence (XRF) and by analyses of organic carbon and elemental carbon for 13 urban
areas from the methods evaluation study for the national PM2 5 speciation network are presented
in Appendix 3B. The terms organic carbon (OC) and elemental carbon (EC) are subject to some
ambiguity, and the meanings of these terms as discussed in Section 2.2.7 and Appendix 2.B.2 are
applied here.
Organic compounds contribute from 10 to 70% of the dry fine particle mass in the
atmosphere (see Appendix 3C). However, concentrations and the composition of organic PM
are poorly characterized, as are its formation mechanisms. Particulate organic matter is an
aggregate of hundreds of individual compounds spanning a wide range of chemical and
thermodynamic properties (Saxena and Hildemann, 1996). Some of the organic compounds are
"semivolatile" (i.e., they have atmospheric concentrations and saturation vapor pressures such
that both gaseous and condensed phases exist in equilibrium in the atmosphere). The presence
of semivolatile or multiphase organic compounds complicates the sampling process. Organic
compounds originally in the gas phase may be absorbed on glass or quartz filter fibers and create
a positive artifact. Conversely, semivolatile compounds originally present in the condensed
phase may evaporate from particles collected on glass, quartz, or Teflon filters creating a
negative artifact. In addition, no single analytical technique is currently capable of analyzing the
entire range of organic compounds present in atmospheric PM. Rigorous analytical methods are
able to identify only 10 to 20% of the organic PM mass on the molecular level (Rogge et al.,
1993), and only about 50% of the condensed phase compounds can be identified in smog
-------
chamber studies of specific compounds (Forstner et al., 1997a,b). Measurement techniques and
associated artifacts are discussed in Section 2.2.3.2. Information on the identification and
concentration of the many different organic compounds identified in atmospheric samples
obtained during the 1990s is given in Appendix 3C.
Summary tables giving the results of 66 field studies that obtained data for the composition
of particles in the PM2 5, PM10_25, or PM10 size ranges were presented in Appendix 6 A of the
1996 PM AQCD. The summary tables included data for mass, OC, EC, nitrate, sulfate, and trace
elements. Data from the studies were presented for the eastern, western, and central
United States. It should be noted that these studies took place at various times and lasted for
various durations over a 20-year period, and there may have been significant changes in the
concentrations of many species between the times when these studies were conducted and now.
These changes resulted from a number of factors (e.g., pollution controls, technological
advances, land use changes, etc).
There were a number of discernible differences in the composition of particles across the
United States evident in the data sets listed in Appendix 6A in the 1996 PM AQCD (cf,
Figures 6-85a to 6-85c). However, these differences can only be discussed in the context of the
uncertainties in the measurements of the main components (sulfate, OC, EC, crustal material,
ammonium, and nitrate). Sulfate, followed by crustal materials, has the smallest uncertainty
associated with its measurement among all the components listed. Sulfate constituted -38% of
PM2 5 in the aerosol composition studies in the eastern United States and was the major
identifiable component of PM25. However, it constituted only ~11% of PM25 in the studies
listed for the western United States. The contribution of crustal materials to PM2 5 ranged from
-4% in the East to -15% in the West. The contribution of unidentified material (possibly
consisting mostly of water of hydration) ranged -23% in the East to 0% in the West. The
contribution of EC to PM25 ranged from -4% in the East to -15% in the West. Organic
compounds constituted -21% of PM25 in the eastern United States, reaching to -39% for the
studies listed in the western United States. However, uncertainties for OC, EC, ammonium, and
nitrate are larger than for sulfate and crustal material. Concentrations of OC were multiplied by
a factor of 1.4 when calculating mass to account for the presence of oxygen, nitrogen, and
hydrogen in the organic compounds. This factor may vary among different areas and may
represent the lowest reasonable estimate for an urban aerosol (Turpin and Lim, 2001).
5-4
-------
In addition, the samples collected in the studies were subject to a number of sampling artifacts
involving the adsorption of gases and the evaporation of volatile components that either formed
on the filters or were present in the ambient particles. The values reported for OC and EC in
filter samples depend strongly on the specific analysis method used (Chow et al., 2002), as
discussed extensively in Chapter 2.
Crustal materials constitute from -52% of PM10_25 in the eastern United States to -70%
of PM10_2 5 in the western United States, as indicated by studies in Appendix 6 A of the 1996 PM
AQCD. The fraction of unidentified material in PM10_2 5 varied from 41% in the eastern United
States to 27% in the western United States. However, in the vast majority of these studies, no
attempt was made to characterize organic components or nitrate in the PM10_2 5 size fraction.
It should also be remembered that a small fraction (typically about 10%) of PM25 is entrained
into the flow of the channel of the dichotomous sampler that collects the PM10_2 5 sample and that
errors may occur during the procedure used to account for this entrainment. Even if analyses of
total OC were available, they would not be able to distinguish between primary biological
particles (PBPs), which include microorganisms and fragments of living things, and organic
compounds of miscellaneous origin in surface deposits on filters. A clear distinction should be
made between PBP and primary OC emitted by organisms (e.g., waxes coating the surfaces of
leaves, seeds, fruits, stems, pollen, fungi, and insects). Indeed, the fields of view of many
photomicrographs of PM samples obtained by scanning electron microscopy are often dominated
by large numbers of pollens, plant and insect fragments, and microorganisms. Bioaerosols such
as pollens, fungal spores, and most bacteria are expected to be found mainly in the coarse size
fraction. However, allergens from pollens can also be found in respirable particles (Monn, 2001;
Taylor et al., 2002). Matthias-Maser (1998) summarized information about the size distribution
of PBP in and around Mainz, Germany in what is perhaps the most complete study of this sort.
Matthias-Maser found that PBP constituted up to 30% of total particle number and volume in the
size range from about 0.35 jim to about 50 jim on an annual basis. Additionally, whereas the
contribution of PBP to the total aerosol volume did not change appreciably with season, the
contribution of PBP to total particle number ranged from about 10% in December and March to
about 25% in June and October.
Data for the chemical composition of particles in a number of national parks and remote
areas have been collected for a number of years by the IMPROVE network. Concentrations
5-5
-------
have been reported for sulfate, nitrate, light-absorbing carbon, OC, and soil components.
With the collection of compositional data by the speciation network, more synoptic (i.e.,
concurrent) coverage will be obtained for these constituents from relatively remote to urban
environments across the United States.
PM10 Concentrations and Trends
Nationwide PM10 annual mean concentrations on a county-wide basis from the AIRS
database for calendar years 1999, 2000, and 2001 are shown in Figure 3-la. Concentrations in
most areas of the country from 1999 through 2001 were below the PM10 annual standard of
50 |ig/m3. Further information about the attainment status of different areas can be found in
EPA's Air Quality Trends Reports. The nationwide median value of county-wide annual
average PM10 concentrations for this three year period (1999 to 2001) was -23 |ig/m3. Those
concentrations flagged as natural events (e.g., resulting from high winds, wildfires, volcanic
eruptions) or exceptional events (e.g., construction, prescribed burning) were not included in the
calculations. Procedures used for calculating the annual means at the site level followed
40 Code of Federal Regulations (CFR) Part 50 Appendix K (requiring 75% completeness per
quarter for all three years). The 98th percentile concentrations from the PM10 monitor showing
the highest value in a given county over the three-year period are shown in Figure 3-lb.
As shown by the blank areas on the maps, the picture is not complete, because some monitoring
locations did not record valid data for all four quarters, or recorded fewer than 11 samples in one
or more quarters, or some counties simply did not have monitors. Similar considerations apply
to the maps shown later for PM2 5 and PM10_2 5. It should also be noted that the area of counties
can be much greater in the West than in the East. As a result, the density of monitors may
appear to be greater in the West and air quality may appear to be worse over much larger areas in
the West than in the East. Concentrations are shown at the county level because this is the
typical scale used in many health outcome studies. Metropolitan statistical area (MSA) or multi-
county scales have also been used in some number of epidemiologic studies, e.g., Schwartz et al.
(1996) or the NMMAPS studies discussed in Chapter 8.
Nationwide trends in annual mean PM10 concentrations from 1992 through 2001 (based on
data obtained at 153 rural sites, 297 suburban sites, and 316 urban sites reporting to AIRS)
are shown in Figure 3-2 (U.S. Environmental Protection Agency, 2002a). Although average
5-6
-------
Concentration f^g/
0 < X < 23
23 < x < 38
x>38
Figure 3-la. 1999-2001 county-wide average annual mean PM10 concentrations (ug/m3)
for counties with PM10 monitors.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
Concentration (|ig/m3)
0 170
Figure 3-lb. 1999-2001 highest county-wide 98th percentile 24-h average PM10
concentrations (ug/m3) for counties with PM10 monitors.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
5-7
-------
30
25
! 20
£;
c
~ 15
(0
+J
c
0)
o 10
o
O
1992 - 2001
Rural Sites (153)
___ Suburban sites (297)
Urban Sites (316)
I I I I
92 93 94 95
96 97
YEAR
98 99 00
01
Figure 3-2. Nationwide trend in ambient PM10 concentration from 1992 through 2001.
Source: U.S. Environmental Protection Agency (2002a).
concentration levels differ among types of sites with higher levels at urban and suburban sites
the nationwide data set shows a decrease of 14% that occurred mainly during the first half of the
record. Concentrations at suburban and urban sites show a consistent decline from 1992 to 2001.
However, concentrations at rural sites appear to have leveled off since about 1996. Figure 3-3
shows the annual mean PM10 trend summarized by EPA region. Decreases in annual average
PM10 concentrations from 1990 to 1999 were largest in the Northwest (9.6 |ig/m3) and smallest
in the south-central United States (1.3 |ig/m3). Analyses of available TSP measurements
obtained since 1950 indicate that mean TSP concentrations appear to have declined by about
2- to 3-fold in urban areas between 1950 and 1980 (Lipfert, 1998).
PM2 5 Concentrations and Trends
Nationwide annual mean PM25 concentrations obtained from data collected during 1999,
2000, and 2001 are shown in Figure 3-4a; and 98th percentile concentrations are shown in
Figure 3-4b. Quantities shown in Figure 3-4a and 3-4b were calculated for individual counties.
-------
20.529
23.500 \| 1 \
—-— A ' J>
ScT. v^
?* \ t11%
18.287
The National Trend
27.745 23.907
1992 2001
f- 14%
Alaska is in
EPA Region 10;
Hawaii, EPA Region 9
and PuertoRico,
EPA Region 2.
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, 1992 through
2001 (ug/m3).
Source: U.S. Environmental Protection Agency (2002a).
Annual mean concentrations shown in Figure 3-4a were calculated on the basis of the latest
8 consecutive quarters containing at least 11 samples per quarter, and results from the highest
monitor were used to produce Figure 3-4b. Data meeting these completeness criteria were
obtained at 897 sites in 592 counties. The median PM2 5 concentration nationwide was about
13 |ig/m3. Annual mean PM25 concentrations were above 17 |ig/m3 at 5% of the sites, mainly in
California and in the southeastern United States. The 98th percentile 24-h average
concentrations (as defined earlier for PM10) were below 47 |ig/m3 at 95% of the sites sampled.
Most of the sites with levels above this value are located in California.
Annual average PM2 5 concentrations obtained as part of several health studies conducted
in various locations in the United States and Canada from the late 1980s to the early 1990s are
shown in Figure 3-5 (Bahadori, 2000a). These studies include the Harvard Six-Cities Study
(Steubenville, OH; Watertown, MA; Portage, WI; Topeka, KS; St. Louis, MO; and Kingston-
5-9
-------
Concentration ((.ig/ma) Y'/.'/••:/ 0 < x < 13 ISSS 13 < x < 17
x> 17
Figure 3-4a. 1999-2001 county-wide average annual mean PM2 5 concentrations (ug/m3)
for counties with PM2 5 monitors.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
Concentration ((.ig/m3)
047
Figure 3-4b. 1999-2001 highest county-wide 98th percentile 24-h average PM2 5
concentrations (jig/m3) for counties with PM2 5 monitors.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
3-10
-------
140-
£
100-
ii 8
-------
Harriman, TN); Particle Total Exposure Assessment Methodology (PTEAM) study (Riverside,
CA); Metropolitan Acid Aerosol Characterization Study (MAACS; Philadelphia, PA;
Washington, DC; and Nashville, TN); South Boston Air Quality and Source Apportionment
Study (Boston, MA); and NPMRMN (Phoenix, AZ). The remaining sites were part of the
24-Cities Study (Spengler et al., 1996).
Sufficient data are not yet available to permit the calculation of nationwide trends of PM2 5
and PM10_2 5; however, some general emerging patterns can be discerned. Darlington et al.
(1997) proposed that the consistent reductions in PM10 concentrations found in a wide variety of
environments ranging from urban to rural may have resulted from common factors or controls
that affected fine particles more strongly than coarse particles. This is because fine particles
have longer atmospheric lifetimes than coarse particles and can be transported over longer
distances and, hence, can affect larger areas. Apart from the IMPROVE network of monitoring
sites located mainly in national parks, the longest time series of PM25 concentration and
composition data have been obtained by the California Air Resources Board (CARB). Their
data show that annual average PM2 5 concentrations decreased from 1990 to 1995 by -50% in the
South Coast Air Basin, 35% in the San Joaquin Valley, 30% in the San Francisco Bay Area, and
35% in the Sacramento Valley (Dolislager and Motallebi, 1999). PM2 5 data were collected
continuously from 1994 to 1998 as part of the Children's Health Study in 12 southern California
communities (Taylor et al., 1998). Data obtained at all sites show decreases in PM25 ranging
from 2% at Santa Maria to 37% at San Dimas/Glendora from 1994 through 1998. These
decreases were accompanied by decreases in major components such as nitrate, sulfate,
ammonium, and acids. Based on the analysis of PM2 5 data sets collected prior to 1990, Lipfert
(1998) found that PM2 5 concentrations appear to have decreased by about 5% per year from
1970 to 1990 in a number of urban areas. These declines were also found to be consistent with
decreases in emissions from combustion sources over that time period.
PM10_2^5 Concentrations
By using AIRS data for 1999, 2000, and 2001 obtained by the PM10 and PM2 5 compliance
networks, it is possible to construct a picture of the distribution of coarse PM across the country.
This is accomplished by pairing data from 208 compliance monitoring sites in 196 counties
where PM10 and PM25 monitors are collocated and subtracting the mass concentrations of PM2 5
3-12
-------
from PM10. Annual mean concentrations were calculated on the basis of the latest 8 consecutive
quarters containing at least 11 samples per quarter. Nationwide annual mean PM10_2 5 values
calculated by this approach are shown in Figure 3-6a. Annual mean PM10_25 concentrations
ranged from ~1 to -50 |ig/m3; the nationwide median concentration was -10 |ig/m3; and 5% of
the sites had mean concentrations > 28 |ig/m3. The higher values occur mainly in the western
United States, particularly in California. The highest county-wide 98th percentile PM10_2 5
concentrations based on this same data set are shown in Figure 3-6b. Highest values in the
western United States are caused by dust raised locally either by natural means or by
anthropogenic activity. It is not clear what the contribution of PBP to these values may be.
Elevated dust levels are also found in southern Florida as the result of dust storms in North
Africa (Section 3.3.3) and trans-Atlantic transport. In many areas, combined errors in the PM25
and PM10 measurements may be similar to or even greater than the calculated PM10_2 5
concentrations. Because of this and other potential problems with this approach (Section 3.2.1),
these results should be viewed with caution.
3.2.1 Seasonal Variability in PM Concentrations
PM25
Aspects of the spatial and temporal variability of PM2 5 concentrations for 1999, 2000,
and 2001 in a number of metropolitan areas across the United States are presented in this and
following subsections. Data for multiple sites in 27 urban areas across the United States were
obtained from the AIRS data base and analyzed for their seasonal variations and for their spatial
correlations and spatial uniformity in concentrations. Selection of these 27 MS As was based on
the criteria that data be available for at least 15 days in each calendar quarter of either a 3-year
period (1999, 2000, and 2001) or a 2-year period (2000 and 2001) at three or more sites within
that MSA. In addition, a maximum of 11 sites per MSA were included for analysis. (In the
Chicago and St. Louis MSAs, the 11 sites having the most observations were selected from a
greater number of qualifying sites.) A number of aspects of the spatial and temporal variability
of the 1999 PM25 data set were presented in Rizzo and Pinto (2001), based in part on analyses
given in Fitz-Simons et al. (2000).
Information regarding the seasonal variability in PM2 5 concentrations in four U.S. MSAs
(Philadelphia, PA; Cleveland, OH; Dallas, TX; and Los Angeles-Long Beach, CA)
3-13
-------
_!/<$ ^*>
Alaska
Concentration (|ig/m3)
o > f
Hawaii V_\
''////, /A 0
-------
Alaska
Concentration (ng/ms) Y/////M 0 < x < 29
29 < x < 83
x>83
Figure 3-6b. 1999-2000 estimated county-wide highest 98th percentile 24-h PM10_25
concentrations (ug/m3) for counties with collocated PM2 5 and PM10 monitors.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
Annual mean PM2 5 concentrations at individual monitoring sites in the MSAs examined
ranged from ~6 to -30 |ig/m3. The lowest values were found in rural portions of the MSAs
examined, typically near the perimeter of the MSA. The two highest mean concentrations were
found in the Riverside and Los Angeles-Long Beach MSAs in southern California, whereas the
three lowest means were found in the Northwest (Portland, OR; Boise, ID; Seattle, WA). These
MSAs along the Eastern seaboard (Washington, DC; Philadelphia, PA; Norfolk, VA) tend to
have lower mean PM2 5 concentrations than those in the north-central United States
(Steubenville, OH; Cleveland, OH; Pittsburgh, PA; Chicago, IL; Detroit, MI; Gary, IN;
Appendix 3 A). Also, average PM2 5 concentrations tended to be lower in 1999, 2000, and
2001 in the urban areas given in Appendix 3 A compared to concentrations observed during
3-15
-------
a. Philadelphia, PA (2000-2001)
ARISID* 340071007 340155001 420170012 420450002 420910013 421010004 421010136
1234 1234 1234 1234 1234 1234 1234
b. Cleveland, OH (1999 - 2001}
ARISID* 390350013 390350038 390350060 390350065 390350066 390351002 390851001 390932003
Mean
Obs
SO
80—
60-
40-
20—
0 —
18.3
368
9.6
I
2
£
1
1
o.:
131
i,s
1
: 1
: 3
• I
'
I
8.4
53
.9
i!i
17.5
9.0
i
'II
!
1
3
f
'
4.7
32
5.4
III
1
3
f
I
5.0 14.0 1
51 342 2
.2 8.4 8
I II !
5.2
98
.a
I
I
1234 128* 1234 1234 1234 1234 123* 123*
Quarter
Figure 3-7a,b. Quarterly distribution of 24-h average PM2 5 concentrations for selected
monitors in the (a) Philadelphia, PA and (b) Cleveland, OH MSAs. Values
for the lowest, lower quartile, median, upper quartile and highest
concentrations are shown in the figures. The AIRS site ID number, annual
mean concentration (ug/m3), number of observations, and standard
deviation of the data are shown above the figures for each site.
3-16
-------
c. Dallas, TX (2000-2001)
ARISID3 480850005 481130020 481130035 481130050 481130057 481130069 481130087
Mean
Obe
SD
40-
11.s
234
S.B
12.4
87T
5.7
12.8
222
5.8
13.3
844
5.8
13.7
215
6.1
12.7
687
S.7
11.7
218
II
1234 1234 1234 1234 1234 1234 1234
Quarter
d. Los Angeles, CA (2000-2001)
ARISID* 060370002 0600371103 060371201 060371301 060372005 060374002
0.
Mean
Obs
SD
100 —
rs—
25 —
0 —
|
•i
•
I
D.E
54'
3.;
ll
I'
I
|
2
(
•
•
2.;
556
3.;
ll
i
!l
| 18.3 23.7 | !
I 217 237 I
j 11.7 13.4 j
ii-iiil
jTfTljm'ifl
ab
1.9
'-I
"I
2
6
1
II,
3.3
21
2.4
"
1
1234 1234 1234 1234
Quarter
1234 1234
Figure 3-7c,d. Quarterly distribution of 24-h average PM2 5 concentrations for selected
monitors in the (c) Dallas, TX 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 mean concentration (ug/m3), number of observations, and
standard deviation of the data are shown above the figures for each site.
3-17
-------
pollution-health outcome studies conducted in those five urban areas where these overlap
(Figure 3-5). It should be noted that there are no data demonstrating the comparability of the
monitors used in the studies shown in Figure 3-5 and the FRM.
The patterns of seasonal variability observed in the MSAs examined are complex. In the
Philadelphia, PA MSA, highest median concentrations occur at all but one site in the first quarter
(cf, Figure 3-7a). First quarter maxima are also observed at all sites in the Detroit, MI and
Chicago, IL MSAs (cf, Figures 3 A-l 1 and 3 A-14). The Cleveland, OH MSA (Figure 3-7b),
in contrast, does not have a clear seasonal pattern. In several MSAs examined in the southern
and eastern United States (Atlanta, GA; Baton Rouge, LA; Birmingham, AL; Columbia, SC;
and Washington, DC), highest median concentrations occur at all sites during the third calendar
quarter (i.e., summer months; cf, Appendix 3A). Sites in Dallas, TX, (Figure 3-7c), as in the
other southern cities mentioned above, generally exhibit third quarter median concentration
maxima. Highest median concentrations occur during the fourth calendar quarter in MSAs in the
western United States as in the Los Angeles, CA MSA (Figure 3-7d), although there are
exceptions at individual sites in the Riverside, CA MSA (Figure 3A-26).
Lowest median concentrations occur mainly during the first or fourth quarters at most sites
in the eastern United States, as well as in Cleveland, OH and Dallas, TX (Figures 3-7b and 3-7c),
whereas some occur during the second quarter (Philadelphia, PA; Figure 3-7a). Moving
westward, the seasonal pattern is not as distinct: lowest median concentrations occur in any
quarter, but usually in the second or third quarter, as in the Chicago, Detroit, and Los Angeles-
Long Beach (Figure 3-7d) MSAs. With the exception of Los Angeles, CA and Riverside, CA,
sites in the West show lowest median concentrations in the second quarter. In most of the MSAs
examined, seasonal variations follow a similar pattern at all of the sites within the MSA; but in a
few MSAs, there are noticeable differences in the seasonal pattern between sites. The large-
scale differences in seasonal variability between MSAs tend to follow differences in the major
categories of PM sources affecting the monitoring sites. Local heating by wood burning during
the colder months is practiced more widely in the western United States than in the eastern
United States. Prolonged winter stagnation events are also more common in western mountain
valleys during winter than in sections of the eastern United States located in relatively flat
topography. These conditions are also conducive to the formation of ammonium nitrate, which
is a major and sometimes the dominant contributor to PM25 in western valleys during the winter
3-18
-------
and in southwestern coastal areas during the fall (Kim et al., 2000). Hence, winter maxima and
greater variability in PM2 5 concentrations across sites are expected in the West due to the
influence of these local sources. On the other hand, photochemical production of secondary PM,
especially sulfate, occurs over wide areas in relatively homogeneous air masses during the
summer months in the eastern United States. Because sulfates (along with associated cations
and water) constitute the major fraction of summertime PM2 5 in the East, there is greater
uniformity in third quarter PM concentrations within eastern MSAs (cf, Appendix 3 A).
Maximum 24-h average concentrations shown in the box plots in Figures 3-7a to 3-7d and
in Figures 3A-1 to 3A-27 do not necessarily follow the same seasonal pattern as the median
concentrations. There is no clear relationship between the maximum and the median
concentrations evident in the Philadelphia, PA data set (Figure 3-7a). In Cleveland, OH
(Figure 3-7b), maximum concentrations occur during the second or fourth quarters, and highest
median concentrations generally occur during the third or first quarters. In Dallas, TX
(Figure 3-7c), maximum concentrations generally occur during the fourth quarter, but highest
median concentrations tend to occur during the second or third quarter. In the Los Angeles-Long
Beach MSA (Figure 3-7d), the maximum and highest median concentrations occur together in
the fourth quarter with the exception of one site. Peak individual concentrations likely reflect
the occurrence of transient events such as forest fires (mainly in the West) or episodes of
secondary PM production (mainly in the East). However, chemical analyses of filter samples or
other evidence should be used to determine specific causes in particular locations.
There have been a few studies that have characterized PM2 5 and PM10 concentrations in
major urban areas. The Metropolitan Aerosol Acidity Characterization Study (MAACS)
(Bahadori et al., 2000b) characterized the levels and the spatial and temporal variability
of PM2 5 PM10, and acidic sulfate concentrations in four cities in the eastern United States
(Philadelphia, PA; Washington, D.C.; Nashville, TN; and Boston, MA). Seasonal variations
in PM25 and PM10 concentrations obtained during the course of this study are shown in
Figure 3-8. The data for the four cities included in MAACS are presented in a box plot showing
the lowest, lowest tenth percentile, lowest quartile, median, highest quartile, highest tenth
percentile, and highest PM2 5 and PM10 values. Mean and highest PM2 5 and PM10 concentrations
peak during the summer in all four cities (in contrast with Figure 3-7a), although the seasonal
pattern in Boston appears to be more nearly bimodal with an additional winter peak.
3-19
-------
£ 'u
2 60-
c
.2 50-
viwl
OS
£ 40-
0>
o
c 30-
o
D
o 20-
•+*
ro
= 10-
",^
(i 0-
n
Philadelphia Washington
0 0
T T
o
o
o
I
o
o
0 ° I
T °
T -
o
o
o
I 1
,
I
0 0
o
•
^ 0
I
o
1 1
o
o
T
1 i 1
o
« T
T • I
A n
1
• JJU-.J.
; ly
o I o
o
o
I
1 1
Nashville
T
o
I _
I
o
o
Boston
I=IPM10
PM2.5
O
"
o
. 1
X (
o
1 I
o
T
JA
1
o
1
I
o
o
o
D
.
o
JL
I
o
1
O
1
r
O
O
J
D
1
I I
SPSU F W SPSU F W SPSU F W SPSU F W
Season
Figure 3-8. Seasonal 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.
SP = spring; SU = summer; F = fall; W = winter.
Source: Bahadori et al. (2000b).
PM1025
Of the 27 MS As selected for analysis of PM25 data (Appendix 3 A), 17 have multiple sites
satisfying the PM10_25 data completeness criteria. A subset of 53 qualifying sites were selected
for analysis of PM10_2 5 data. Each qualifying site has 11 or more observations per calendar
quarter for 12 consecutive quarters (1999 to 2001, 2 MSAs), eight quarters (1999 + 2000 or
2000 + 2001, 7 MSAs) or four quarters (2000 or 2001, 8 MSAs). In addition, data from sites
within the same MSA represent the same year (or years) of observations, so that intersite
comparisons within an MSA are based on the same reporting period. However, comparisons
between different MSAs can involve different annual reporting periods. As can be seen by
3-20
-------
comparing Figures 3-7a,b,c,d and Figures 3-9a,b,c, the number of observations used for
summarizing PM10_2 5 data is much lower than that measured for PM2 5. Unfortunately, fewer
qualifying MSAs for PM10.2 5 result in geographical coverage that is more spotty than with PM2 5.
Five of the seven MSAs representing the eastern United States for PM25, for example, did not
meet completeness criteria for PM10_2 5, leaving only Tampa, FL and Columbia, SC to represent
the East. Secondly, it can readily be seen, e.g., for Cleveland, OH (Figure 3-9a) and, also, in
Figures 3A-28 through 3A-44, that a number of PM10_25 concentrations are negative. For
example, in 2001, about one-fourth of all PM10_25 concentrations at the three qualifying sites in
the Gary, IN MSA were negative. (The negative estimates have been included in the calculation
of mean concentrations.) There are a number of reasons for the negative concentration
estimates, many of which arise because the ratios of PM25 to PM10 are based on two independent
measurements. Measurement imprecision plays a role when the ratios are large and
concentrations are small. Differences in the behavior of semivolatile components in the two
samplers could occur; and the results may also reflect errors in sampler placement, field,
laboratory, or data processing procedures. Thus, caution should be exercised when attempting to
interpret results for PM10_2 5 based on current network collocated PM2 5 and PM10 monitors.
Annual mean PM10_2 5 concentrations at individual monitoring sites range from ~6 |ig/m3
(Portland, OR) to -33 |ig/m3 (Riverside, CA). (Gary, IN sites were excluded because of
numerous negative PM10_2 5 concentrations, and one site in Riverside, CA affected by a local dust
event was excluded.) The three highest annual mean concentrations are observed in Riverside,
CA; Los Angeles-Long Beach, CA; and Salt Lake City, UT; and the lowest (excluding Gary)
were observed in Portland, OR. For the remaining MSAs, there does not appear to be a
significant geographical trend associated with the annual mean PM10_2 5 concentration. Within
MSAs, the lowest concentrations are frequently observed at sites near the perimeter of the MSA,
although it must be noted that the number of sites is limited.
Within each MSA, collocated PM2 5 and PM10_2 5 concentrations, averaged over the same
years, were compared. The mean PM2 5:PM10_25 ratio was calculated for each of the 17 MSAs
using as many sites as possible. The median mean PM25:PM10_25 ratio for the 17 MSAs was 1.2.
For eight MSAs (Tampa, FL; Columbia, SC; Louisville, KY; Chicago, IL; Gary, IN; Milwaukee,
WI; Steubenville, OH; Portland, OR), the mean PM2 5 concentration exceeded the mean PM10_2 5
concentration. For an additional eight MSAs, the PM2 5 and PM10_2 5 concentration means were
3-21
-------
a. Cleveland, OH (2000 - 2001)
AIRSIDf 390350013 390360038 390360046 390360060 390360065 390851001
Mean
Obs
SD
ISO-
'S QQ-
E
OJ
*¥ 50-
Q
EL
0-
-5Q~
!
•
.
I
tA
16
7.4
I!
'
|
1
f
1
I!
8.6
14
1.6
!i
'I
!
16.8
112
9.0
'ill
!
!
21.3 16.8 7.2
113 111 109
16.4 9.5 4.9
iWfiit,,
Quarter
b. Dallas, TX (2001)
AIRS IDS 481130020 481130035 481130050 481130057
Mean
Obs
SD
^)
o
s"
11.2
60
5.4
12.9
66
8.T
14.5
56
8.4
19.1
65
10.5
234 12
Quarter
Figure 3-9a,b. Quarterly distribution of 24-h average PM10_2 5 concentrations for selected
sites in the (a) Cleveland, OH; (b) Dallas, TX 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.
3-22
-------
c. Los Angeles, CA (2001)
AIRSID* 060370002 060371002 060371103 060374002
Mean
Obs
SD
60 H
50-
n 40-
« 30-
o
20-
10-
O-l
24.1
56
11.7
15.3
56
6.7
21.4
57
8.7
16.1
53
6.6
1234 1234 1234 1234
Quarter
Figure 3-9c. Quarterly distribution of 24-h average PM10_2 5 concentrations for selected
sites in the 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: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
the same (within one SD). Salt Lake City was the only MSA for which the mean PM10_2 5
concentration exceeded the mean PM2 5 concentration.
Information regarding the seasonal variability in PM10_2 5 concentrations in three MSAs
(Cleveland, OH; Dallas, TX; Los Angeles-Long Beach, CA) is summarized in Figures 3-9a
through 3-9c. The figures show lowest, lower quartile, median, upper quartile, and highest
concentrations for each calendar quarter of 2000 and 2001 for the Cleveland, OH MSA (six
sites); and of 2001 for the Dallas, TX and Los Angeles, CA MSAs (four sites each). Data for
3-23
-------
these and an additional 13 MS As, as well as additional descriptions of the data, are given in
Appendix 3 A. The seasonal pattern for PM10_2 5 median concentrations is different from that
of PM25 (Figures 3-7a,b,c,d, and Appendix 3A). Most seasonal median maxima in PM10_25 occur
during the second or third calendar quarters, i.e., spring and summer months (45% and 36% of
all sites, respectively) as seen, for example, at most sites in Figures 3-9. Contributions from
bioaerosols during the spring blooming season, which mostly affect PM10_2 5 concentrations,
might be a significant contributing factor in the second quarter PM10_2 5 median maxima in some
regions of the United States. Lowest median concentrations occur mainly in the first or fourth
quarters 62% and 23% of all sites, respectively). Cleveland, OH (Figure 3-9a) and Tampa, FL,
where lowest medians generally are observed in the third quarter, are exceptions (Figure 3 A-29).
At no site was the lowest median PM10_2 5 concentration observed in the second quarter. In
comparison to PM10_2 5, seasonal PM2 5 median maxima mostly occurred in the first or third
quarters and PM2 5 median minima are mostly observed in the second and fourth quarters.
With few exceptions, collocated PM2 5 and PM10_2 5 seasonal medians at individual sites peak in
different quarters. Likewise, at a given site, the lowest median concentrations for PM2 5 and
PM10_2 5 rarely occur in the same quarter. In MSAs with several PM10_2 5 sites, the seasonal
patterns are typically seen at all sites within the MSA. In the Dallas, TX MSA, the maximum
and minimum estimated PM10_2 5 concentrations both occur in the first quarter at all four sites.
The ratio of PM2 5:PM10_25 seasonal median concentrations peak in the first quarter for
MSAs in the central and north-central United States and tend to peak in the fourth quarter for
western states. The largest ratios of PM25:PM10_25 seasonal median concentrations are observed
in the central and north-central MSAs (Chicago, Cleveland, Detroit, and St. Louis), whereas
smaller ratios are found in the western and southern United States.
As can be seen from Figure 3-9a and Figures 3 A-40 and 3 A-43, the maxima for PM10_2 5
concentrations exceeded 100 |ig/m3 in Cleveland, OH and Salt Lake City, UT, and 500 |ig/m3 in
Riverside, CA. This latter value was related to a dust storm. In several urban areas (Cleveland,
OH; Detroit, MI; Chicago, IL; Dallas, TX; and Riverside, CA) maxima in PM10_2 5 concentrations
were larger than those for PM2 5. However, there is no clear geographic pattern.
The results described above should be viewed with caution because of inherently large
errors in the technique used to derive them.
3-24
-------
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). Also shown
are concentrations predicted from the log-normal distribution, using geometric mean values and
standard deviations derived from the data. 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 |ig/m3
for PM2 5 and 8.6 ± 7.5 |ig/m3 for PM10. Maximum day-to-day concentration differences are
54.7 |ig/m3 for PM25 and 50.4 |ig/m3 for PM10.
350
geometric mean = 15.2 |jg/m3
on = 1.69
50 60
Concentration (|jg/m3)
Figure 3-10. Frequency distribution of 24-h average PM25 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.
Source: Bahadori et al. (2000b).
3-25
-------
Different patterns are observed in data collected elsewhere in the United States. PM2 5
concentrations obtained in Phoenix, AZ from 1995 through 1997 (Zweidinger et al., 1998) are
summarized in Figure 3-11; and frequency distributions of PM2 5 concentrations obtained in
Phoenix are shown in Figure 3-12. Mean ± SD day-to-day concentration differences in this data
set are 2.9 ±3.0 |ig/m3; the maximum day-to-day concentration difference was 23 |ig/m3. PM25
and PM10_2 5 data were obtained with dichotomous samplers at a number of sites in California on
a sampling schedule of every 6 days from 1989 through 1998. Histograms showing the
frequency distribution of the entire set of PM25 and PM10_25 concentrations obtained by the
CARB network of dichotomous samplers from 1989 to 1998 are shown in Figures 3-13 and
3-14. Also shown are log-normal distributions generated by using geometric means and
standard deviations derived from the data as input. Although the data for both size fractions
appear to be reasonably well simulated by the function, data obtained at individual locations may
not be. Data showing the seasonal variability of PM2 5 obtained at Riverside-Rubidoux are
summarized in box plot form in Figure 3-15. The frequency distribution of PM2 5 concentrations
obtained at Riverside-Rubidoux from 1989 to 1994 is shown in Figure 3-16. It can be seen that
the data are not as well fit by a log-normal distribution as are data shown in Figure 3-10, partly
as the result of a significant number of days when PM25 values exceed 100 |ig/m3.
40
30
c
o
~ 20
0)
o
E
O
to
CM
10-
Phoenix, AZ
PM2.5
(n = 876)
n
Y
T
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: Zweidinger etal. (1998).
3-26
-------
200
to 150 H
"5.
E
<55ioo
M—
O
0
PM2.5
geometric mean = 10.5 pg/m3
a, = 1.70
5 10 15 20 25 30 35 40
Concentration (|jg/m3)
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: Zweidinger et al. (1998).
3000
0
PM
2.5
geometric mean = 12,8 pg/m3
on = 2.29
t t t
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Concentration (|jg/m3)
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.
3-27
-------
3000
2500-
PM-,0-2.5
geometric mean = 15.7 |jg/m3
ern = 2.26
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Concentration (|jg/m3)
Figure 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.
IUU
140-
-, 120-
co
E
5 100-
c
5 80 -
s
c
g 60-
c
o
0 40-
20 -
0 -
Fine
(n = 382)
T
/
^
¥
I
-1 L
•
r
n
• «
w
V i \
»
t
i i
Jan - Mar Apr - Jun Jul - Sept Oct - Dec
1st Qtr 2nd Qtr 3rd Qtr 4th Qtr
Figure 3-15. Concentrations of 24-h average PM25 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.
3-28
-------
100
PM
2,5
geometric mean = 26.6 |jg/m3
a =2.10
20 40
60 80 100 120 140
Concentration (|jg/m3)
160 180 200
Figure 3-16. Frequency distribution of 24-h average PM25 concentrations measured
at the Riverside-Rubidoux site from 1989 to 1994.
An examination of the data from the four MAACS cities, Phoenix, AZ, and Riverside, CA,
indicates that substantial differences exist in aerosol properties between the eastern cities
(MAACS) and the western cities (Phoenix, AZ; Riverside, CA). Fine-mode particles account for
most of the PM10 mass observed in the MAACS cities and appear to drive the daily and seasonal
variability in PM10 concentrations in the East, whereas coarse-mode particles represent a larger
fraction of PM10 mass in Phoenix and Riverside and drive the seasonal variability in PM10 seen in
the West. The average ratio of PM2 5 to PM10 concentrations is much larger in the MAACS cities
of Philadelphia, PA (0.72), Washington, DC (0.74), and Nashville, TN (0.63) than in either
Phoenix, AZ (0.34) or Riverside, CA (0.49). Differences between median and maximum
concentrations in any size fraction are much larger at the Riverside site than at either the
MAACS or Phoenix sites. Many of these differences could reflect the more sporadic nature of
dust suspension at Riverside. In addition, the seasonal variability of PM2 5 concentrations seen in
Phoenix and Riverside appears to be different from that observed in the MAACS cities. These
considerations demonstrate the hazards in attempting to extrapolate conclusions about the nature
of variability in aerosol characteristics inferred at one location to another.
3-29
-------
3.2.2 Diurnal (Circadian) Variability in PM Concentrations
The variability of PM concentrations on time scales shorter than a day can, in principle, be
characterized by measurements made by continuous samplers (e.g, TEOMs and p-gauge
monitors that are currently used to provide Air Quality Index [AQI] information to the public).
A description of these methods was provided in Section 2.2.9. However, as shown in Chapter 2,
continuous methods are subject to artifacts in large part because of the heating of their inlets to
remove water, which results in the loss of components such as ammonium nitrate and, also
semivolatile organic compounds (see Sections 2.2.2.1 and 2.2.3 for further details concerning the
chemistry of volatilizable components). Consequently, caution should be used in interpreting
results obtained by these techniques. It should be remembered that the FRMs are also subject to
artifacts; therefore, caution should also be exercised in interpreting results obtained by them.
The composite diurnal variation of PM25 concentrations obtained across the continental
United States by 31 TEOM and P-gauge monitors reporting to AIRS in 1999 is shown in
Figure 3-17. As can be seen in that figure, there is a distinct pattern with maxima occurring
during the morning and evening. Notable exceptions to this pattern occur in California, where
broad nighttime maxima and daytime minima occur, which may be related to the use of
p-gauges with unheated inlets there. It should be noted that, in examining the diurnal variations
shown in Figure 3-17, there is substantial day-to-day variability in the diurnal profile of PM25
measured at the same location, which is smoothed out after a suitably long averaging period is
chosen. The large ratio of the interquartile range to the median values supports the view that
there is substantial variability in the diurnal profiles.
The diurnal variability of PM components is determined by interactions between variations
in emissions, the rates of photochemical transformations, and the vertical extent and intensity of
turbulent mixing near the surface. Wilson and Stockburger (1990) characterized the diurnal
variability of sulfate and lead (Pb) in Philadelphia. At that time, Pb was emitted mainly by
motor vehicles. Pollutants emitted mainly by motor vehicles, such as carbon monoxide (CO),
show two distinct peaks occurring during the morning and evening rush hours (see Chapter 3,
U.S. Environmental Protection Agency, 2000b). Pollutants, such as sulfate, which are
transported long distances in the free troposphere (i.e., above the planetary boundary layer), tend
to be mixed downward and have their highest concentrations during the afternoon when the
intensity and vertical extent of turbulent mixing (and chemical oxidation) are greatest.
3-30
-------
30-
CO
£
2.
C
O
'-4-1
^^
d>
O
C
0
o
in
r>i
0.
o-
6685 RRR7 „„„„
I
03
*
—
3
6140 67m 6708 6709
-
•K
' — '
— -
+
_
.
I
' '
1 ,
~r
_
^
—
70
+
—
e
7
-^"
69
-
9
__^-
-
1
-
e
~~-~
64
_i
—
5
6
57
+
—
D
£
~-
54
+
—
6576 6571 6605 664
—
_
_
~
—
3 6649
~
—
-
69
+
_
e
0
^ — •
70
+
_
5 £
70
j_
T
—
1 6
69
4-
-
3 c
oy
i.
T^
—
0
e
70
_
_
4
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).
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
concentrations produced by residential heating (e.g., from wood burning), on the other hand,
reach maximum levels during the night when mixing heights are much lower than during the
day.
Although the interquartile ranges for hour-to-hour changes in PM2 5 concentrations shown
in Figure 3-17 encompass several |ig/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 |ig/m3. Maximum positive excursions were much
larger, ranging from 27 |ig/m3 in the Northeast up to 220 |ig/m3 in the Southwest and with
maximum excursions in other regions all less than 125 |ig/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
during passage through the sampler inlets and on the TEOM measurement elements. These
3-31
-------
factors add uncertainty to the interpretation of the hour-to-hour changes that are observed, as
discussed in Chapter 2. However, because of the tendency of these monitoring instruments to
lose material by evaporation, the concentrations reported during excursions probably represent
lower limits for the true values that were present.
3.2.3 Relationships Among Particulate Matter in Different Size Fractions
Relations Among PM2 s, PM10_2^S, andPM10
Data obtained in 1999 by collocated PM2 5 and PM10 FRM monitors have been used to
calculate the ratio of PM2 5 to PM10 concentrations and correlations among PM25, PM10_25,
and PM10 concentrations. Results are shown in Table 3-1 for each of the seven aerosol
characteristic regions identified in Chapter 6 of the 1996 PM AQCD. As can be seen from the
table, the ratio of PM2 5 to PM10 concentrations tends to be higher in the eastern United States
than in the western United States. This general pattern and the values are consistent with that
found for the studies included in Appendix 6A of 1996 PM AQCD. In that compilation based on
the results of studies using dichotomous samplers, the mean ratio of PM25 to PM10 was 0.75 in
the East, 0.52 in the central United States, and 0.53 in the western United States. Although a
large number of paired entries have been included in Table 3-1, seasonal variations and annual
averages in a number of regions could not be determined from the data set because of data
sparseness, mainly during the early part of 1999. It also can be seen in Table 3-1 that the ratio
of PM25 to PM10 was > 1 for a few hundred measurements. There are a number of reasons for
these results, as mentioned in Section 3.2.1 in the discussion on PM10_25 concentrations.
Ultrafine Particle Concentrations
Data for characterizing the concentrations of ultrafme particles (< 0.10 jim Da) and the
relations between ultrafme particles and larger particles are sparse. Although ultrafme particles
dominate particle number concentrations, they make very minor contributions to PM2 5 mass.
For example, Cass et al. (2000) found that particles between 0.056 and 0.1 jam Da contributed
only 0.55 to 1.16 |ig/m3 at several sites in southern California. Perhaps the most extensive data
set for ultrafme particle properties is that described by Woo et al. (2001) for a site located 10 km
to the northwest of downtown Atlanta, GA. Size distributions from 3 to 2000 nm were measured
every 12 min for 24 months beginning in August 1998. Approximately 89% of the total number
3-32
-------
TABLE 3-1. DISTRIBUTION OF RATIOS OF PM?. TO PM1(1 AND CORRELATIONS BETWEEN PM?. AND PM
10
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
Percentiles
Region
Northeast
Southeast
Industrial Midwest
Upper Midwest
Southwest
Northwest
Southern California
Mean
0.70
0.70
0.70
0.53
0.38
0.50
0.47
Total:
Sites
45
76
92
39
23
73
36
384
Values
1433
2823
4827
1446
701
3300
1813
16343
95
0.97
1.27
1.09
0.92
0.51
0.67
0.70
90
0.95
1.06
0.88
0.84
0.51
0.65
0.57
75
0.77
0.74
0.78
0.62
0.47
0.56
0.55
50
0.67
0.63
0.68
0.49
0.40
0.49
0.48
25
0.60
0.54
0.59
0.44
0.31
0.44
0.44
10
0.51
0.46
0.51
0.34
0.23
0.39
0.31
5
0.48
0.43
0.47
0.24
0.23
0.36
0.24
PM25:PM10
0.72a
0.69a
0.71a
0.35a
0.63a
0.69a
0.70a
Correlations
PM25:PM1025
0.02
-0.04a
0.17a
-0.02
0.49a
0.07a
0.19a
PM10.2.5:PM10
0.71a
0.69a
0.81a
0.93a
0.99a
0.77a
0.83a
aResults considered to be significantly different from zero at the a = 0.01 level.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
-------
of particles were found to be smaller than 100 nm; whereas 26% were < 10 nm. Concentrations
tend to be lower during the summer than during the winter. No correlation was found between
number concentration and either volume or surface area for particle sizes up to 2 jim. Because
the total number of particles is concentrated in the smallest size ranges, these results also
indicate that fine particle mass does not correlate with the number of ultrafine particles. The
high time resolution of the measurements allows some inferences to be made about the possible
sources of the ultrafine particles. The number of particles > 10 nm tends to peak during the
morning rush hour around 8:00 am and then to decrease through the day and to increase again
after 6:00 pm, consistent with a traffic-related source. Particles < 10 nm tend to peak during the
mid-afternoon, consistent with nucleation involving products of active photochemistry
(McMurry et al., 2000). More direct relationships between particle mass observed in different
size ranges can be obtained using multi-stage impactors. Keywood et al. (1999) found a
correlation between PM25 and PM015 of-0.7, whereas they found correlations of-0.96
between PMl and PM2 5 and between PM2 5 and PM10 based on samples collected by MOUDIs
(Multiple Orifice Uniform Deposit Impactors) in six Australian cities.
3.2.4 Relationships Between Mass and Chemical Component Concentrations
Time series of elemental composition data for PM2 5 based on X-ray fluorescence (XRF)
analyses have been obtained at a number of locations across the United States. Time series of
components of the OC fraction of the aerosol have not yet been obtained. The results of XRF
analyses for the composition of the inorganic fraction of PM25 and PM10_25 are presented in
Table 3-2 for Philadelphia, PA and in Table 3-3 for Phoenix, AZ. Frequency distributions
for PM2 5 concentration data collected at these sites were shown in Figures 3-10 and 3-11.
All XRF analyses were performed at the same U.S. EPA X-ray spectrometry facility in Research
Triangle Park, NC. Data shown in the first column of Table 3-2 are based on analyses of filters
collected over 3 years (April 1992 to April 1995, labeled a) at the Castor Avenue Laboratory at
the PBY site in southwestern Philadelphia. These data and data for PM10 were collected using
Harvard impactors. Data for PM2 5 and PM 10_2 5 shown in the second and third columns were
obtained by the City of Philadelphia from July 25 to August 14, 1994, using a modified
dichotomous sampler (VAPS). The samples at the Phoenix site were collected in 1996 and 1997
using the same type of dichotomous sampler used in the shorter term study in Philadelphia.
3-34
-------
TABLE 3-2. CONCENTRATIONS (ng/m3) OF PM2 5, PM10 2 5, AND SELECTED ELEMENTS (ng/m3) IN THE
PM2 5 AND PM10 2 5 SIZE RANGES WITH STANDARD DEVIATIONS (SD) AND CORRELATIONS BETWEEN
ELEMENTS AND PM2 5 MASS IN PHILADELPHIA, PA*
n = 1105 Cone (ng/m3) ± SD (unc) r
PM25' 17 ± 0.9 (0.8) x 103 1
Al 4.0 ±56 (31) 0.1
Si 116 ±107 (21) 0.51
P 8.6 ±14 (10) 0.31
S 2100 ± 1610(143) 0.92
Cl 5.1 ±35 (3.4) -0.01
K 60.4 ± 45 (4.7) 0.5
Ca 47 ± 33 (4.2) 0.39
Ti 4.9 ±5.2 (4.1) 0.44
V 8.8 ±8.7 (1.8) 0.37
Cr 0.7 ±1.1 (0.7) 0.15
Mn 3.1 ±2.2 (0.8) 0.39
Fe 109 ±71 (10.5) 0.5
Co 0.1 ±1.8 (1.4) 0.04
Ni 7.3 ±8.4 (1.4) 0.22
Cu 4.8 ±4.9 (1.1) 0.25
Zn 36.9 ±44 (3.7) 0.21
As 0.6 ±1.4 (1.2) 0.18
Se 1.5 ±1.3 (0.6) 0.63
Br 5.0 ±11.7 (0.9) 0.11
Pb 17.6 ±22 (2.5) 0.19
n = 20 Cone (ng/m3) ± SD (unc) r
PM252 29.8 ± 14.7 (1.1) x 103 1
Al 109 ±61 (21) 0.15
Si 191 ± 134 (26) 0.22
P 15 ± 4.3 (2.7) 0.72
S 3190 ±1920 (207) 0.91
Cl 23 ±28 (5.5) 0.19
K 68 ±21 (6.4) 0.31
Ca 63 ± 33 (9.0) -0.02
Ti 8.7 ± 4.7 (9.0) 0.47
V 9.7 ±7.1(2.9) 0.38
Cr 1.4 ±1.2 (2.9) 0.09
Mn 3.2 ±1.5 (1.6) 0.43
Fe 134 ±49 (0.5) 0.48
Co 0.8 ±0.7 (8.5) 0.58
Ni 8.5 ±5.6 (0.3) 0.61
Cu 7.7 ±3.8 (0.7) 0.22
Zn 56 ± 37 (4.8) 0.22
As 0.4 ±1.0 (1.0) -0.02
Se 1.3 ±0.8 (0.4) 0.65
Br 14 ±12 (1.3) 0.21
Pb 28 ± 24 (2.4) 0.26
n = 20 Cone (ng/m3) ± SD (unc) r
PM10.252 8.4 ± 2.9 (0.4) x 103 1
Al 325 ±241(99) 0.89
Si 933 ±652 (231) 0.9
P 28 ±9.4 (7.1) 0.78
S 38 ±45 (71) -0.15
Cl 47 ±48 (5.8) -0.11
K 100 ±66 (10) 0.81
Ca 421 ±192 (31) 0.81
Ti 30 ±17 (5.6) 0.9
V 3.2 ±2.2 (1.5) 0.66
Cr 1.0 ±5.0 (0.9) 0.43
Mn 6.3 ±4.1(0.6) 0.9
Fe 352 ± 156 (24) 0.9
Co -0.2 ±0.5 (0.3) -0.10
Ni 2.0 ±1.4 (0.3) 0.08
Cu 14 ±12 (1.1) -0.05
Zn 52 ±43 (4.7) -0.03
As 0 ± 0.5 (0.5) 0.07
Se -0.1 ±0.2 (0.2) -0.24
Br 3.0 ±2.5 (0.5) -0.10
Pb 13 ±11 (1.3) 0.1
'Data obtained at the Presbyterian home (PBY) site in Philadelphia from April 1992 to April 1995 with Harvard impactors.
2Data obtained at the Castor Avenue Laboratory, North Central Philadelphia, from July 25 to August 14, 1994 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
Ti
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
0.8
-0.01
0.38
0.69
0.64
0.5
0.4
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)
5350 ±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
0.9
0.9
0.51
0.76
0.91
0.9
0.38
0.7
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: Calculations based on data obtained by Zweidinger et al. (1998).
These data are shown to give an idea of the range of concentrations found in studies conducted
more recently than those shown in Appendix 6A of the 1996 PM AQCD. The speciation
network should provide more thorough coverage of the composition of particles in the PM25 size
3-36
-------
range across the United States. Results from the pilot study for the speciation network are given
in Appendix 3B.
As can be seen from inspection of Tables 3-2 and 3-3, the analytical uncertainty (given in
parentheses next to concentrations) as a fraction of the absolute concentration is highly variable.
It exceeds the concentration for a number of trace metals whose absolute concentrations are low,
whereas it is very small for abundant elements such as sulfur.
Sulfur (S) is the major element analyzed in the PM25 size fraction in the two Philadelphia
studies and is highly correlated with PM2 5; however, its abundance is roughly two orders of
magnitude lower in the PM10_2 5 size range and is negatively correlated with PM10_2 5. The
concentrations of the crustal elements Al, Si, K, Ca, and Fe are much higher in the PM10_25 size
range than in the PM2 5 size range and are well correlated with PM10_2 5. A number of trace
elements (e.g., Cr, Co, Ni, Cu, Zn, As, Se, and Pb) are detectable in the two PM2 5 data sets, and
the concentrations of many of these elements are much greater than the uncertainty in their
determination. Except for Co, As, and Se (which are not detected in the PM10_2 5 samples), the
concentrations of many elements (Cr, Zn, and Pb) are comparable in the PM25 and PM10_25 size
ranges. The concentration of Cu is significantly higher in the PM10_25 size range, whereas the
concentration of Ni is smaller in the PM10_2 5 size range than in the PM25 size range. Problems
with detecting and quantifying concentrations of Co, As, and Se can result in artificial reduction
of their correlation coefficients with PM25.
There are a number of distinct differences between the PM25 sets for Philadelphia and
Phoenix. For instance, sulfate and associated cations and water that would be expected to
correspond to the measurement of S appear to constitute a major fraction of the composition of
the PM in the Philadelphia data set; whereas they appear to constitute a much smaller fraction of
the PM in the Phoenix data set. The highest PM2 5 values were observed in Philadelphia during
episodes driven by high sulfate abundances, whereas those in Phoenix were driven by raised soil
dust. The concentration of S in Phoenix is much lower in the Phoenix PM2 5 data set than in
either Philadelphia PM2 5 data set, even though it represents the most abundant element and it is
only weakly correlated with PM2 5. This is in marked contrast to the data shown for Philadelphia
and also to data found at other eastern sites. It is not clear what the reasons are for this finding.
As in Philadelphia, the concentration of S in Phoenix is higher in the PM2 5 size range than in
the PM10_2 5 size range. Trace metals (e.g., Cr, Co, Ni, Cu, Zn, As, and Pb) are not well
3-37
-------
correlated (0.04 < r < 0.25) with PM2 5 in the Philadelphia data set, whereas they are more
variably correlated (0.01 < r < 0.69) with PM25 in the Phoenix data set. The uncertainty in the
concentration measurement most probably plays a role in determining a species' correlation
with PM2 5, especially when the analytical uncertainty is high relative to concentration as it is for
a number of elements in the data shown in Tables 3-2 and 3-3. Concentrations of Al, Si, K, Ca,
and Fe are again much higher in the PM10_2 5 size range than in the PM2 5 size range and are
strongly correlated with PM10_2 5 in both data sets.
There are also similarities in the PM2 5 data sets for Philadelphia and Phoenix. Crustal
elements are not as well correlated with PM2 5 as they are with PM10_2 5 in both data sets. The
concentrations of trace metals (Cr, Ni, Cu, and Zn) in PM2 5 are similar in Philadelphia and
Phoenix. It can also be seen that their concentrations are of the same order of magnitude in
both PM2 5 and PM10_2 5. Concentrations of Cu are noticeably higher in PM10_2 5 than in PM2 5 in
both Philadelphia and Phoenix. These results are consistent with those of many monitoring
studies shown in Appendix 6A of the 1996 PM AQCD, which also show that concentrations of
these metals are of the same order of magnitude in both size fractions and that concentrations of
Cu tend to be higher in PM10_2 5 than in PM2 5.
One study suggests that the partitioning of trace metals between the fine and coarse
fractions varies with PM concentration. Salma et al. (2002) determined the size distribution of a
number of trace elements at four sites characterizing environments ranging from the urban
background to an urban traffic tunnel in Budapest, Hungary. Although S, K, V, Ni, Cu, Zn, As,
and Pb were found mainly in the fine fraction at the urban background site, their mass median
aerodynamic diameters increased with increasing PM concentrations until they were all found
mainly in the coarse fraction in the traffic tunnel. They also found that Na, Mg, Al, Si, P, Ca, Ti,
Fe, Ga, Sr, Zr, Mo, and Ba were concentrated mainly in the coarse fraction at all four sites and
that their mass median aerodynamic diameters increased with increasing PM concentrations.
The mean concentration of Pb observed in the methods evaluation study for the speciation
network was only about 5 ng/m3 in Philadelphia during the first half of 2000 (Appendix 3B),
whereas its concentration was about three times higher during studies conducted during the early
1990s (Table 3-3). In a study conducted in the greater Philadelphia area during the summer of
1982, Dzubay et al. (1988) found concentrations of Pb of about 250 ng/m3, or about fifty times
higher than observed in 2000. The mean Pb concentration was about 3 ng/m3 at the Phoenix site
3-38
-------
included as part of the same methods evaluation study for the speciation network; however, the
mean Pb concentration was 39 ng/m3 during an earlier study conducted during 1989 and 1990 in
Phoenix (Chow et al., 1991). These changes in Pb concentrations are consistent with those in
many other urban areas for which monitoring studies have been conducted during the late 1970s
and 1980s (cf, Appendix 6 A of the 1996 PM AQCD) and for which there are data given in
Appendix 3B. It should be remembered that the older studies were conducted while Pb was still
used as a gasoline additive. The ratio of Pb in PM25 to Pb in PM10_2 5 was also much higher in
the older studies than in the more recent ones, reflecting the importance of combustion as its
source. Smaller decreases are apparent in the concentrations of other trace metals such as Cu,
Ni, and Zn between studies conducted in the early 1980s and in the methods evaluation study for
the speciation network conducted in 2000.
Some indication of the sources of metals such as Pb, Cu, Cd, and Zn in current ambient
PM2 5 and PM10_2 5 samples can be obtained by examining their sources in urban runoff. The
sources of these elements in urban runoff were found to be the weathering of building surfaces,
motor vehicle brake and tire wear, engine oil and lubricant leakage and combustion, and wet and
dry atmospheric deposition (Davis et al., 2001). Once deposited on the ground, these elements
can be resuspended with other material as PM2 5 and PM10_2 5, although research is needed into
the mechanisms of how this is accomplished. Wind-abrasion on building siding and roofs
(coatings such as Pb paint and building material such as brick, metal, and wood siding); brake
wear (brake pads contain significant quantities of Cu and Zn); tire wear (Zn is used as a filler in
tire production); and burning engine oil could all produce particles containing these metals,
especially Zn.
Data for the chemical composition of ambient ultrafine particles are sparse. In a study
conducted at several urban sites in Southern California, Cass et al. (2000) found that the
composition of ultrafine particles ranged from 32 to 67% organic compounds, 3.5 to 17.5%
elemental carbon, 1 to 18% sulfate, 0 to 19% nitrate, 0 to 9% ammonium, 1 to 26% metal oxides,
0 to 2% sodium, and 0 to 2% chloride. Thus carbon, in various forms, was found to be the major
contributor to the mass of ultrafine particles. However, ammonium was found to contribute 33%
of the mass of ultrafine particles at one site in Riverside. Iron was the most abundant metal
found in the ultrafine particles. Chung et al. (2001) found that carbon was the major component
of the mass of ultrafine particles in a study conducted during January of 1999 in Bakersfield,
3-39
-------
CA. However, in the study of Chung et al., the contribution of carbonaceous species (OC and
EC; typically 20 to 30%) was much lower than that found in the cities in Southern California.
They found that calcium was the dominant cation, accounting for about 20% of the mass of
ultrafme particles in their samples. Sizable contributions from silicon (0 to 4%) and aluminum
(6 to 14%) were also found. Further studies, including scanning electron microscopy, may be
needed to quantify the role of coarse particle bounce from the upper stages of their MOUDI
impactor.
Gone et al. (2000) measured the size distribution of trace elements from 0.056 jim to
1.8 |im Da in Pasadena, CA and in the Great Smoky Mountains National Park, TN. They found
that elements identified as being of anthropogenic origin had mass median diameters < 1 jim
PM; whereas elements of crustal origin generally had a mass median diameter > 1 jim.
Concentrations of trace metals were much higher in the accumulation mode than in the ultrafme
mode in both study areas. In PMb 76% of Cr, 95% of Fe, 94% of Zn, 89% of As, and 79% of
Cd at the Tennessee site were found in the accumulation mode; and 70% of Fe, 85% of Zn, 92%
of As, and 84% of Cd were found in the accumulation mode in Pasadena. Fe was the most
abundant metal found in the ultrafme particles. The abundance of crustal elements, such as Al,
declined rapidly with decreasing particle size at both locations; and Al in PMX probably
represented the lower tail of the coarse PM mode. However, on two days at Pasadena there were
increases in the concentration of Al in ultrafme particles that were associated with increases in
Sc and Sm. The latter two elements originate exclusively from crustal material (Gone et al.,
2000).
3.2.5 Spatial Variability in Particulate Matter and its Components
PM2S
Aspects of the spatial variability of PM2 5 concentrations on the urban scale are examined
in this section. Intersite correlation coefficients for PM2 5 can be calculated based on the results
of FRM monitors placed at multiple sites within MS As across the United States. Pearson
correlation coefficients (r) calculated for pairs of monitoring sites in the Philadelphia, PA;
Cleveland, OH; Dallas, TX; and Los Angeles, CA MSAs are shown in Table 3-4. The 90th
percentile value, P90, of the absolute differences (in |ig/m3) between the two sites is shown in
parentheses below r along with the coefficient of divergence (COD), and the number of
3-40
-------
TABLE 3-4a-d. MEASURES OF THE SPATIAL VARIABILITY OF
PM2 5 CONCENTRATIONS WITHIN SELECTED METROPOLITAN
STATISTICAL AREAS
(a) Philadelphia,
PA
Sitel.D. # 340071007 340155001 420170012
340071007
340155001
420170012
420450002
420910013
421010004
421010136
Mean
Obs
SD
1 0.91 0.93
(6.3,0.14) (5.2,0.15)
170 167
1 0.84
(7.5,0.19)
176
1
Key
Airs Site I.D. #
Pearson r
(90th %-tile difference in concentration,
coefficient of divergence)
number of observations
14.78 14.59 14.11
197 208 217
9.2 8.65 8.47
420450002
0.87
(6.9,0.19)
183
0.88
(7.4,0.18)
194
0.85
(7.5,0.16)
199
1
15.92
230
8.65
420910013
0.88
(5.0,0.16)
176
0.83
(7.1,0.18)
184
0.88
(4.7,0.13)
200
0.87
(6.1,0.15)
208
1
14.2
221
8.93
421010004
0.94
(4.6,0.15)
163
0.89
(7.4,0.17)
169
0.94
(4.9,0.11)
177
0.94
(5.1,0.11)
187
0.90
(4.6,0.11)
181
1
15.72
610
9.18
421010136
0.93
(5.1,0.14)
166
0.85
(6.9,0.18)
173
0.89
(5.3,0.13)
180
0.88
(4.0,0.12)
193
0.87
(4.5,0.11)
185
0.96
(3.3, 0.08)
550
1
15.81
616
9.26
(b) Cleveland, OH
Sitel.D. # 390350013 390350038 390350060
390350013
390350038
390350060
390350065
390350066
390351002
390851001
390932003
Mean
Obs
SD
1 0.91 0.96
(7.1,0.13) (3.3,0.12)
320 322
1 0.92
(6.9,0.14)
306
1
Key
Airs Site I.D. #
Pearson r
(90th %-tile difference in concentration,
coefficient of divergence)
number of observations
18.34 20.16 18.39
368 931 353
9.59 11.5 9.9
390350065
0.94
(5.4,0.10)
314
0.89
(9.4,0.15)
296
0.93
(5.1,0.14)
309
1
17.47
340
8.95
390350066
0.92
(7.2,0.16)
300
0.85
(13.1,0.21)
290
0.90
(8.15,0.19)
300
0.96
(4.7,0.14)
295
1
14.74
332
8.45
390351002
0.88
(9.0,0.18)
308
0.84
(12.9,0.21)
304
0.87
(8.6, 0.20)
310
0.91
(5.4,0.16)
310
0.88
(4.9,0.16)
304
1
15
351
8.16
390851001
0.89
(10.7,0.21)
308
0.84
(14.3, 0.23)
294
0.88
(10.8,0.22)
307
0.90
(7.8, 0.20)
306
0.91
(5.8,0.15)
295
0.89
(6.0,0.18)
303
1
14
342
8.44
390932003
0.92
(8.1,0.17)
265
0.892
(11.2,0.18)
256
0.90
(8.9,0.18)
256
0.91
(7.7,0.18)
264
0.91
(5.8,0.13)
247
0.87
(5.3,0.18)
261
0.90
(6.6,0.15)
275
1
15.22
298
8.8
3-41
-------
TABLE 3-4a-d (cont'd). MEASURES OF THE SPATIAL VARIABILITY OF
PM2 5 CONCENTRATIONS WITHIN SELECTED METROPOLITAN
STATISTICAL AREAS.
(c) Dallas, TX
Site I.D. #
480850005
481130020
481130035
481130050
481130057
480850005
481130020
Mean
Obs
SD
(d) Los Angeles
Site I.D. #
60370002
60371103
60371201
60371301
60372005
60374002
Mean
Obs
SD
480850005 481130020 481130035 4
1 0.92 0.94
(3.5,0.11) (3.6,0.11) (
220 204
1 0.95
(3.2, 0.08) (
212
1
(
Key
Airs Site I.D. #
Pearson r
(90th %-tile difference in concentration,
coefficient of divergence)
number of observations
11.54 12.42 12.83
234 677 222
5.62 5.69 5.79
CA
60370002 60371103 60371201
1 0.87 0.76
(10.7,0.18) (14.6,0.23) (
581 208
1 0.86
(12.8, 0.20) (
205
1
(
Key
Airs Site I.D. #
Pearson r
(90th %-tile difference in concentration,
coefficient of divergence)
number of observations
20.91 22.48 18.32
641 656 217
13.35 13.47 11.65
81130050 481130057
0.94 0.89
4.3,0.13) (6.3,0.15)
213 195
0.94 0.92
3.3,0.09) (4.1,0.11)
603 205
0.97 0.93
2.0, 0.06) (3.9, 0.09)
203 191
1 0.94
(2.7, 0.08)
199
1
13.34 13.67
644 215
5.79 6.14
50371301 60372005
0.68 0.95
17.9,0.25) (6.2,0.14)
229 212
0.89 0.93
10.1,0.12) (7.1,0.11)
222 207
0.76 0.85
18.1,0.24) (12.1,0.18)
212 197
1 0.78
(13.2,0.18)
214
1
23.73 20.09
237 220
13.43 11.92
480850005
0.94
(3.7,0.10)
218
0.95
(2.5, 0.07)
635
0.97
(1.9,0.06)
207
0.98
(2.2, 0.06)
608
0.95
(3.1,0.08)
198
1
12.68
687
5.67
60374002
0.60
(18.1,0.26)
553
0.80
(13.6,0.17)
563
0.66
(18.2,0.24)
197
0.95
(8.1,0.11)
216
0.62
(15,0.20)
204
1
20.31
621
12.42
481130020
0.94
(3.1,0.10)
189
0.97
(2.3, 0.07)
207
0.94
(3.6,0.10)
185
0.94
(4.2,0.12)
196
0.91
(5.5,0.14)
182
0.96
(3.0, 0.09)
198
1
11.67
216
5.43
3-42
-------
observations used in the calculation of r, P90 and COD is given on the third line. The COD was
used by Wongphatarakul et al. (1998) as a measure of the degree of similarity between 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 and those for another 23 MSAs are given along with maps in
Appendix 3 A. As the concentrations of PM2 5 at two sampling sites become more alike, the
COD approaches zero; as the concentrations diverge, the COD approaches one.
The four MSAs shown in Table 3-4(a-d) were chosen to illustrate different patterns of
spatial variability across the United States. In addition, air-pollution health-outcome studies
have been performed in a few of these MSAs. It can be seen from inspection of Table 3-4 that
correlation coefficients vary over a wide range in the MSAs shown. Correlations between sites
in the Philadelphia, PA, Cleveland, OH, and Dallas, TX MSAs are all high and span a relatively
narrow range (0.82 to 0.97). However, correlations between sites in the Los Angeles-Long
Beach MSA are lower than in the three other MSAs and span a wider range of values (0.60 to
0.95). If the monitoring site in Lancaster, CA were included, correlations would be even lower.
This site was omitted because it did not meet completeness criteria for 2001. The extension of
these analyses to include the relevant consolidated MSAs (CMSAs) would also produce a
number of sites that are even less well correlated with each other, in part, because some included
sites are located outside of urban airsheds.
Correlation coefficients between pairs of sites in the other 23 MSAs given in Appendix 3 A
for the most part fall within the range of values given in Table 3-4. In four MSAs (Columbia,
SC; Norfolk, VA; Grand Rapids, MI; and Baton Rouge, LA), intersite correlations are all greater
than 0.9. In nine others (Cleveland, OH; Louisville, KY; Chicago, IL; Milwaukee, WI;
Philadelphia, PA; Detroit, MI; Kansas City, MO; Dallas, TX; and Salt Lake City, UT), they are
1 The COD for this purpose is defined as follows:
1
'* \p
p
I
7 = 1
f *ij
Uy
- x& "
+ ^'J
2
(3-1)
where Xy and X;k represent the 24-h average PM2 5 concentration for day i at site j and site k and p is the number
of observations.
3-43
-------
all greater than 0.8. Correlations between sites in the other MSAs examined tend to be lower
and span a broader range than for the MSAs mentioned above.
Seven pairs of collocated monitors in seven MSAs (Columbia, SC; Dallas, TX; Detroit,
MI; Grand Rapids, MI; Louisville, KY; Steubenville, OH; Washington, DC) provide an
indication of the performance of collocated monitors (see Table 3A-1). Mean values of r, P90,
and COD for these seven pairs of monitors are 0.986, 1.63 |ig/m3, and 0.060, thus suggesting that
most of the intersite variability seen is not due to sampler imprecision.
There are no strong regional patterns evident in the data given in Appendix 3 A except that
correlations tend to be higher between monitoring sites in MSAs in the southeastern United
States than between monitoring sites in other regions.
A number of factors affect intersite correlations within MSAs. These include field
measurement and laboratory analysis errors, placement of monitors close to active sources,
placement of monitors in outlying areas, placement of monitors in locations that are isolated
topographically from other monitors, placement of monitors in areas outside of local
atmospheric circulation regimes (e.g., land-sea breezes), and transient local events (e.g.,
thunderstorms or sporadic emissions). In several MSAs (i.e., Atlanta, GA, Seattle, WA, and
Los Angeles-Long Beach, CA) at least one site is remote from the others (by at least 100 km),
is physically separated from them by mountains, and is really neither part of the urban area nor
the urban airshed. Correlations between concentrations at these sites and others tend to be lower
than among the other sites, and concentration differences tend to be larger. It should be noted
that outlying sites such as these are included in many epidemiologic time-series studies without
any differential weighting (e.g., with respect to the exposed population or spatial differences in
susceptibility or with regard to compositional differences). Although it is frequently the case
that distance between sites in urban areas is largely responsible for the spatial variability that is
observed, there are a few instances for which correlations are higher and differences in
concentrations are lower for sites that are located farthest apart. This situation may arise because
these sites are influenced more by the regional background of secondary PM rather than by local
sources, but there are not any consistent distances evident below which correlations and
differences in concentrations tend towards some limiting values. Still, it is generally the case
that outlying sites are characterized by lower annual mean concentrations.
3-44
-------
Indications of land use (commercial, industrial, residential, agricultural, forest) and
location of sites (urban/city center, suburban, rural) are given in the AIRS data base. Categories
such as urban/city center can refer to very different conditions in Columbia, SC, and Chicago,
IL. Also, it should not be automatically assumed that concentrations measured at sites
categorized as industrial are dominated by local emissions. The PM2 5 monitoring sites are
generally deployed to capture potential population exposures in a variety of environments as
opposed to monitoring for compliance as it exists around local sources. It should be
remembered that much of PM2 5 is secondary in origin. The widespread formation of secondary
PM, coupled with the long atmospheric lifetime of PM2 5, ensures some measure of uniformity in
the correlations of PM25 across urban areas. Correlations between many site pairs classified as
industrial can be high, even though they are separated by large distances (as in the Seattle MSA).
Some indication of the variability of primary PM25 produced by local sources can be
obtained by examining the variability of CO, which is produced mainly by mobile sources (U.S.
Environmental Protection Agency, 2000b), and by the variability in EC concentrations (Kinney
et al., 2000). Carbon monoxide is relatively inert on the urban scale, and its distribution is
governed by the spatial pattern of its emissions and the subsequent dispersion of these emissions
not by photochemistry. Carbon monoxide concentrations are at least a factor of three higher
near urban centers than in surrounding rural areas within the four consolidated metropolitan
statistical areas examined in the EPA document, Air Quality Criteria for Carbon Monoxide (CO
AQCD; U.S. Environmental Protection Agency, 2000b). The correlations of CO within the
urban areas examined in that document were all low to moderate. Therefore, it might be
expected that primary PM2 5 produced by local traffic should be at least as heterogeneous as CO
in a given urban area. Elemental carbon is a significant component of diesel exhaust (cf,
Appendix 3D). Kinney et al. (2000) measured EC and PM2 5 concentrations at four sites located
on sidewalks of streets characterized by varying exposures to diesel emissions in upper
Manhattan (Harlem, NY). Whereas the mean PM2 5 concentrations varied by about one-third
from 37 to 47 |ig/m3 at the four sites, mean EC concentrations varied by a factor of four from
1.5 to 6.2 |ig/m3. The corresponding ratios of EC to PM25 ranged from 0.039 to 0.14. Although
EC constituted a relatively small fraction of PM25 in this study, spatial variability in its sources
(diesel and gasoline fueled vehicles, resuspended road dust, and cooking) contributed, on
average, about one-third of the spatial variability observed in PM25 concentrations. Further
3-45
-------
analyses are needed to determine whether the remaining variability could be attributed to other
local and city-wide sources. Because the effects of emissions from local point sources on
receptor sites depend strongly on wind direction, correlations involving contributions from local
sources can be much lower than from area sources (much as motor vehicle traffic) or from
regionally dispersed sources (such as the photochemical production of secondary organic PM
and sulfate).
The difference in mean PM25 concentrations between the site with the lowest and the site
with the highest mean concentration range in all MSAs included in Appendix 3A ranges from
0.4 |ig/m3 (Baton Rouge) to about 8 |ig/m3 (Pittsburgh). Six MSAs (Chicago, Seattle, Cleveland,
St. Louis, Detroit, and Pittsburgh) show maximum intersite differences in the annual mean larger
than 6 |ig/m3. In the Seattle MSA, there is one monitoring site (Figure 3A-23a) that is separated
from the remaining sites by topography and has much lower mean PM2 5 concentrations, much
smaller seasonal variability in concentrations, and much lower maximum concentrations than
these other sites. However, the annual mean concentrations at all the other sites within the
Seattle MSA are within 3 |ig/m3 of each other. Differences in annual mean concentrations are
also larger between sites located in different MSAs but within the same CMSA. For example,
in the consolidated MSA of Los Angeles-Riverside, the range of annual mean PM2 5 values is
extended from -20 |ig/m3 in the urban area of Los Angeles county to -29 |ig/m3 in Riverside
County. Large differences in annual mean concentrations within a given area reflect differences
in source or meteorological or unique topographic characteristics affecting sites, whereas very
small differences found in some areas may mainly be the result of measurement imprecision.
While high correlations of PM25 provide an indication of the spatial uniformity in temporal
variability (directions of changes) in PM2 5 concentrations across urban areas, they do not imply
uniformity in the PM25 concentrations themselves. The 90th percentile difference in
concentrations (P90) and the coefficient of divergence (COD) are used here to give a more
quantitative indication of the degree of spatial uniformity in PM25 concentrations across urban
areas. A COD of zero implies that both data sets are identical, and a COD of one indicates
that two data sets are completely different. The calculation of the Pearson correlation coefficient
(r), P90, and COD allows for distinctions between pairs of sites to be made based on various
combinations of these parameters. Figure 3-18 shows examples of the varying degree of
heterogeneity in concentrations between pairs of sites that are highly correlated (r > 0.9 for all
3-46
-------
Columbia, SC 1999 & 2000
8 12°1
o
c 100-
^
= 80-
o
o
0 60-
"5
fc 40-
£ 20-
3
z o-
I
'
0' 30-i
o
o
I 25-
3 20-
u
o
0 15-
1 1°~
1 5"
n
1
\
O>' T,
co 14 -,
N%> ^
Chicago, IL2000
17-031-2001 vs. 17-031-4201
1
1
r = 0.94
COD = 0.14
P90 = 5.5 |jg/m3
l.ll .
> j> j ^\^VVV\^VN
Detroit, Ml 2000
26-099-0009 vs. 26-163-0033
r = 0.93
COD = 0.22
P90= 12.7|jg/m3
1
II.
nil. ,i . i
&
&
N"^ N" \ N^ N" ^
Concentration Difference (|jg/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 I.D. 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.
3-47
-------
three site pairs). The increase in the spread of concentrations between the chosen site-pairs is
reflected in increases in both P90 and COD. Pairs of sites showing high correlations and
CODs < 0.1 and P90s < 4 |ig/m3 (as in Columbia, SC, Figure 3-7a) indicate homogeneity in
both PM2 5 concentrations and in their temporal variations. Presumably, sites such as these are
more strongly affected by regional than local sources. Pairs of sites showing low correlations,
values of P90 > 10 |ig/m3 and CODs > 0.2, as in Los Angeles, CA (Table 3-5), indicate
heterogeneity in both PM2 5 concentrations and in their temporal variations. Note that the
extended urban area or the CMSA includes Riverside County, as well as Los Angeles County.
Even lower correlations and a greater degree of heterogeneity in PM25 concentrations were
found in the extended CMSA. Pairs of sites showing high correlations (r > 0.9) and CODs > 0.2
and P90s > 10 |ig/m3 indicate heterogeneity in concentrations but homogeneity in their day-to-day
changes. Selected pairs of sites in the Cleveland MSA show moderate to high correlations
coupled with CODs > 0.2 and P90s > 10 |ig/m3 (Table 3-4), suggesting moderate homogeneity in
day-to-day changes but significant spatial heterogeneity in concentrations.
The effect of local point sources on intersite variability can be seen at several sites among
those listed in Table 3-4(a-d) and Appendix 3A. Sites 39-035-0038 (Cleveland, OH; Table 3-4
and Figure 3A-8), 18-089-0022 (Gary, IN; Figure 3A-15), 55-079-0043 (Milwaukee, WI;
Figure 3A-13), and 17-119-0023 (St. Louis, MO; Figure 3A-17) are designated as "source
oriented" in the AIRS data base in contrast to the "population exposure" objective associated
with most of the MSA sites. PM25 concentrations at these sites are weakly correlated with other
sites within the MSA as evidenced by low correlation coefficients and large P90s and CODs even
though some of the neighboring sites may be located short distances away. Other sites
designated as "source oriented" in Chicago, Milwaukee, and St. Louis do not show clear
evidence that local sources are contributing to intersite variability. Conversely, in the Tampa,
FL MSA, pairs of sites are only moderately correlated (0.7 < r < 0.87), but the distribution of
concentrations is rather homogeneous (COD < 0.14 and P90 < 5 |ig/m3; Figure 3 A-7). Thus, a
number of different combinations of spatial uniformity in PM2 5 concentrations, and correlations
of these concentrations are found.
Values of P90 for absolute differences in concentrations between sites span a wide range in
the data set given in Appendix 3 A. In many instances they can be quite low, only about a
few |ig/m3. Such cases are found mainly in the eastern United States. The largest P90 values
3-48
-------
TABLE 3-5. MEASURES OF THE SPATIAL VARIABILITY OF
PM102 5 CONCENTRATIONS WITHIN SELECTED METROPOLITAN
STATISTICAL AREAS
(a) Cleveland, OH
Site 390350013 390350038
390350013 1 0.67
(23.2, 0.26)
182
390350038 1
390350045
0.67
(28.5, 0.28)
95
0.65
(16.1,
0.22)
90
390350045
Key
390350060 Airs site j D #
Pearson r
(90th %-tile difference in concentration
390350065 coefficient of divergence)
number of observations
390851001
MEAN 26.36 18.63
Obs 216 614
SD 17.38 11.6
;
L
16.76
112
8.96
390350060
0.73
(17.9,0.22)
97
0.73
(11.7,0.18)
93
0.66
(18.1,0.23)
94
1
21.35
113
16.39
390350065 390851001
0.62 0.41
(27.0,0.31) (40.0,0.60)
98 94
0.69 0.44
(13.9, 0.62) (24.9, 0.53)
90 89
0.71 0.49
(10.6,0.31) (19.9,0.50)
102 99
0.74 0.31
(15.4,0.38) (28.0,0.59)
93 94
1 0.22
(20.4, 0.55)
99
1
16.79 7.15
111 109
9.49 4.94
(b) Dallas, TX
Site 481130020 481130035
481130020 1 0.79
(4.5, 0.17)
54
481130035 1
481130050
0.71
(9 3
0 99s!
\^.^, .j.~~j
55
0.69
(7.8,0.18)
50
481130050
481130057
MEAN 11.22 12.86
Obs 60 55
SD 5.35 6.66
-
L
14.46
56
6.44
481130057
0.66
(16 5 0 32)
54
0.60
(13.2,0.30)
50
0.69
(13.5,0.24)
50
1
19.12
55
10.55
Key
Airs Site I.D. #
Pearson r
(90th %-tile difference in
concentration,
coefficient of divergence)
number of observations
(c) Los Angeles, CA
Site 60370002 60371002
60370002 1 0.82
(19.0, 0.24)
49
60371002 1
60371103
0.63
(15.5,
0.18)
49
0.74
(11.5,0.21)
49
60371103
60374002
MEAN 24.1 15.33
Obs 56 56
SD 11.67011257 6.68
:
L
21.44
57
8.65
60374002
0.58
(17.3,0.27)
45
0.54
(11.5,0.25)
47
0.57
(12.5,0.22)
45
1
16.08
53
6.61
Key
Airs Site I.D. #
Pearson r
(90th %-tile difference in
concentration,
coefficient of divergence)
number of observations
3-49
-------
were associated with a single site in Pittsburgh and reached as high as 21 |ig/m3 (Figure 3A-9).
Excluding this site, large P90 values are found mainly in the western United States. Values of P90
> 18 |ig/m3 are found in the Riverside and Los-Angeles-Long Beach MS As. Maximum
differences in concentrations between sites can be much larger than shown in Figure 3-18 and
have been larger than 100 |ig/m3 on several occasions in the Atlanta, GA and Los Angeles-Long
Beach, CA MSAs. Rizzo and Pinto (2001) and Fitz-Simons et al. (2000) examined correlations
between sites located even farther apart than those examined here based on the 1999 AIRS data
set for PM2 5. They found that in a number of MSAs, PM25 concentrations are still well
correlated (r > 0.7) up to distances of 100 km or more. Leaderer et al. (1999) found r = 0.49
between sites outside of homes and a regional background monitor located from 1 to 175 km
away in southwestern Virginia. PM2 5 tends to be correlated over much larger areas in the East
than in the West, mainly because the terrain tends to be flatter over wider areas in the East
(Rizzo and Pinto, 2001). As a result, there is a greater opportunity for mixing of emissions
among dispersed source regions. Many large urban areas in the West are surrounded by
mountains. The presence of more rugged terrain in the West leads to greater confinement of
emissions from large urban areas. Other factors such as differences in the composition and
amount of emissions of precursors and in the rates of photochemical oxidation of these
emissions in the atmosphere also play a role.
There is also evidence for inter-annual variability in the spatial variability in PM2 5
concentrations. The median year-to-year changes in intersite r (0.03), P90 (-0.75 |ig/m3), and
COD (-0.015) from 1999 to 2000 do not differ significantly from zero for all the site pairs
considered in Appendix 3 A. The year-to-year changes in the spatial variability of PM25
concentrations in a number of MSAs (e.g., the Columbia, SC; Grand Rapids, MI; Milwaukee,
WI; Baton Rouge, LA; Kansas City, MO; Boise, ID; and Portland, OR MSAs) are similar and
are smaller than those found in the Cleveland, OH; Salt Lake City, UT; and San Diego, CA
MSAs. The ranges in these parameters are largest for a number of individual site-pairs,
especially those involving sites that are remote from the others in their MSAs. In these MSAs
(e.g., Atlanta, GA; Los Angeles, CA; and Seattle, WA MSAs), some sites may be located in
different airsheds from the remaining sites. Year-to-year changes in parameters describing
spatial variability in PM2 5 concentrations tend to be larger when sites in different counties within
a given MSA are considered rather than when sites in the same county are considered. There are
3-50
-------
a number of factors that can account for inter-annual variability in these parameters, such as
changes in patterns in the emissions of primary PM25; in the transport and rates of
transformation of secondary PM25 precursors in field measurement; and analysis procedures.
Some additional data for indicating the stability with respect to year-to-year changes in
spatial variability are available from earlier studies. For example, a comparison between data
obtained during the summers of 1992 and 1993 (Wilson and Suh, 1997), shown in Figure 3-19,
and data obtained during the summer of 1994 (Pinto et al., 1995) (cf, Table 3-8) in Philadelphia,
PA, suggests that inter-site correlations of PM25 have remained high and have changed very little
between the two study periods.
1.0
0.9
0.8
£ 0.7
•§ 0.6
SEE
0)
O n e
o
04
o 0.3
O
0.2
0.1
Correlations of PM Exposure Indicators
Philidelphia, Summer, 1992-93, 7 Sites
o.o
A
^
v/ \/
y v
PM
10
10-2.5
+ PM
*PM
Average r
2.5
10
0.90
0.86
APM10-2.5 °-38
4 8 12 16 20 24
Distance Between Sites (km)
0 Not Significant, all other correlations significant (P < 0.05)
28
32
Figure 3-19. Intersite correlation coefficients for PM25, PM10, and PM10_25.
Source: Wilson and Suh (1997).
3-51
-------
PM10.2.5
Intersite correlations of PM10_25 concentrations obtained during the summers of 1992 and
1993 in Philadelphia, PA (Wilson and Suh, 1997) are shown in Figure 3-19. As can be seen,
correlations of PM10_25 are substantially lower than those for PM25.
Intersite correlation coefficients can also be calculated for PM10_2 5 based on the AIRS data
set as shown in Table 3-5 for the Cleveland, OH, Dallas, TX, and Los Angeles, CA MS As.
However, data for analyzing the spatial variability of PM10_25 are more limited than for PM2 5;
therefore, fewer urban areas could be characterized in Appendix 3 A (Figures 3 A-28 to 3 A-44).
Whereas PM2 5 concentrations were found to be highly correlated between sites in the Detroit,
MI MSA (Table 3-4), estimated PM10_25 concentrations are noticeably less well correlated.
Likewise, correlations of PM10_2 5 in the Chicago, IL MSA are also lower than those for PM2 5.
In contrast, correlations of PM10_25 concentrations between several pairs of sites in the
Los Angeles-Long Beach partial MSA are higher than those for PM2 5.
The interpretation of these results is not straightforward, as concentrations of PM10_25 are
generated by taking the difference between collocated PM2 5 and PM10 monitors. Consequently,
caution must be exercised when viewing them. Errors in the measurement of PM2 5 and PM10
may play a large role in reducing apparent correlations of PM10_25 such that collocated
PM10_2 5"measurements" may be expected to be poorly correlated (White, 1998). Indeed, several
estimates of concentrations are negative. Negative PM10_25 concentrations also lead to artifacts
in the calculation of CODs. In cases where these artifacts cause a division by zero or a very
small number in the calculation of CODs, dashes are used in Figures 3 A-28 through 3 A-44.
These results imply that negative concentrations can be almost equal in absolute magnitude to
positive concentrations in the same MSA. The possible causes of these errors are essentially the
same as those discussed in Section 3.2.1 with regard to the occurrence of PM25 to PM10 ratios
greater than one. There are also physical bases for expecting that PM10_2 5 concentrations may be
more variable than those for PM2 5. PM10_2 5 is mainly primary in origin, and its emissions are
spatially and temporally heterogenous. Similar considerations apply to primary PM2 5, but much
of PM25 is secondary, and sources of secondary PM are much less spatially and temporally
variable. Dry deposition rates of particles depend strongly on particle size. Whereas all particles
may be brought to the surface by turbulent motions in the atmosphere, gravitational settling
becomes more important with increasing particle size. Gravitational settling can effectively
3-52
-------
limit the horizontal distance a particle can travel. For example, 10 jim Da particles suspended in
a hypothetical 1 km deep planetary boundary layer can be removed within a few hours, but 1 |im
Da particles can remain suspended in the atmosphere for up to 100 to 1,000 times longer before
being dry deposited. (Estimated atmospheric lifetimes were based on deposition velocities given
in Lin et al. [1994] for typical wind speeds.) The findings of larger correlations of PM10_25
between several site pairs in the Los Angeles basin (cf, Figures 3 A-25/26 and Figures 3 A-42/43)
and one other site pair in the St. Louis, MO-IL MSA (cf, Figures 3 A-17 and 3A-37) are
anomalous in light of the discussion above. However, these findings could have resulted from
differences between the spatial and temporal behavior of sources of PM2 5 and PM10_2 5 in these
locations. Because of negative values, CODs were not calculated.
PM Components
Three methods for comparing the chemical composition of aerosol databases obtained at
different locations and times were discussed by Wongphatarakul et al. (1998). Log-log plots of
chemical concentrations obtained at pairs of sampling sites accompanied by the coefficient of
divergence (COD) were examined as a way to provide an easily visualized means of comparing
two data sets2. Examples comparing downtown Los Angeles with Burbank and with
Riverside-Rubidoux are shown in Figures 3-20 and 3-21. As the composition of two sampling
sites become more similar, the COD approaches zero; as their compositions diverge, the COD
approaches one. Correlation coefficients calculated between components can be used to show
the degree of similarity between pairs of sampling sites.
In addition to calculating correlation coefficients for total mass or for individual
components, correlation coefficients for characterizing the spatial variation of the contributions
from given source types can also be calculated by averaging the correlation coefficients of the
set of chemical components that represent the source type. Correlation coefficients showing the
2 The COD for two sampling sites is defined as follows:
(3-2)
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.
3-53
-------
102'
co
E 10° •
O)
c
re
.a
3 10-1-
CO
10-2-
io-3-
COD = 0.099
TOT>
io-3
SO,
2-
K
1Q-2
Downtown Los Angeles (ng/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).
3-54
-------
o>
x
3
O
P
!5
3
oc
10-d
1Q-
Downtown Los Angeles (jig/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).
3-55
-------
spatial relationships among PM2 5 (total) and contributions from different source categories
obtained 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. Also,
in that study, 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); and 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 SO42 represent secondary PM components. The averages 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.
Correlation coefficients in Philadelphia air based on data obtained at four sites for PM2 5
(total), crustal components (Al, Si, Ca, and Fe), the major secondary component (sulfate),
organic carbon (OC), and elemental carbon (EC) are shown in Table 3-7. Because these data
were obtained after Pb had been phased out of gasoline, a motor vehicle contribution could not
be estimated from the data. Pb also is emitted by discrete point sources, such as the Franklin
smelter. Concentrations of V and Ni were often beneath detection limits; so, the spatial
variability in PM due to residual oil combustion was not estimated. Sulfate in aerosol samples
collected in Philadelphia arises mainly from long-range transport from regionally dispersed
sources (Dzubay et al., 1988). This conclusion is strengthened by the high correlations in sulfate
between different monitoring sites and the uniformity in sulfate concentrations observed among
the sites. Widespread area sources (e.g., motor vehicle traffic) also may emit pollutants that are
correlated between sites provided that traffic patterns and emissions are similar throughout the
area under consideration.
Landis et al. (2001) found relatively high correlations between PM2 5 (r = 0.97), sulfate
(r = 0.99), OC (r = 0.97), EC (r = 0.83), NaCl (r = 0.83), and nitrate (r = 0.83) measured at two
sites located several km apart in the Baltimore, MD area. Concentrations of crustal material
(r = 0.63) and the sum of total metal oxides (r = 0.76) were not as well correlated. These results
are consistent with those for another eastern city, Philadelphia, PA, given in Table 3-7. The
results presented above for Philadelphia, PA; Baltimore, MD; and Los Angeles, CA, indicate
3-56
-------
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)
Sampling Site
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
rtotal
-0.027
0.051
0.066
0.095
0.12
0.76
0.852
0.77
0.827
0.808
0.704
0.731
0.88
0.842
0.928
I"
crustal
0.034
0.075
0.105
0.143
0.568
0.599
0.633
0.649
0.653
0.825
sec
0.768
0.888
0.749
0.804
0.854
0.79
0.737
0.909
0.817
0.96
rmv
0.492
0.504
0.579
0.556
0.669
0.688
0.714
0.861
0.719
0.871
residual oil
0.17
0.15
0.161
0.233
0.533
0.491
0.295
0.482
0.378
0.606
Source: Wongphatarakuletal. (1998).
TABLE 3-7. CORRELATION COEFFICIENTS FOR SPATIAL VARIATION OF
PM25 MASS AND DIFFERENT COMPONENTS FOR PAIRS OF SAMPLING
SITES IN PHILADELPHIA (1994)
Sampling Site
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.
r,o,
0.92
0.93
0.93
0.98
0.95
0.95
crustal
0.52
0.47
0.57
0.67
0.9
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
rEC
0.84
0.77
0.89
0.82
0.79
0.63
rpb
0.43
-0.07
0.11
0.2
0.47
0.11
Source: Pinto etal. (1995).
3-57
-------
that secondary PM components are more highly correlated than primary components and may be
more highly correlated than total PM2 5. These results suggest that the correlation of PM
concentrations across an urban area may depend on the relative proportions of primary and
secondary components of PM at individual sites. Sampling artifacts affecting the measurement
of nitrate and organic carbon can obscure these relations and may depress correlations between
sites.
Kao and Friedlander (1995) examined the statistical properties of a number of PM
components in the South Coast Air Basin (Los Angeles area). They found that, regardless of
source type and location within their study area, the concentrations of nonreactive, primary
components of PM10 had approximately log-normal frequency distributions with constant values
of the geometric standard deviations (GSDs). However, aerosol constituents of secondary origin
(e.g., SO42 , NH4+, and NO3 ) were found to have much higher GSDs. Surprisingly, the GSDs of
organic (1.87) and elemental (1.74) carbon were both found to be within 1 SD (0.14) of the mean
GSD (1.85) for nonreactive primary species, compared to GSDs of 2.1 for sulfate, 3.5 for nitrate,
and 2.6 for ammonium. These results suggest that most of the OC seen in ambient samples in
the South Coast Air Basin was of primary origin. Pinto et al. (1995) found similar results for
data obtained during the summer of 1994 in Philadelphia. Further studies are needed to
determine if these relations are valid at other locations and to what extent the results might be
influenced by sampling artifacts such as the evaporation of volatile constituents during or after
sampling.
The use of correlations between OC and EC and between OC to EC ratios based on a
comparison between values measured in source emissions and ambient observations has also
been suggested as a means to distinguish between secondary and primary sources of OC (Turpin
and Huntzicker, 1995; Strader et al., 1999). Ratios of OC to EC from combustion sources are
typically < 3 and may even be < 1 in diesel emissions (cf, Appendix 3D). Cabada et al. (2002)
concluded that secondary organic PM can contribute from 10 to 35% of total organic PM on an
annual basis, with values > 50% during the summer and 0% during the winter months in
Pittsburgh, PA based on chemistry-transport model (CTM) results and comparison with
emissions inventory values of OC to EC ratios. All of these inferences are subject to
considerable uncertainty in the methods for measuring OC, as discussed in Chapter 2.
Ambiguity also arises in the ratio method, as the ratios may change due to chemical reactions
3-58
-------
occurring during aging of the particles. However, little work has been done on this issue. The
ratio can be greater than nine in emissions from wildfires, and transport from distant fires can
skew results unless this transport is taken into account (cf, Appendix 3D). Modeling studies
that rely on OC to EC ratios in emissions inventories to predict the amount of secondary OC may
be biased towards higher ratios of secondary OC because emissions of primary biologic particles
are not included in the inventories. Additional concerns arise from uncertainties in the
mechanism of formation of secondary OC from gaseous biogenic and anthropogenic precursor
emissions and the uncertainty in those emissions (Section 3.3.1). It is clear, however, that
secondary organic PM is being formed in the atmosphere (Blando et al., 1998 and Appendix 3C).
Few studies have compared aerosol composition in urban areas to that in nearby rural
areas. One exception is Tanner and Parkhurst (2000), which found that sulfate constituted a
larger fraction of fine particle mass at rural sites in the Tennessee Valley PM25 monitoring
network than did OC. For urban sites, the situation was largely reversed: OC constituted a
larger fraction of aerosol mass than sulfate. Future systematic comparisons of urban-rural
differences in aerosol properties should be facilitated with implementation of the national
speciation network and continued operation of the IMPROVE network.
3.3 SOURCES OF PRIMARY AND SECONDARY PM
Information about the nature and relative importance of sources of ambient PM is
presented in this section. Table 3-8 summarizes anthropogenic and natural sources for the major
primary and secondary aerosol constituents of fine and coarse particles. Anthropogenic sources
can be further divided into stationary and mobile sources. Stationary sources include those such
as: fuel combustion for electrical utilities, residential space heating, and industrial processes;
construction and demolition; metals, minerals, and petrochemicals; wood products processing;
mills and elevators used in agriculture; erosion from tilled lands; waste disposal and recycling.
Mobile or transportation-related sources include direct emissions of primary PM and secondary
PM precursors from highway vehicles and non-road sources as well as fugitive dust from paved
and unpaved roads. In addition to fossil fuel combustion, biomass in the form of wood is burned
for fuel. Vegetation is burned to clear new land for agriculture and for building construction, to
dispose of agricultural and domestic waste, to control the growth of animal or plant pests, and
3-59
-------
TABLE 3-8. CONSTITUENTS OF ATMOSPHERIC PARTICLES AND THEIR MAJOR SOURCES1
Sources
Primary (PM < 2.5 urn) Primary (PM > 2.5 urn)
Aerosol
species Natural Anthropogenic Natural Anthropogenic
Sulfate Sea spray Fossil fuel combustion Sea spray —
(so42-)
Nitrate — — — —
(NOf)
Secondary PM Precursors
Natural
Oxidation of reduced sulfur
gases emitted by the oceans and
wetlands and SO2 and H2S
emitted by volcanism and forest
fires
Oxidation of NOX produced by
soils, forest fires, and lighting
(PM < 2.5 urn)
Anthropogenic
Oxidation of SO2
emitted from fossil
fuel combustion
Oxidation of NOX
emitted from fossil
Minerals
fuel combustion and
in motor vehicle
exhaust
Erosion and
reentrainment
Fugitive dust from
paved and unpaved
roads, agriculture,
forestry, construction,
and demolition
Erosion and
reentrainment
Fugitive dust, paved
and unpaved road dust,
agriculture, forestry,
construction, and
demolition
Ammonium —
(NH4+)
Organic Wildfires
carbon (OC)
Elemental Wildfires
carbon
(EC)
Metals Volcanic
activity
Bioaerosols Viruses and
bacteria
Prescribed burning,
wood burning, motor
vehicle exhaust, and
cooking
Motor vehicle exhaust,
wood burning, and
cooking
Fossil fuel combustion,
smelting, and brake
wear
—
Soil humic matter
Erosion, reentrainment,
and organic debris
Plant and insect
fragments, pollen, fungal
spores, and bacterial
agglomerates
— Emissions of NH3 from wild Emissions of NH3
animals, and undisturbed soil from animal
husbandry, sewage,
and fertilized land
Tire and asphalt wear Oxidation of hydrocarbons Oxidation of
and paved road dust emitted by vegetation (terpenes, hydrocarbons emitted
waxes) and wild fires by motor vehicles,
prescribed burning,
and wood burning
Tire and asphalt wear — —
and paved road dust
— — —
— — —
'Dash (-) indicates either very minor source or no known source of component.
-------
to manage forest resources (prescribed burning). Also shown are sources for precursor gases,
the oxidation of which forms secondary PM. The atmospheric chemical processes producing
secondary PM are described in Section 3.3.1.
In general, the sources of fine PM are very different from those for coarse PM. Some of
the mass in the fine size fraction forms during combustion from material that has volatilized in
combustion chambers and then recondensed before emission into the atmosphere. Some
ambient PM25 forms in the atmosphere from photochemical reactions involving precursor gases.
That PM formed by the first mechanism is referred to as primary, and the PM formed by the
second mechanism is referred to as secondary. PM10_2 5 is mainly primary in origin, as it is
produced by the abrasion of surfaces or by the suspension of biological material. Because
precursor gases undergo mixing during transport from their sources, it is difficult to identify
individual sources of secondary PM constituents. Transport and transformations of precursors
can occur over distances of hundreds of kilometers. The coarse PM constituents have shorter
lifetimes in the atmosphere; so their effects tend to be more localized. Only major sources for
each constituent within each broad category shown at the top of Table 3-8 are listed. Not all
sources are equal in magnitude. Chemical characterizations of primary particulate emissions for
a wide variety of natural and anthropogenic sources (as shown in Table 3-8) were given in
Chapter 5 of the 1996 PM AQCD. Summary tables of the composition of source emissions
presented in the 1996 PM AQCD and updates to that information are provided in Appendix 3D.
The profiles of source composition were based in large measure on the results of various studies
that collected signatures for use in source apportionment studies.
Natural sources of primary PM include windblown dust from undisturbed land, sea spray,
and plant and insect debris. The oxidation of a fraction of terpenes emitted by vegetation and
reduced sulfur species from anaerobic environments leads to secondary PM formation.
Ammonium (NH4+) ions, which play a major role in regulating the pH of particles, are derived
from emissions of ammonia (NH3) gas. Source categories for NH3 have been divided into
emissions from undisturbed soils (natural) and emissions that are related to human activities
(e.g., fertilized lands, domestic and farm animal waste). There is ongoing debate about
characterizing emissions from wildfires (i.e., unwanted fire) as either natural or anthropogenic.
Wildfires have been listed in Table 3-8 as natural in origin, but land management practices and
other human actions affect the occurrence and scope of wildfires. For example, fire suppression
3-61
-------
practices allow the buildup of fire fuels and increase the susceptibility of forests to more severe
and infrequent fires from whatever cause, including lightning strikes. Similarly, prescribed
burning is listed as anthropogenic, but can be viewed as a substitute for wildfires that would
otherwise occur eventually on the same land.
The transformations that gaseous precursors to secondary PM undergo after being emitted
from the sources shown in Table 3-8 are described in Section 3.3.1. Aspects of the transport of
primary PM and secondary PM, including the transport of material from outside the United
States, are described in Section 3.3.2. A brief introduction to the deposition of particles is also
given in Section 3.3.2, and a more detailed discussion of deposition processes is presented in
Chapter 4. Methods to infer contributions from different source categories to ambient PM using
receptor models and the results of these modeling efforts are given in Section 3.3.3. Estimates of
emissions of primary PM and precursors to secondary PM from major sources are presented in
Section 3.3.4. A discussion of the uncertainties associated with these emissions is given in
Section 3.3.5.
3.3.1 Chemistry of Secondary PM Formation
Precursors to secondary PM have natural and anthropogenic sources, just as primary PM
has natural and anthropogenic sources. The major atmospheric chemical transformations leading
to the formation of particulate nitrate and sulfate are relatively well understood; whereas those
involving the formation of secondary aerosol OC are less so and are still subject to much current
investigation. A large number of organic precursors are involved and many of the kinetic details
still need to be determined. Also, many of the actual products of the oxidation of hydrocarbons
have yet to be identified.
Formation ofSulfates and Nitrates
A substantial fraction of the fine particle mass, especially during the warmer months of the
year, is secondary sulfate and nitrate formed as the result of atmospheric reactions. Such
reactions involve the gas phase conversion of SO2 to H2SO4 (which forms liquid particles)
initiated by reaction with OH radicals and aqueous-phase reactions of SO2 with H2O2, O3, or O2
(catalyzed by Fe and Mn). These heterogeneous reactions may occur in cloud and fog droplets
or in films on atmospheric particles. NO2 can be converted to gaseous HNO3 by reaction with
3-62
-------
OH radicals during the day. At night, NO2 also is oxidized to nitric acid by a sequence of
reactions initiated by O3 that produce nitrate radicals (NO3) and dinitrogenpentoxide (N2O5) as
intermediates. Both H2SO4 and HNO3 react with atmospheric ammonia (NH3). Gaseous NH3
reacts with gaseous HNO3 to form particulate NH4NO3. Gaseous NH3 reacts with H2SO4 to form
acidic HSO4 (in NH4 HSO4) as well as SO42 in (NH4)2SO4. In addition, acid gases such as SO2
and HNO3 may react with coarse alkaline particles to form coarse secondary PM containing
sulfate and nitrate. Examples include reactions with basic compounds resulting in neutralization
(e.g., CaCO3 + 2HNO3 - Ca(NO3)2 + H2CO3t) or with salts of volatile acids resulting in release
of the volatile acid (e.g., SO2 + 2NaCl + H2O - Na2SO3 + 2HC1T).
If particulate NH4NO3 coagulates with an acidic sulfate particle (H2SO4 or HSO4 ),
gaseous HNO3 will be released, and the NH3 will increase the neutralization of the acidic sulfate.
Thus, in the eastern United States, where PM tends to be acidic, sulfate usually constitutes a
larger fraction of PM mass than nitrate. However, in the western United States, where
higher NH3 and lower SO2 emissions permit complete neutralization of H2SO4, the concentration
of nitrate could be higher than that of sulfate as it is in areas such as the Los Angeles Basin and
the San Joaquin Valley. As SO2 concentrations in the atmosphere in the eastern United States
are reduced, the NH3 left in the atmosphere after neutralization of H2SO4 will be able to react
with HNO3 to form NH4NO3. Therefore, a reduction in SO2 emissions, especially without a
reduction in NOX emissions, could lead to an increase in NH4NO3 concentrations (West et al.,
1999; Ansari and Pandis, 1998). Thus, the sources, trends, and possible environmental effects
of NH4NO3 are of interest for both the western and eastern United States.
Chemical reactions of SO2 and NOX within plumes are an important source of H+, SO42,
and NO3. These conversions can occur by gas-phase and aqueous-phase mechanisms.
In power-plant or smelter plumes containing SO2 and NOX, the gas-phase chemistry depends on
plume dilution, sunlight, and VOCs either in the plume or in the ambient air mixing into and
diluting the plume. For the conversion of SO2 to H2SO4 in the gas-phase in such plumes during
summer midday conditions in the eastern United States, the rate typically varies between 1 and
3%/h, but in the cleaner western United States the rate rarely exceeds 1%/h. For the conversion
of NOX to HNO3, the gas-phase rates appear to be approximately three times faster than the SO2
conversion rates. Winter rates for SO2 conversion are approximately an order of magnitude
lower than summer rates.
3-63
-------
The contribution of aqueous-phase chemistry to particle formation in point-source plumes
is highly variable, depending on the availability of the aqueous phase (wetted aerosols, clouds,
fog, and light rain) and the photochemically generated gas-phase oxidizing agents, especially
H2O2 for SO2 chemistry. The in-cloud conversion rates of SO2 to SO42 can be several times
larger than the gas-phase rates given above. Overall, it appears that SO2 oxidation rates to SO42
by gas-phase and aqueous-phase mechanisms may be comparable in summer, but aqueous-phase
chemistry may dominate in winter. Further details concerning the chemistry of SO2 and NOX in
power plant plumes can be found in Hewitt (2001).
In the western United States, markedly higher SO2 conversion rates have been reported in
smelter plumes than in power plant plumes. The conversion occurs predominantly by a gas-
phase mechanism. This result is attributed to the lower NOX in smelter plumes. In power plant
plumes, NO2 depletes OH radicals and competes with SO2 for OH radicals.
In urban plumes, the upper limit for the gas-phase SO2 conversion rate appears to be about
5%/h under the more polluted conditions. For NO2, the rates appear to be approximately three
times faster than the SO2 conversion rates. Conversion rates of SO2 and NOX in background air
are comparable to the peak rates in diluted plumes. Neutralization of H2SO4 formed by SO2
conversion increases with plume age and background NH3 concentration. If the NH3 levels are
more than sufficient to neutralize H2SO4 to (NH4)2SO4, the HNO3 formed from NOX conversions
may be converted to NH4NO3.
Formation of Secondary Organic Particulate Matter (SOPM)
Atmospheric reactions involving VOCs such as alkanes, alkenes, aromatics, cyclic olefms,
and terpenes (or any reactive organic gas that contains at least seven carbon atoms) yield organic
compounds with low-saturation vapor pressures at ambient temperature. Such reactions may
occur in the gas phase, in fog or cloud droplets (Graedel and Goldberg, 1983; Faust, 1994), or
possibly in aqueous aerosols (Aumont et al., 2000). Reaction products from the oxidation of
reactive organic gases also may nucleate to form new particles or condense on existing particles
to form SOPM. Organic compounds with two double bonds or cyclic olefms may react to form
dicarboxylic acids, which, with four or more carbon atoms, also may condense. Both biogenic
and anthropogenic sources contribute to primary and secondary organic PM (Grosjean, 1992;
3-64
-------
Hildemann et al., 1996; Mazurek et al., 1997; Schauer et al., 1996). Oxalic acid was the most
abundant organic acid found in PM2 5 in California (Poore, 2000).
Although the mechanisms and pathways for forming inorganic secondary PM are fairly
well known, those for forming SOPM are not as well understood. Ozone and the OH radicals
are thought to be the major initiating reactants. However, HO2 and NO3 radicals also may
initiate reactions; and organic radicals may be nitrated by HNO2, HNO3, or NO2. Pun et al.
(2000) discuss formation mechanisms for highly oxidized, multifunctional organic compounds.
The production of such species has been included in a photochemical model by Aumont et al.
(2000), for example. Understanding the mechanisms of formation of SOPM is important
because SOPM can contribute in a significant way to ambient PM levels, especially during
photochemical smog episodes. Experimental studies of the production of secondary organic PM
in ambient air have focused on the Los Angeles Basin. Turpin and Huntzicker (1991, 1995) and
Turpin et al. (1991) provided strong evidence that secondary PM formation occurs during
periods of photochemical ozone formation in Los Angeles and that as much as 70% of the OC
in ambient PM was secondary in origin during a smog episode in 1987. Schauer et al. (1996)
estimated that 20 to 30%, on an annually averaged basis, of the total organic carbon PM in the
< 2.1 |im size range in the Los Angeles airshed is secondary in origin.
Pandis et al. (1992) identified three mechanisms for formation of SOPM: (1) condensation
of oxidized end-products of photochemical reactions (e.g., ketones, aldehydes, organic acids, and
hydroperoxides), (2) adsorption of semivolatile organic compounds (SVOCs) onto existing solid
particles (e.g., poly cyclic aromatic hydrocarbons), and (3) dissolution of soluble gases that can
undergo reactions in particles (e.g., aldehydes). The first and third mechanisms are expected to
be of major importance during the summer when photochemistry is at its peak. The second
pathway can be driven by diurnal and seasonal temperature and humidity variations at any time
of the year. With regard to the first mechanism, Odum et al. (1996) suggested that the products
of the photochemical oxidation of reactive organic gases are semivolatile and can partition
themselves onto existing organic carbon at concentrations below their saturation concentrations.
Thus, the yield of SOPM depends not only on the identity of the precursor organic gas but also
on the ambient levels of OC capable of adsorbing the oxidation products.
Haagen-Smit (1952) first demonstrated that hydrocarbons irradiated in the presence of NOX
produce light scattering aerosols. The aerosol-forming potentials of a wide variety of individual
3-65
-------
anthropogenic and biogenic hydrocarbons were compiled by Pandis et al. (1992), based mainly
on estimates made by Grosjean and Seinfeld (1989) and on data from Pandis et al. (1991) for
p-pinene and from Izumi and Fukuyama (1990) for aromatic hydrocarbons. Zhang et al. (1992)
examined the oxidation of a-pinene. Pandis et al. (1991) found no aerosol products formed in
the photochemical oxidation of isoprene, although they and Zhang et al. (1992) found that the
addition of isoprene to reaction mixtures increased the reactivity of the systems studied. Further
details about the oxidation mechanisms and SOPM yields from various reactive organic gases
are given in the above studies. Estimates of the production rate of SOPM in the Los Angeles
airshed are provided in the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996).
More recently, Odum et al. (1997a,b) have found that the aerosol-formation potential of
whole gasoline vapor can be accounted for solely by summing the contributions of the individual
aromatic compounds in the fuel. In general, data for yields for SOPM formation can be broken
into two distinct categories. The oxidation of toluene and aromatic compounds containing ethyl
or propyl groups (i.e., ethylbenzene, ethyltoluene, n-propylbenzene) produced higher yields of
SOPM than did the oxidation of aromatic compounds containing two or more methyl groups
(i.e., xylenes, di-, tri-, tetra-methylbenzenes). Yields in the first group ranged from about 7 to
10%; in the second group, yields ranged from 3 to 4% for OC concentrations between 13 and
100 |ig/m3. Reasons for the differences in SOPM yields found between the two classes of
compounds are not clear.
There have been a few recent studies that have examined the composition of secondary
organic PM. Edney et al. (2001) carried out a smog chamber study to investigate the formation
of multifunctional oxygenates from photooxidation of toluene. The experiments were carried
out by irradiating toluene/propylene/NOx/air mixtures in a smog chamber operated in the
dynamic mode and analyzing the collected aerosol by positive chemical ionization GC-MS after
derivatization of the carbonyl oxidation products. The results of the GC-MS analyses were
consistent with the formation of semivolatile multifunctional oxygenates, including hydroxy
diones as well as triones, tetraones, and pentaones. The authors also suggested that some of
these compounds could be present in SOPM in the form of polymers.
Jang and Kamens (200la) employed a number of analytical approaches, including GC-MS
detection of volatile derivatives of carbonyl, hydroxy, and acid compounds in SOPM formed in
the irradiation of toluene/NOx mixtures. A wide range of substituted aromatics, nonaromatic
3-66
-------
ring-retaining and ring-opening products were detected. Newly identified ring-opening
oxycarboxylic acids detected included glyoxylic acid; methylglyoxylic acid; 4-oxo-2-butenoic
acid; oxo-C5-alkenoic acids; dioxopentenoic acids; oxo-C7-alkadienoic acids; dioxo-C6-alkenoic
acids; hydroxydioxo-C7-alkenoic acids; and hydroxytrioxo-C6-alkanoic acids. Other newly
identified compounds included methylcyclohexenetriones; hydroxymethylcyclohexenetriones;
2-hydroxy-3-penten-l,5-dial, hydroxyoxo-C6-alkenals; hydroxy-C5-triones, hydroxydioxo-C7-
alkenals; and hydroxy-C6-tetranones. Included among these compounds were a number of the
hydroxy polyketones detected by Edney et al. (2001).
Recent laboratory and field studies support the concept that nonvolatile and semivolatile
oxidation products from the photooxidation of biogenic hydrocarbons contribute significantly to
ambient PM concentrations in both urban and rural environments. The oxidation of a variety of
biogenic hydrocarbons emitted by trees and plants, such as terpenes (
-------
aerosol reactivity for a number of aromatic and biogenic compounds for four initial mixtures.
Incremental aerosol reactivity ranged from 0.133 to 10.352 |ig/m3/ppb and varied by almost a
factor of two depending on the initial mixture.
A number of multifunctional oxidation products produced by the oxidation of biogenic
hydrocarbons have been identified in laboratory studies (Yu et al., 1998; Glasius et al., 2000;
Chri staffer sen et al., 1998; Koch et al., 2000; and Leach et al., 1999). Many of these compounds
have subsequently been identified in field investigations (Yu et al., 1999; Kavouras et al., 1998,
1999a,b; Pio et al., 2001; and Castro et al., 1999). Most studies of the formation of secondary
organic aerosol formation from terpenes have focused on their reactions with ozone. There have
been many fewer studies dealing with the oxidation of terpenes initiated by OH radicals. Larsen
et al. (2001) found that the major aerosol products produced ultimately from the reaction of OH
radicals with monoterpenes with endocyclic double bonds (a-pinene, 3-carene) were C10
keto-carboxylic acids (such as pinonic and caronic acids); whereas the major products from the
oxidation of monoterpenes with exocyclic double bonds (p-pinene) were C9-dicarboxylic acids
(such as pinic acid), and the major product from the oxidation of limonene (which has both
endo- and exocyclic double bonds) was 3-acetyl-6-oxo-heptanal (keto-limonaldehyde). A large
number of related aldehydes, ketones and acids were also found in their experiments. However,
the total yields of condensable products are much lower than for the corresponding reactions
with ozone. For example, yields of C9-dicarboxylic acids, C10-hydroxy-keto-carboxylic acids,
and C10-hydroxy-keto-aldehydes from the reaction of ozone with mono-terpenes with endocyclic
double bonds ranged form 3 to 9%, whereas they ranged only from 0.4 to 0.6% in the reaction
with OH radicals. Likewise, the reaction of monoterpenes with exocyclic double bonds with O3
produced much higher yields (1 to 4%) of C8- and C9-dicarboxylic acids than did their reaction
with OH radicals (0.2 to 0.3%). Apart from the complex products noted above, it should be
remembered that much simpler products, such as formaldehyde and formic acid, are also formed
in much larger yields from the same reactants (e.g., Winterhalter et al., 2000). Compounds such
as these also contribute to the formation of secondary organic aerosol according to the
mechanisms given in Pandis et al. (1992) and mentioned earlier in this section.
It is worth noting that the dicarboxylic acids and hydroxy-keto-carboxylic acids have very
low vapor pressures and may act as nucleating species in OH- and O3-terpene reactions (Larsen
et al., 2001). The rate coefficient for reaction of a-pinene with OH radicals is approximately a
3-68
-------
factor of 106 greater than for its reaction with O3, based on data given in Atkinson (1994). The
daytime average concentration of O3 is typically a factor of 106 greater than that for OH radicals
in polluted boundary layers, whereas the above-mentioned yields of aerosol products are roughly
a factor often greater in the O3-initiated reaction than in the corresponding OH radical reaction.
The foregoing analysis suggests that the O3-initiated reaction may be more important than the
OH-initiated reaction for the formation of aerosol products. During the day, new particles may
be generated indoors through the infiltration of ambient O3; and, because ambient O3 is also
present at night in lower concentrations, new particles may be generated under these conditions
at lower rates. For example, Wainman et al. (2000) found that O3 can react with limonene
released by air fresheners in indoor environments to produce substantial quantities of submicron
particles. The corresponding reaction involving OH radicals outdoors at night is expected to be
negligible by comparison because of the very low OH concentrations present. Sarwar et al.
(2002) estimated indoor OH radical concentrations and suggested that OH in indoor
environments is produced mainly by reactions of O3 transported from outdoors and terpenes
emitted from indoor sources. They reported that indoor OH levels (1-5 x 10s OH/cm3) are
usually lower than typical urban outdoor daytime OH levels (1-5 x 106 OH/cm3). However, they
can be greater than typical urban outdoor night time OH levels (1-5 x 104 OH/cm3). Although
much progress has been made in determining the importance of anthropogenic and biogenic
hydrocarbons for the formation of SOPM, further investigations are needed to accurately assess
their overall contributions to PM2 5 concentrations.
Reactions of organic compounds either in particles or on the surface of particles have only
come under study during the past 20 years. Tobias and Ziemann (2000) reported evidence for
the formation of relatively stable low-volatility peroxy hemiacetals from reactions of
hydroperoxides with aldehydes and ketones on the surface of secondary organic particles.
Not long after the publication of these results, Jang and Kamens (200 la) suggested, based on
results of their outdoor Teflon chamber studies of SOPM formation from irradiation of
toluene/propylene/NOx/air mixtures, that carbonyls and hydroxy compounds (either within or on
the surface of aromatic SOPM) could react together to form larger and less volatile hemiacetals
and acetals. They also proposed that dissolved carbonyls could undergo further reactions
leading to the formation of a polymer, a mechanism that has also been suggested by Edney et al.
(2001). Jang and Kamens (2001b) carried out a series of screening experiments to assess
3-69
-------
whether volatile carbonyl compounds adsorbed onto particles could undergo heterogeneous
reactions forming low vapor pressure compounds. Experiments were carried out in which
aldehydes were introduced in Teflon bags in the dark in the presence of a seed aerosol containing
either ammonium sulfate or a mixture of ammonium sulfate and sulfuric acid. The increase in
the aerosol volume was then measured using a scanning mobility particle sizer. The aldehydes
employed for the study included glyoxal, butanal, hexanal, octanal, and decanal. Increased
organic aerosol yields were found in the presence of the ammonium sulfate seed aerosol for each
of the carbonyls; the highest yield was found for hexanal followed in decreasing order by
glyoxal and then octanal. The presence of the acidified sulfate salt significantly increased the
yields even further. In a number of other experiments, 1-decanol was added to the carbonyl-
aerosol system to investigate the possible formation of hemiacetals and/or acetals. Again, the
volume of aerosol increased in both the presence of ammonium sulfate aerosol and the acidified
salt with a significantly larger yield found in the presence of acidity.
To explain their findings for acid-catalyzed carbonyl reactions, Jang and Kamens (2001a,b)
proposed a chemical mechanism in which the dissolved carbonyl first undergoes a protonization
reaction forming an adduct that can react with water to form its hydrate (1,1-dihydroxy gem-
diol). The adducts can then react with OH groups of the gem-diol, forming higher molecular
weight and less volatile dimers that are subject to further reactions. In principal, this process,
which the authors refer to as a "zipping reaction" can lead to the formation of polymers.
However, because the individual reactions are reversible, the process can be reversed by an
unzipping reaction. The zipping process could serve as an important mechanism for SOPM
formation by converting volatile oxidation products, including glyoxal and methyl glyoxal, into
low-volatility compounds. On the other hand, the unzipping process that could take place during
the workup of the aerosol samples could be responsible for the detection of highly volatile
oxidation products in SOPM (including glyoxal and methyl glyoxal) reported by Edney et al.
(2001), Cocker et al. (2001), and Jang and Kamens (2001a). While these processes may take
place in the absence of significant acidity, the experimental results suggest that the processes are
likely enhanced by acid-catalyzed reactions. Further research is needed to determine the
importance of the mechanisms proposed above for the ambient atmosphere.
Sampling and characterizing PM in the ambient atmosphere and in important
microenvironments is required to address important issues in exposure, toxicology, and
3-70
-------
compliance. Currently, it is not possible to fully quantify the concentration, composition, or
sources of the organic components. Many of the secondary organic aerosol components are
highly oxidized, difficult-to-measure multifunctional compounds. Additional laboratory studies
are needed to identify such compounds, strategies need to be developed to sample and measure
such compounds in the atmosphere, and models of secondary organic aerosol formation need to
be improved and added to air quality models in order to address issues related to human
exposure.
A high degree of uncertainty is associated with all aspects of the calculation of secondary
organic PM concentrations. This is compounded by the volatilization of OC from filter
substrates during and after sampling as well as potential positive artifact formation from the
absorption of gaseous hydrocarbon on quartz filters. Significant uncertainties always arise in the
interpretation of smog chamber data because of wall reactions, sampling artifacts, and the use of
unrealistically high concentrations of reactants. Limitations also exist in extrapolating the
results of smog chamber studies to ambient conditions found in urban airsheds and forest
canopies. Concentrations of terpenes and NOX are much lower in forest canopies (Altshuller,
1983) than the levels commonly used in smog chamber studies. The identification of aerosol
products of terpene oxidation has seldom been a specific aim of field studies, making it difficult
to judge the results of model calculations of secondary organic PM formation.
Uncertainties also arise because of the methods used to measure biogenic hydrocarbon
emissions. Khalil and Rasmussen (1992) found much lower ratios of terpenes to other
hydrocarbons (e.g., isoprene) in forest air than were expected based on their relative emissions
strengths and rate coefficients for reaction with OH radicals and O3. In many cases, reactions
with these species are capable of reducing the concentrations of monoterpenes to beneath
minimum detection levels as has been found by others in a wide range of North American forest
ecosystems (Guenther et al., 1996; Helmig et al., 1998; Geron et al., 2000). Thus, making
judgments about the importance of additional loss processes can be a highly problematic
exercise, given uncertainties in obtaining relevant OH radical and O3 concentrations and in the
reaction rate coefficients. The ocimenes and sesquiterpenes are estimated to have half lives of
seconds to minutes in the presence of ambient O3 levels, while the pinenes and isoprene have
lifetimes ranging from several hours to days. Reaction with hydroxyl radicals is the major sink
for isoprene. Khalil and Rasmussen offered two explanations for their findings: (1) either the
3-71
-------
terpenes were being removed rapidly by some heterogeneous process, or (2) emissions were
enhanced artificially by feedbacks caused by the bag enclosures they used. The isoprene
emission rates observed by Khalil and Rasmussen (1992) are reasonably consistent with those
found by Geron et al. (2000a and 2001), given inevitable self-shading of much of the foliage
within the bag enclosures and also the lower position of the foliage within the accessible portion
of the sampled tree crowns. Temperatures were sufficiently warm (>30 °C) and sunny to induce
high isoprene emission, although cloudiness from a passing thunderstorm could have markedly
reduced these emissions from some of the species. Disturbance or elevated temperatures may
also have induced elevated monoterpene emissions from several of the species sampled by
Khalil and Rasmussen, because emissions rates reported by them are in many cases 2 to 20 times
higher than those reported by others (including those who performed enclosure studies that did
not cause elevated temperatures, as well as micrometeorological flux studies that did not disturb
the forest canopy). Thus, the somewhat reduced isoprene emissions, combined with elevated
monoterpene emissions, can indeed affect the comparison of ambient isoprene versus
monoterpene emissions. However, monoterpene compounds recently have been found to
undergo heterogenous reactions on the surface of acid aerosol particles. Further work is needed
to assess the importance of these reactions for ambient monoterpene concentrations and for the
rate of production of secondary organic PM in forest ecosystems.
3.3.2 Source Contributions to Ambient PM Determined by Receptor Models
Receptor models are perhaps the primary means used to estimate the contributions of
different source categories to PM concentrations at individual monitoring sites. Dispersion
models (i.e., three-dimensional chemistry and transport models) are formulated in a prognostic
manner (i.e., they attempt to predict species concentrations using a tendency equation that
includes terms based on emissions inventories, atmospheric transport, chemical transformations,
and deposition). Receptor models are diagnostic in their approach (i.e., they attempt to derive
source contributions based either on ambient data alone or in combination with data from the
chemical composition of sources). These methods have the advantage that they do not invoke all
of the uncertainties inherent in emissions inventories or in parameterizing atmospheric transport
processes in grid point models.
3-72
-------
There are two main approaches to receptor modeling. Receptor models such as the
chemical mass balance (CMB) model (Watson et al., 1990a) relate source category contributions
to ambient concentrations based on analyses of the composition of ambient PM and source
emissions samples. This technique has been developed for apportioning source categories of
primary PM and was not formulated to include the processes of secondary PM formation. In the
second approach, various forms of factor analysis are used that rely on the varying mix of
species present in observations of compositional data to derive both the composition of sources
and the source contributions. Few of these techniques actually utilize the information contained
in the temporal sequences of the data. Standard approaches such as factor analysis or Principal
Component Analysis (PCA) can apportion only the variance, but not the mass, in an aerosol
composition data set. The other techniques described below, PMF and UNMIX, do apportion
mass, however. Positive matrix factorization (PMF) is a recently developed multivariate
technique (Paatero and Tapper, 1993, 1994) that overcomes many of the limitations of standard
techniques, such as PCA, by allowing for the treatment of missing data and data near or below
detection limits. This is accomplished by weighting elements inversely according to their
uncertainties. Standard methods such as PCA weight elements equally regardless of their
uncertainty. Solutions also are constrained to yield non-negative factors. Both the CMB and the
PMF approaches find a solution based on least squares fitting and minimize an object function.
Both methods provide error estimates for the solutions based on estimates of the errors in the
input parameters. It should be remembered that the error estimates often contain subjective
judgments. For a complete apportionment of mass, all of the major sources affecting a
monitoring site must be sampled for analysis by CMB, whereas there is no such restriction in the
use of PMF.
Among other approaches, the UNMIX model takes a geometric approach that exploits the
covariance of the ambient data to determine the number of sources, the composition and
contributions of the sources, and the uncertainties (Henry, 1997). A simple example may help
illustrate the approach taken by UNMIX. In a two-element scatter plot of ambient Al and Si, a
straight line and a high correlation for Al versus Si can indicate a single source for both species
(soil), while the slope of the line gives information on the composition of the soil source. In the
same data set, Fe may not plot on a straight line against Si, indicating other sources of Fe in
addition to soil. More importantly, the Fe-Si scatter plot may reveal a lower edge. The points
3-73
-------
defining this edge represent ambient samples collected on days when the only significant source
of Fe was soil. Success of the UNMIX model hinges on the ability to find these "edges" in the
ambient data from which the number of source types and the source compositions are extracted.
UNMIX uses principal component analysis to find edges in m-dimensional space, where m is the
number of ambient species. The problem of finding edges is more properly described as finding
hyperplanes that define a simplex. The vertices at which the hyperplanes intersect represent pure
sources from which source compositions can be determined. However, there are measurement
errors in the ambient data that "fuzz" the edges, making them difficult to find. UNMIX employs
an "edge-finding" algorithm to find the best edges in the presence of error. UNMIX does not
make explicit use of errors or uncertainties in the ambient concentrations, unlike the methods
outlined above. This is not to imply that the UNMIX approach regards data uncertainty as
unimportant, but rather that the UNMIX model results implicitly incorporate error in the ambient
data. The underlying philosophy is that the uncertainties are often unquantifiable and, hence,
it is best to make no a priori assumptions about what they are.
In addition to chemical speciation data, Norris et al. (1999) showed that meteorological
indices could prove useful in identifying sources of PM responsible for observed health effects
(specifically asthma) associated with exposure to ambient PM. They examined meteorology
associated with elevated pollution events in Spokane and Seattle, WA and identified a
"stagnation index" that was associated with low wind speeds and increases in concentrations of
combustion-related pollutants. Their factor analysis also identified a meteorological index (low
relative humidity and high temperatures) that was associated with increases in soil-derived PM
as well as a third factor (high temperatures and low relative humidity) that was associated with
increases in concentrations of particulate sulfate and nitrate species (Norris, 1998). In addition
to these variables, past investigations have also used cooling and heating degree days.
Ondov (1996) examined the feasibility of using sensitive isotopic and elemental tracer
materials to determine the contributions of petroleum-fueled sources of PM10 in the San Joaquin
Valley, CA. Costs of these experiments are affected not only by the expense for tracer materials,
but also by the sensitivities of the analytical methods for each tracer, as well as their background
levels. Suarez et al. (1996) used iridium as a tracer to tag emissions from diesel-burning
sanitation trucks in Baltimore and determined the size distribution of soot from the trucks.
3-74
-------
A number of specialty conference proceedings, review articles, and books have been
published that provide greater detail about source category apportionment receptor models than
described in the 1996 PM AQCD. A review of the various methods used to apportion PM in
ambient samples among its source categories was given in Section 5.5.2 of the 1996 PM AQCD.
The collection of the source category characterization profiles shown in Appendix 3D has been
motivated in many cases by the need to use them in receptor modeling applications.
The results of several source apportionment studies are discussed in this section to provide
an indication of the relative importance of different sources of ambient PM across the United
States. First, results obtained mainly by using the CMB approach for estimating contributions
to PM2 5 from different source categories at monitoring sites in the United States are discussed
and presented in Table 3-9. More recent results using the PMF approach are included for
Phoenix, AZ. Results obtained at a number of monitoring sites in the central and western United
States by using the CMB model for PM10 are shown in Table 3-10. The sampling sites represent
a variety of different source characteristics within different regions of Arizona, California,
Colorado, Idaho, Illinois, Nevada, and Ohio. Definitions of source categories also vary from
study to study. The results of the PM10 source apportionment studies were given in the 1996 PM
AQCD and are presented here to allow easy comparison with results of PM25 source
apportionment studies. Chow and Watson (2002) present a detailed comparison of numerous
studies using the CMB model performed mainly after 1995. Their comprehensive review
includes the results of a number of studies that have been published as reports, such as the
Northern Front Range Air Quality Study, in addition to those that have appeared in peer-
reviewed j ournals.
There are several differences between the broadly defined source categories shown at the
tops of Tables 3-9 and 3-10. These differences reflect the nature of sources that are important
for producing the fine and coarse PM shown in Table 3-8. They also are related to
improvements in the ability to distinguish between sources of similar nature (e.g., diesel and
gasoline vehicles, meat cooking, and vegetation burning). The use of organic tracers allows
motor vehicle emissions to be broken down into contributions from diesel and gasoline vehicles.
In studies where this distinction cannot be made, the source type is listed as 'total motor
vehicles' in the tables. The studies that were reported to be able to distinguish gasoline- from
diesel-fueled vehicles reported mixed results for the contributions from gasoline and diesel
3-75
-------
TABLE 3-9. RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
-2.5
Sampling Site
Pasadena, CA 19821
Downtown LA, CA 19821
West LA, CA 19821
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 WNWj 1986-875
Grover City, IL NNWj 1986-875
Reno, NV Summer 19986
Phoenix, AZ Summer 1995-987
Phoenix, AZ Winter 1995-987
Measured Total
PM25 Motor
Concentration Vehicles
28.2 —
32.5 —
24.5 —
42.1 —
39.5 24.5
52 16
63 13
27 8.5
28.3 9.2
26 5.8
—
_
2.4
—
7.8 68
8.3 —
13.8 —
Gasoline Road
Diesel Vehicles Dust, Soil
18.8 5.7 12.4
35.7 6.5 11.1
18 5.7 12.2
12.8 0.7 13.1
— — 1.2
— — <3
— — <3
— — 4.4
— — 3.2
— — 2.7
— — 2.3
_ _ _
— 5.1
— — 3.1
— — 14.5
10.9 36.2 1.8
14.5 38.9 1.1
% Contribution
Vegetation Secondary Secondary
Burning Sulfate Nitrate
9.6 20.9 7.4
5.8 20.3 9.2
11 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 —
4112
15 — —
8.9 — —
Misc. Misc.
Source 1 Source 2
5.3" 9.2b
3.7" 9.2b
4.1" 9.4b
4.5" 12.1b
_ _
— —
2.28 1.9h
2.58 2.5h
0.88 1.5h
9.7k 3.01
11. 6k 11.91
2.8k —
3.41 3.0"
0.6" —
20.8° 4.9r
9.5" 4.5r
Misc. Misc. Total %
Source 3 Source 4 Allocated
8.5C l.ld 98.9
5.2C 0.6d 107.3
8.2C 1.6d 102.1
4.5C 0.5d 87.9
— — 84.9
— — <80
— — <85
0.41 — 99.3
0.71 — 99.8
0.41 — 95.8
1.28 — 99.4
4.18 4.6m 91.2
— — 98.8
— — 96
— — 100.1
6.7s 3.6" 99.9
18.7s 4.1" 100.2
'Schauer et al. (1996)
2Motallebi (1999)
3Magliano et al. (1998)
4Dzubayetal(1988)
5Gloveretal. (1991)
'Gillies et al. (2000)
'Ramadan et al. (2000)
"Secondary and other organic compounds
"•Secondary ammonium
°Meat cooking
dVegetative detritus
'Value represents sum of diesel and gasoline
vehicle exhaust
'Including associated cations and water
incinerators
hOil fly ash
'Fluidized catalyst cracker
jWind direction
kLead smelter
'Iron works
"Copper smelter
"Coal power plant
"As ammonium sulfate
pAs ammonium nitrate
"Sea salt
"Wood burning
8Nonferrous smelting
-------
TABLE 3-10. RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
10
% Contribution
Measured
PMIO
Sampling Site Concentration
Craycroft, AZ
Winter 1989-19901
Hayden 1, AZ 19861
Hayden2, AZ 19861
Rillito, AZ 19882
Bakersfield, CA
1988-19893
Crows Landing, CA
1988-19893
Fellows, CA
1988-19893
Fresno, CA
1988-19893
Indio, CA4
Kern 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
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
87.4
114.8
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
Primary
Construction
0
1.9b
6.8b
17.4b
2
0.0
2.6
0.0
5.2
4.2
0.0
0.0
0.0
0.0
4.6
3.9
Primary
Motor
Vehicle
Exhaust
35.5
0.0
0.0
1.5f
9.7
4.2
3.8
9.5
7.6
4.6
9.8s
13.7
44.5
10.9
6.4s
15.1
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
Secondary
Ammonium
Sulfate
3
3.8
6.8
0
6.9
5.3
9.3
5
6.2
6.9
15.4
23.6
4
7.5
7.3
8.3
Secondary
Ammonium
Nitrate
2.6
0.0
0.0
0
16
12.4
13.7
14.5
7.1
3.1
17.7
1.7
24.1
33.4
24.4
23.9
Misc.
Source 1
s.r
70.5C
47.5C
14.68
1.3m
1.0m
12.8"1
0.4m
0.3J
1.0m
0.2J
0.2j
0.0s
0.5J
0.3j
0.0>
Misc.
Source 2
0.0
4.811
0.0
0.0
1.9"
1.9"
2.6"
1.9"
1.7"
3.1"
3.9"
4.8h
2.8h
2.0h
1.1"
4.4"
Misc.
Source 3
0
1.0e
1.7
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
Misc. Total %
Source 4 Allocated
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
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
-------
TABLE 3-10 (cont'd). RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
10
oo
% Contribution
Measured
PMIO
Sampling Site Concentration
Rubidoux, CA
Fall 19876
Rubidoux, CA 19887
San Nicolas Island, CA
Summer 19876
Stockton, CA 19893
Pocatello, ID 1990s
S.Chicago, IL 19869
S.E. Chicago, IL
198810
Reno,NV 1986-87"
Sparks, NV 1986-87"
Follansbee, WV 199 112
Mingo, OH 199112
Steubenville, OH
199112
112
87
17.4
62.4
100.0
80.1
41
30.0
41
66.0
60.0
46
Primary
Geological
17.1
55.2
9.2
55.1
8.3
34
35.9V
49.7
36.8
15.2
20
18
Primary
Construction
14.4
0.0
0
0.8
7.5"
3
0.0
0.0
0.0
0.0
0.0
0.0
Primary
Motor
Vehicle
Exhaust
27.1
11.7
5.2
8.3
0.1
3.5
2.2f
33.3
28.3
53
23.3
30.4
Primary
Vegetative
Burning
0.0
0.0
0.0
7.7
0.0
0.0
0.0
6.3
32.7
0.0
6.8
1.7
Secondary
Ammonium
Sulfate
1.9
6.1
21.3
5
0
19.2s
18.8
4.3
6.6
24.2
25
30.4
Secondary
Ammonium Misc. Misc.
Nitrate Source 1 Source 2
28.2 0.0> 1.0"
24.9 0.& 1.7h
2.9 0.0s 24.7h
11.2 l.lm 2.9"
0 0 0.0
— 18.9' 2.7"
— 2.0' 0.7h
2 0.0 0.0
2.2 0.0 0.0
— 14.1' 0.0
— 5.7' 18.3X
— 8.3' 10.9X
Misc.
Source 3
0
6.6°
0
0.0k
84. lr
0.0
2.7"
0.0
0.5k
0.0
0.0
0.0
Misc.
Source 4
0
0.0
0
0
0
0.0
18.88
0.0
0.0
0.0
0
0.0
Total %
Allocated
89.7
106.8
63.3
92.1
100
81.3
81.1
95.6
107.1
106.5
99.1
99.7
'Chowetal. (1992a)
2Garfield; Ryan et al. (1988)
3Jail; Ryan etal. (1988)
"Thanukos etal. (1992)
5Chowetal. (1992b)
6Kim etal. (1992)
'Gray etal. (1988)
8Watson etal. (1994)
'Chowetal. (1992c)
10Houcketal. (1992)
"Hopke etal. (1988)
12Vermette etal. (1992)
"Chowetal. (1988)
"Skidmore etal. (1992)
"Smelter background aerosol
bCement plant sources, including
kiln stacks, gypsum pile, and
kiln area
'Copper ore
''Copper tailings
'Copper 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
kSecondary organic carbon
'Biomass burning
"Primary crude oil
"NaCl + NaN03
"Lime
pRoad sanding material
'Asphalt industry
"Phosphorus/phosphate industry
sRegional sulfate
'Steel mills
"Refuse incinerator
"Local road dust, coal yard road
dust, and steel haul road dust
"Incineration
"Unexplained mass
-------
vehicles. In the study in southern California, diesels made a larger contribution to PM2 5 than did
gasoline-fueled vehicles. On the other hand, the study in Phoenix, AZ, found the opposite result.
Meat cooking is also distinguished from vegetation burning in more recent studies although both
are considered to be part of biomass burning. Vegetation burning consists of contributions from
residential fuel wood burning, wildfires, prescribed burning, and burning of agricultural and
other biomass waste. Miscellaneous sources of fine particles include contributions from
combustion sources, whereas miscellaneous sources of coarse particles consist of contributions
from soil and sea spray and industrial processing of geological material (e.g., cement
manufacturing). Although a large number of elements and chemical components are used to
differentiate among source categories and although there can be a large number of source types
affecting a given site, only a few broadly defined source types are needed to account for most of
the mass of PM2 5 and PM10. At any given site, < 5 source types account for > 65% of the mass
of PM25 (Table 3-9); and <5 source types account for > 65% of the mass of PM10 (Table 3-10).
Secondary sulfate is the dominant component of PM25 samples collected in the studies of
Dzubay et al. (1988) and Glover et al. (1991). Both studies found that sulfate at their monitoring
site arose from regionally dispersed sources. Sulfate also represented the major component
of PM25 found in monitoring studies in the eastern United States shown in Appendix 6 A of the
1996 PM AQCD. Primary and secondary organic components also make major contributions
to PM2 5 and in many areas may represent a larger component than sulfate. Contributions from
road dust and soils are relatively minor, typically constituting < 10% of PM25 in the studies
shown in Table 3-9. Studies in the western United States shown in Table 3-9 have found larger
contributions from motor vehicles, fugitive dust, and ammonium nitrate. The most notable
difference in the relative importance of major source categories of PM25 shown in Table 3-9 and
PM10 shown in Table 3-10 involves crustal material, (e.g., soil, road dust), which represents
about 40% on average of the total mass of PM10 in the studies shown in Table 3-10. The fraction
is higher at sites located away from specific sources, such as sea spray or smelters. Emissions of
crustal material are concentrated mainly in the PM10_2 5 size range.
In Table 3-10, the magnitude of primary motor vehicle exhaust contributions is highly
variable and ranges from 5 to 40% of average PM10. Vehicle exhaust contributions are also
variable at different sites within the same study area. The mean value and the variability of
motor vehicle exhaust contributions reflects the proximity of sampling sites to roadways and
3-79
-------
traffic conditions during the time of sampling. Many studies were conducted during the late
1980s when a portion of the vehicle fleet still used leaded gasoline. Pb and Br in motor vehicle
emissions facilitated the distinction of motor vehicle contributions from other sources. Vehicles
using leaded fuels have higher emission rates than vehicles using unleaded fuels. Pb also
poisons automobile exhaust catalysts and produces adverse human health effects. As a result, Pb
has been eliminated from vehicle fuels. Organic species such as w-pentacosane through
w-nonacosene, cholestanes, ergostanes, sitostanes, and hopanes have replaced Pb as a source
marker for motor vehicle emissions (e.g., Schauer and Cass, 2000). In their comprehensive
review of CMB modeling studies undertaken since 1995, Chow and Watson (2002) note that,
in 22 studies, fossil fuel combustion was found to be a large contributor to PM2 5 and PM10
concentrations, with most of the contributions to primary PM originating from the exhaust of
diesel and gasoline vehicles.
Marine aerosol is found, as expected, at coastal sites such as Long Beach (average 3.8% of
total mass) and San Nicolas Island (25%). These contributions to PM10 are relatively variable
and are larger at the more remote sites. Individual values reflect proximity to local sources.
Of great importance are the contributions from secondary ammonium sulfate in the eastern
United States and ammonium nitrate in the western United States. Secondary ammonium sulfate
is especially noticeable at sites in California's San Joaquin Valley (Bakersfield, Crows Landing,
Fellows, Fresno, and Stockton) and in the Los Angeles area.
Because many source apportionment studies address problems in compliance with the
National Ambient Air Quality Standards (NAAQS) and other air quality standards, samples
selected for chemical analysis are often biased toward the highest PM10 mass concentrations in
the studies shown in Table 3-10. Thus, the average source contribution estimates shown in
Table 3-10 are probably not representative of annual averages and may not be representative of
a large spatial area for some source-dominated studies. For example, the study by Motallebi
(1999) considered only days when the PM10 concentration was greater than 40 |ig/m3. Quoted
uncertainties in the estimated contributions of the individual sources shown in Tables 3-9 and
3-10 range from 10 to 50%. Errors can be much higher when the chemical source profiles for
different sources are highly uncertain or are too similar to distinguish one source from another.
Very few source apportionment studies using the CMB modeling technique have examined
the spatial variability of source contributions at different sites within an urban area. As can be
3-80
-------
seen from Table 3-9, Dzubay et al. (1988) found a uniform distribution of sulfate at the NE
Airport in Philadelphia, PA, downtown Camden, NJ, and Clarksboro, NJ during the summer of
1982. The farthest distance between two monitoring sites (NE Airport and Clarksboro) was
-40 km. Magliano et al. (1998) examined the spatial variability of PM10 source contributions at
a number of sites in Fresno and Bakersfield, CA during the winter of 1995-1996 and reported
values for 1 day (December 27, 1995). During that day, mobile sources contributed from 13.0 to
15.8 |ig/m3, vegetation burning 5.1 to 11.1 |ig/m3, ammonium sulfate 2.4 to 3.4 |ig/m3, and
ammonium nitrate 19.3 to 24.6 |ig/m3 to PM10 at the sites in Bakersfield. Mobile sources
contributed 13.9 to 22.5 |ig/m3, vegetation burning 8.2 to 15.7 |ig/m3, ammonium sulfate 1.8 to
2.3 |ig/m3, and ammonium nitrate 14.5 to 18.9 |ig/m3 at the sites in Fresno. All of these
components are expected to be found mainly in the PM2 5 size fraction. As can be seen, source
contributions at different sites varied by factors of 1.2 to 2.2 in Bakersfield and by factors of
1.3 to 1.9 in Fresno on that day.
The receptor modeling methods outlined above do not explicitly include consideration of
the distances between PM sources and the receptor site. Information about the relative
importance of sources as a function of distance may be available from examination of data
obtained by continuous monitoring methods. For example, concentration spikes are expected to
be the result of transport from nearby sources because turbulent mixing in the atmosphere would
not allow them to persist for very long. Short duration spikes in the time series of concentrations
are assumed to result from emissions from local sources (0.1 to 1 km away) in this method.
Contributions from sources located farther away are determined by comparisons between
baselines measured at different sites. Details such as these are also lost in integrated 24-h
samples. Watson and Chow (2001) used time series of black carbon (BC) obtained by
aetholometers over 5-min intervals to estimate the contributions from sources located < 1 km
away, 1 to 5 km away, and > 5 km away from a monitoring site in downtown Mexico City.
They found that most of the BC was produced by sources scattered throughout the city and that
sources located less than 1 km away from the site contributed only about 10% to BC
concentrations even in the presence of local sources such as buses and trucks.
3-81
-------
3.3.3 Background Concentrations of PM in the United States
This section contains information about the concentrations of "background" PM that are
relevant for policy setting. For the purposes of this document (and consistent with the 1996
PM AQCD), policy relevant background (PRB) concentrations are those that would result in the
United States from emissions from natural sources worldwide plus anthropogenic sources
outside of North America. In other words, they are the concentrations that would exist if
anthropogenic emissions in North America were zero. PRB concentrations of PM2 5, PM10_2 5,
and PM10 may be viewed as coming from two conceptually separate components: a reasonably
consistent "baseline" component and an episodic component. The baseline component consists
of the contribution from natural sources within North America and from transport of natural and
anthropogenic sources outside of North America that is generally well characterized by a
reasonably consistent distribution of daily values each year, although there is variability by
region and season. The episodic component consists of infrequent, sporadic contributions from
natural high-concentration events over shorter periods of time (e.g., hours to several days ) both
within North America (e.g., volcanic eruptions, large forest fires, dust storms) and outside of
North America (e.g., transport related to dust storms from deserts in North Africa and China and
storms at sea). These episodic natural events, as well as events like the uncontrolled biomass
burning in Central America and southern Mexico, are essentially uncontrollable and do not
necessarily occur in all years. The same processes that are responsible for transporting these
emissions also transport anthropogenic emissions from outside North America.
It is impossible to reliably estimate PRB concentrations solely by examining mass
measurements of PM25, PM10_2 5, and PM10. Rather, this section examines available monitoring
data, source apportionment analysis, and information about various source contributions to
inform a characterization of these two components of PRB PM concentrations and the associated
uncertainties and limitations in estimating PRB concentrations. Appendix 3E presents data from
relatively remote IMPROVE sites to provide some rough upper limits on PRB concentrations for
the United States.
Annual average natural background concentrations of PM10 have been estimated to range
from 4 to 8 |ig/m3 in the western United States and 5 to 11 |ig/m3 in the eastern United States.
Corresponding PM2 5 levels have been estimated to range from 1 to 4 |ig/m3 in the western
United States and from 2 to 5 |ig/m3 in the eastern United States (U.S. Environmental Protection
3-82
-------
Agency, 1996). The range of the natural background concentration of PM10_25 can be roughly
estimated by subtracting the minimum PM2 5 value [1 |ig/m3 (western U.S.) and 2 |ig/m3 (eastern
U.S.)] from the maximum PM10 value [8 |ig/m3 (western U.S.) and 11 |ig/m3 (eastern U.S.)] and
by subtracting the maximum PM25 value [4 |ig/m3 (western U.S.) and 5 |ig/m3 (eastern U.S.)]
from the minimum PM10 value [4 |ig/m3 (western U.S.) and 5 |ig/m3 (eastern U.S.)]. These
calculations indicate that the average PM10_2 5 concentration is < 1 to 7 |ig/m3 in the western
United States and < 1 to 9 |ig/m3 in the eastern United States. The estimated natural background
concentrations given above do not include contributions from long-range transport from sources
outside North America. [The range of annual mean PM2 5 concentrations at IMPROVE network
sites in the western United States (cf, Appendix E) is consistent with the range of values in the
lowest 5th percentile annual mean PM2 5 concentrations for specific sites in the AIRS data base
(2.8 |ig/m3 to 6.9 |ig/m3)]. However, PM25 concentrations are much higher at sites in the eastern
United States than at sites in the western United States. At most IMPROVE sites in the western
United States, the annual mean concentration of PM10_2 5 is higher than that of PM25, and daily
average PM2 5 concentrations are moderately correlated (r = 0.72) with PM10_2 5 concentrations.
In contrast, PM2 5 concentrations are higher than those of PM10_25 at IMPROVE sites in the
eastern United States, and PM2 5 concentrations are only weakly correlated (r = 0.26) with those
of PM10_2 5. As discussed above, peak 24-h average concentrations due to natural sources may be
substantially higher than the annual or seasonal average concentrations from natural sources,
especially within areas affected by wildfires and dust storms.
Long-Range Transport from Outside North America
Windblown dust from dust storms in the Sahara desert has been observed in satellite
images as plumes crossing the Atlantic Ocean and reaching the southeast coast of the United
States (e.g., Ott et al., 1991). Dust transport from the deserts of Asia across the Pacific Ocean
also occurs (Prospero, 1996). Most dust storms in the deserts of China occur in the spring
following the passage of strong cold fronts after the snow has melted and before a surface
vegetation cover has been established. Strong winds and unstable conditions result in the rapid
transport of dust to altitudes of several kilometers, where it is transported by strong westerly
winds out over the Pacific Ocean (Duce, 1995). Satellite images were used to track the progress
3-83
-------
of a dust cloud from the Gobi desert to the northwestern United States during the spring of 1998
(Husar et al., 2000).
Satellite images obtained at visible wavelengths cannot track mineral dust across the
continents because of a lack of contrast between the plume and the underlying surface. Other
means must be used to track the spread of North African dust through the eastern United States.
Perry et al. (1997) used two criteria (PM25 soil concentration > 3 |ig/m3 and Al:Ca > 3.8) to
distinguish between soil of local origin from soil originating in North Africa in characterizing
the sources of PM in aerosol samples collected in the IMPROVE network. North African dust
has been tracked as far north as Illinois (Gatz and Prospero, 1996) and Maine (Perry et al.,
1997). The analysis of Perry et al. (1997) indicates that incursions of Saharan dust into the
continental United States have occurred, on average, about three times per year from 1992 to
1995. These events persist for -10 days on average, mainly during the summer. Large-scale
dust events typically cover from 15 to 30% of the area of the continental United States and result
in -8.7 ± 2.3 |ig/m3 increases of PM25 levels throughout the affected areas during these events,
with mean maximum dust contributions of 19.7 ± 8.4 |ig/m3 and a peak contribution of 32 |ig/m3
to 24-h average PM2 5 levels.
As can be expected, the frequency of dust events is highest in the southeastern United
States. About half of these events are observed only within the state of Florida and are
associated with dense hazes in Miami (Figure 3-22) during the summer (Prospero et al., 1987).
North African dust is the dominant aerosol constituent in southern Florida during the summer,
whereas soil dust constitutes only a minor fraction of PM during the remainder of the year
(Prospero, 1999). Approximately 33 to 50% of the mass of the particles reaching southern
Florida have aerodynamic diameters < 2.5 jim (Prospero et al., 2001). During episodes when
daily total dust concentrations ranged up to 100 |ig/m3, it can be seen that daily PM2 5 values of
up to 50 |ig/m3 could have resulted in Miami, FL.
Husar et al. (2001) documented the transport of dust from the Gobi and Taklimakan deserts
to North America during April 1998. The PM10 concentration averaged over 150 stations in
Washington, Oregon, California, Nevada, and Idaho reporting data to AIRS was 65 |ig/m3
between April 26 and May 1, compared to about 20 |ig/m3 during the rest of April and May.
Data from several networks indicated that PM10 concentrations were over 100 |ig/m3 in central
British Columbia, Washington State, and Oregon. The highest PM concentrations observed
3-84
-------
74 76 78 80 82 84 86 88 90 92 94 96
Figure 3-22. Monthly average Saharan dust components in all size fractions 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).
were 120 |ig/m3 for PM10 and 50 |ig/m3 for PM2 5 at Chilliwack Airport in northwestern
Washington State (Figure 3-23). Aircraft measurements made over the northwestern United
States were consistent with a mass median diameter of the dust being between 2 and 3 jim.
Desert dust deposited over oceans provides nutrients to marine ecosystems (Savoie and
Prospero, 1980), and such dust deposited on nutrient-depleted soils also provides nutrients to
terrestrial ecosystems, e.g., Hawaiian rain forests (Chadwick et al., 1999). Microorganisms,
including fungi and bacteria, have been found attached to African dust particles in the U.S.
Virgin Islands (Griffin et al., 2001). The fungus, Aspergillus sydowii, which has been connected
to the death of coral reefs, has been identified in air samples collected in the Caribbean during
African dust transport events (Smith et al., 1996; Shinn et al., 2000). Measurements of the
3-85
-------
140
120 -•
CO
E 100 -
O 80 -
§ 60 H
O
o
40 -
20 -
4/25 4/27 4/29 5/1
Time (days)
5/3
5/5
Figure 3-23. PM2 5 and PM10 concentrations measured at Chilliwack Airport, located in
southwestern British Columbia, just before and during the Asian desert dust
episode of April and May 1998.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
composition of Saharan dust in Miami indicate enhancements of nitrate, non-sea salt sulfate,
ammonium, and trace metals above concentrations expected for clean marine air, suggesting
pollution emitted in Europe and North Africa as sources (Prospero, 1999). It is likely that many
other constituents will be found associated with dust from outside North America as more
measurements are made. It should be noted that, as North African dust and associated material
are transported northward through the United States during the summer, they are added to the
mixture of primary and secondary PM generated domestically.
The transport of PM from uncontrolled biomass burning in Central America and southern
Mexico resulted in anomalously high PM levels observed in southern Texas and generally
elevated PM concentrations throughout the entire central and southeastern United States during
3-86
-------
the spring and early summer of 1998. Biomass burning for agricultural purposes occurs
normally during the spring of each year in Central America and southern Mexico. During the
spring of 1998, fires burned uncontrollably because of abnormally hot and dry conditions
associated with the intense El Nino of 1997-1998. PM10 concentrations observed in the southern
Rio Grande Valley were elevated substantially during the passage northward of the biomass
burning plume produced by these fires, as shown in Figure 3-24. Elevated PM10 concentrations
also were found as far north as St. Louis, MO (Figure 3-25). As can be seen from Figure 3-24
and Figure 3-25, the elevations in PM concentrations were limited in duration. Uncontrolled
wildfires occur in the United States every year, but their effects on air quality throughout the
United States still need to be evaluated systematically. These fires can be widespread, and the
frequency of their occurrence can vary markedly from year to year. For example, -26,000 km2 of
forested land were consumed during 2000, but only a small fraction of this area was burnt during
2001 in the western United States. Wildfires also occur throughout the boreal forests of Canada.
Wotawa and Trainer (2000) suggested that the plume from fires occurring in the Northwest
Territories of Canada in early July 1995 may have extended throughout most of the eastern
United States, resulting in elevated levels of CO and O3. Simple scaling of their calculated
excess CO concentrations because of the fires, by the ratio of emission factors of PM25 to CO,
indicates that the excess PM2 5 concentrations in the plume may have ranged from ~5 |ig/m3 in
the Southeast and to close to 100 |ig/m3 in the northern Plains States.
Sources Within North America
It is much more difficult to determine 24-h PRB concentrations in the absence of specific
events such as those noted above, because contributions from anthropogenic sources located
either nearby or elsewhere within North America can contribute substantially to observed values
and perhaps overwhelm the contributions from PRB sources. Source-apportionment modeling
techniques (described in Section 5.5 of the 1996 PM AQCD and updated in Section 3.3.2 of this
document) can be helpful for this purpose. It is more likely that PRB concentrations can be
more accurately determined in the West than in the East because contributions from pollution
sources can overwhelm those from background sources in the East, and background source
contributions can be lost within the errors of their source contributions. However, it should be
noted that source apportionment techniques such as PMF and UNMIX are not able to distinguish
3-87
-------
600
5/5 5/6 5/7 5/8 5/9 5/105/115/125/135/145/155/165/175/185/195/205/215/225/235/245/255/265/275/28
Time (days)
Figure 3-24. Time series of 24-h average PM10 concentrations observed in the Rio Grande
Valley during May 1998.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
between bioaerosol and organic PM from whatever source, mainly because of analytical
limitations. In addition, background PM source contributions are contaminated by contributions
from pollution sources during transport from source to receptor monitoring site. Thus,
background source contributions derived by these methods should be regarded as upper limits on
the true values.
Eastern United States
The use of data for PM2 5, PM10.2 5, and PM10 from IMPROVE sites in the East will
generally result in gross overestimates of PRB concentrations because of extensive
contamination by anthropogenic sources. Song et al. (2001) derived source contributions to
PM2 5 concentrations measured at Washington, DC, Brigantine, NJ, and Underbill, VT using
PMF. They found that woodsmoke (consisting of contributions from wildfires and residential
wood burning) could contribute 0.93 ± 0.97 |ig/m3; soil 0.11 ± 0.22 |ig/m3; and sea spray
0.90 ±1.0 |ig/m3 on an annual basis at Brigantine, NJ throughout the 1989 to 1999 sampling
period. They also derived contributions of 1.2 ± 0.9 |ig/m3 from woodsmoke; 0.32 ± 0.61 |ig/m3
-------
200
180 -
160-
140-
?T 12° ~
E
3 100-
o
s
Q- 80-
60-
40-
20-
Smoke
Event
PM1024-h Standard
5 6 7 8 9 10 11 12 13 14 15 16 17
May 1998
Figure 3-25. PM10 concentrations observed in St. Louis, MO, during May 1998.
Source: Aerometric Information Retrieval System (AIRS; U.S. Environmental Protection Agency, 2002b).
from soil; and 0.05 ± 0.05 |ig/m3 from sea spray at Underbill, VT from 1989 through 1999. The
"background" sources contribute about 7% to annual average PM2 5 concentrations at Brigantine
and about 12% at Underbill. The daily time series at the NJ and VT sites show striking
variability in background components, characterized by spikes. Maximum daily values during
these spikes are in the range of several |ig/m3. Song et al. concluded that such spikes in
concentrations are likely caused by transient meteorological events, e.g., storms or transport of
dust from northern Africa or other distant regions or by events such as wild fires. Contributions
from all of these sources should be regarded as upper limits because of entrainment of pollutant
emissions during transport from source to receptor of the background source emissions. Care
should also be taken to ensure that distinctions are made between contributions from wildfires
3-89
-------
versus those from residential wood burning on the basis of seasonality or observations of smoke
plumes when using these results.
Pun et al. (2002) used a regional scale chemistry-transport model (CMAQ) to simulate O3
and PM2 5 concentrations arising from natural sources alone in model domains centered over
Tennessee and over the mid-Atlantic region for several days in July 1995. These calculations
were performed for meteorological conditions that resulted in high ambient ozone concentrations
in the eastern United States. They found that natural sources contributed about 1.7 |ig/m3 to
Washington, DC, ( ranging from about 0.6 to 3.1 |ig/m3 in the mid-Atlantic domain) and about
1.2 |ig/m3 to Nashville, TN, on a 24 h average basis. The formation of secondary organic PM
from biogenic precursors may be expected to be maximized for these conditions; however, their
contribution was estimated to be at most 15% of natural PM25 or less than 1 |ig/m3. The largest
contributions in both cases came from natural PM2 5 that was advected in from other regions of
the United States. In addition to the sources considered above, contributions to both primary and
secondary PM from events such as volcanic eruptions or geothermal activity are highly sporadic.
The spatial and temporal distributions of secondary PM produced by background sources shown
in Table 3-8 still remain to be investigated.
Western United States
As mentioned earlier, it is impossible to obtain reliable estimates of PRB concentrations
solely on the basis of measurements of PM25, PM10_25, or PM10. It is preferable to quantify
contributions from both background and non-background sources by using compositional data in
techniques such as source apportionment modeling. However, of those measured throughout the
United States, the concentrations observed at several RRMS in the western United States
probably come closest to what PRB concentrations might be in the West. In addition to looking
at PM2 5, PM10, and PM10_2 5, the PM2 5 and PM10 mass with the sulfate component removed was
examined to obtain better insight into non-anthropogenic background. However, it should be
noted at the outset that an undetermined fraction of the sulfate will be PRB in nature as the result
of long range transport from Asia, more local crustal sources or volcanic emissions, or from sea-
spray (especially at coastal sites). Data from 19 IMPROVE sites are examined in Appendix 3E
(including two eastern sites). The ranges of annual average non-sulfate PM25 and non-sulfate
PM10 as well as PM10_2 5 (calculated as the difference between PM10 and PM2 5) concentrations are
3-90
-------
shown in Table 3-El and the corresponding ranges for the 90th percentile concentrations are
summarized in Table 3-E2. The IMPROVE monitors did not measure coarse fraction sulfate.
It must be recognized at the outset that these concentrations will only provide upper limits, and it
is not clear over what spatial scale these concentrations can be extrapolated. However, at nine of
the 16 sites in the western United States summarized in Appendix 3E, annual mean
concentrations are within the range of PRB values estimated in the 1996 PM AQCD and given
earlier in this section. At the other seven sites in the western United States, they are consistent
with, although slightly above, the annual average values defined earlier in the 1996 PM AQCD.
The annual mean sulfate concentrations at these western IMPROVE sites range significantly
above the PRB values estimated in the 1996 PM AQCD.
Some screening should be performed to rule out transport from urban areas or other
sources of anthropogenic PM on a day-by-day basis. It is important to note that there is much
smaller relative variability in PM2 5 concentrations (from the minimum to the P90 level) on a
year-to-year basis at the western IMPROVE sites than at the eastern IMPROVE sites. This may
be related to a lack of discrete transport events affecting most samples at the western sites
compared to the eastern sites; but does not rule out the possibility that the western sites are
receiving highly diluted contributions from many distant urban sources. Further inquiry is
needed to address these issues. Likewise, data for PM10_2 5 concentrations at RRMS can be used
for similar purposes.
Summary
It can be concluded from the above discussion that 24-h policy relevant background (PRB)
concentrations are highly variable both spatially and temporally. Contributions to PRB
concentrations from external sources (e.g., Asian and North African dust storms and Central
American wildfires) can be significant on an episodic, but probably not on an annual basis.
More local sources of primary PRB PM are also likely to be episodic, reflecting the occurrence
of volcanic eruptions, wildfires, and storms that raise dust and sea spray. The influence from
events such as these can be felt over thousands of square kilometers. Very little work has been
done to quantify the magnitude and variability of contributions from the production of secondary
PM. However, the one modeling study cited above found values < 1 |ig/m3 as secondary PM.
Perhaps the greatest possibility for estimating these concentrations comes from the application of
3-91
-------
source apportionment techniques such as PMF (positive matrix factorization) to time series of
species compositional data obtained at relatively remote monitoring sites to minimize
interference from anthropogenic sources. In the absence of such results, some useful estimates
may be obtained by examining the time series of PM concentrations at such sites with screening
to eliminate days when concentrations are influenced by anthropogenic sources.
The above discussions do not explicitly consider the contributions from PBP. However,
it should be noted there is very little data available for concentrations of PBP, and their
contributions are likely to be highly spatially and temporally variable, especially on a seasonal
basis. For example, one study conducted in Mainz, Germany (Mathias-Maser, 1998) found that
PBP could contribute from about 17 to 20% of total aerosol volume and from 9 to 30% of total
particle number in the size range from 0.35 jim to 50 jim, depending on season. Also, pollen
can, at times, be a dominant contributor to ambient PM particles larger than 10 jim. However,
rupture of pollen grains (especially during warm temperature and high humidity or rainfall
conditions) can result in release of many smaller-size airborne cytoplasmic fragments ranging
down to around 0.3 to 0.4 |im (see Appendix 7B of Chapter 7 for more information on pollen
and other bioaerosols).
3.3.4 Emissions Estimates for Primary PM, and Precursors to Secondary
PM (SO2, NOX, VOCs, and NH3) in the United States
In principle, source contributions to ambient PM also could be estimated on the basis of
predictions made by chemistry-transport models (CTMs) or even on the basis of emissions
inventories alone. Uncertainties in emissions inventories have arguably been regarded as
representing the largest source of uncertainty in CTMs (Calvert et al., 1993). Apart from
uncertainties in emission inventories, a number of other factors limit the ability of an emissions
inventory-driven CTM to determine the effects of various sources on particle samples obtained
at a particular location. CTM predictions represent averages over the area of a grid cell, which
in the case of CMAQ (Community Model for Air Quality) and MAQSIP (Multiscale Air Quality
Simulation Platform), ranges from 16 km2 (4 km x 4 km) to 1296 km2 (36 km x 36 km). CMAQ
and MAQSIP constitute the CTMs within the overall ModelsS framework, which also includes
emissions processors, the meteorological model, and decision-support modules. The
contributions of sources to pollutant concentrations at a monitoring site are controlled strongly
3-92
-------
by local conditions that cannot be resolved by an Eulerian grid-cell model. Examples would be
the downward mixing of tall-stack emissions and deviations from the mean flow caused by
buildings. The effect of local sources at a particular point in the model domain may not be
predicted accurately because their emissions would be smeared over the area of a grid cell or if
the local wind fields at the sampling point deviated significantly from the mean wind fields
calculated by the model. CTMs also have problems in predicting pollutant concentrations
because of uncertainties in vertical mixing and in predicting concentrations of pollutants from
stationary combustion sources resulting from uncertainties in estimates of plume rise. CTMs are
an integral part of air quality management programs and are reviewed in the NARSTO Fine
Particle Assessment (NARSTO, 2002).
Estimated emissions of primary PM25 from different sources in the United States are
summarized in Table 3-11, and estimated emissions of precursors to secondary PM25 (SO2, NOX,
VOCs, and NH3) are summarized in Table 3-12. These estimates provide a rough overview of
the relative importance of major PM sources in the United States. The emissions estimates are
based on information presented in the EPA National Air Pollutant Emission Trends Report,
1990-1999 (U.S. Environmental Protection Agency, 2001), to which the reader is referred for
detailed tables showing trends in PM25 emissions from a number of source categories from 1990
to 1999. Detailed descriptions of the methodology for constructing emissions inventories for
criteria pollutants, quality assurance procedures, and examples of calculations of emissions can
be found in U.S. Environmental Protection Agency (1999). Although uncertainties associated
with the estimates in the National Air Pollutant Emission Trends Report are not given therein, a
discussion of uncertainties in emissions estimates is given in Section 3.3.5.
For the sake of completeness, an attempt was made to supplement the information given
in the emissions tables in the Trends Report (which concentrates mainly on anthropogenic
emissions) with information about emissions from natural sources. Details regarding the
composition of the emissions of primary PM2 5 from the source categories shown in Table 3-11
are summarized in Appendix 3D, where available. Fugitive dust emissions are estimated to
constitute over 50% of nationwide primary PM2 5 emissions according to Table 3-11. However,
there are a number of issues concerning the methods for obtaining relevant emissions factor data
for fugitive dust in field studies, as discussed in Section 3.3.5. An estimate of the production
of PM2 5 from wind erosion on natural surfaces was not included in Table 3-11, because this
3-93
-------
TABLE 3-11. EMISSIONS OF PRIMARY PM2, BY VARIOUS SOURCES IN 1999
Source
On-road vehicle
exhaust
Non-road vehicle
exhaust
Fossil fuel
combustion
Industrial
processes
Biomass burning
Waste disposal
Fugitive dust
Windblown dust
Other
Total
Emissions
(10'kg/y)
0.21
0.37
0.36
0.35
1.2
0.48
3.3
NA1
0.02
6.2
Major PM
Components
Organic compounds,
elemental carbon
Organic compounds,
elemental carbon
Crustal elements,
trace metals
Metals, crustal material,
organic compounds
Organic compounds,
elemental carbon
Organic compounds,
trace metals
Crustal elements
Crustal elements
Organic compounds,
elemental carbon
Notes
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%).
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).
source is highly sporadic, occurs during periods of high winds, and the resulting emissions are
too highly uncertain to be included here. As can be seen from a comparison of entries in Tables
3-11 and 3-12, estimates of emissions of potential precursors to secondary PM formation are
considerably larger than those for estimates of primary PM25 emissions in the United States.
The emissions of SO2, NOX, and NH3 that are converted to PM should be multiplied by factors of
1.5, 1.35, and 1.07, respectively, to account for their chemical form in the aerosol phase.
Estimating a factor for VOCs is somewhat less straightforward. Turpin and Lim (2001)
recommend factors ranging from 1.4 to 2 to account for the conversion of OC to oxygen- and
3-94
-------
TABLE 3-12. EMISSIONS OF PRECURSORS TO SECONDARY PM2 5 FORMATION
BY VARIOUS SOURCES IN 1999
Precursor
Emissions Secondary PM
(109 kg/y) Component
Notes
SO,
Anthropogenic
VOCs
Biogenic
VOCs1
NH,
17 Sulfate
26 Nitrate
16 Various mainly
unidentified
compounds of 'OC'
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
(4%); chemical manufacturing (3%); waste disposal,
recycling, and other minor sources (4%); livestock (73%);
and fertilizer application (16%).
'Includes estimates of natural sources from Guenther et al. (2000).
2Emissions expressed in terms of NO2.
Source: Adapted from U.S. Environmental Protection Agency (2001).
nitrogen-containing compounds in the aerosol phase. There is some additional uncertainty in
assigning a factor for the fraction of VOCs that are converted into OC in the aerosol phase.
These factors are all greater than 1 and further underscore the potential importance of secondary
PM precursor emissions relative to primary PM emissions. However, the emissions of
precursors cannot be translated directly into rates of PM formation. Dry deposition and
precipitation scavenging of some of these gaseous precursors and their intermediate oxidation
products occur before they are converted to PM in the atmosphere, and most of the VOCs are
oxidized to CO2 rather than PM. In addition, some fraction of these gases are transported outside
of the domain of the continental United States before being oxidized. Likewise, emissions of
3-95
-------
these gases from areas outside the United States can result in the transport of their oxidation
products into the United States.
As discussed in Section 3.3.1, the photochemical oxidation of SO2 leads to the production
of sulfate, whereas that of NOX leads ultimately to particulate-phase nitrite and nitrate. Due to
uncertainties it is difficult to calculate the rates of formation of SOPM from the emissions of
VOC precursors. Smog chamber and laboratory studies discussed in Section 3.3.1 indicate that
anthropogenic aromatic compounds and biogenic terpenoid compounds have the highest
potential for forming SOPM; and, as can be seen from Table 3C-1, the dominant compounds
tend to be those derived from these categories. Each of the source categories capable of emitting
VOCs shown in Table 3-12 has components capable of forming SOPM, although in small yields
(ranging typically up to several percent, cf, Section 3.3.1). The oxidation of lighter organic
compounds leads ultimately to the formation of CO and CO2. As discussed by Pandis et al.
(1991) and in Section 3.3.1, soluble gas phase compounds, such as formaldehyde (CH2O), other
aldehydes, organic acids, etc., formed during the oxidation of a wide variety of hydrocarbons,
can be incorporated into suspended particles. Although isoprene is a major component of
biogenic emissions, its oxidation has not been found to result in the formation of new particles;
whereas the oxidation of monoterpenes has. However, it should be remembered that soluble gas
phase species such as CH2O are formed during the oxidation of isoprene.
The emissions estimates shown in this section are based on averaged annual totals.
However, annual averages do not reflect the variability of a number of emissions categories on
shorter time scales. Residential wood burning in fireplaces and stoves, for example, is a
seasonal practice that reaches its peak during cold weather. Cold weather also affects motor
vehicle exhaust PM emissions, both in terms of chemical composition and emission rates (e.g.,
Watson et al., 1990b; Huang et al., 1994). Agricultural activities such as planting, fertilizing,
and harvesting are also seasonal. Forest fires occur mainly during the local dry season and
during periods of drought. Maximum dust production by wind erosion in the United States
occurs during the spring, whereas the minimum occurs during the summer (Gillette and Hanson,
1989). Efforts are being made to account for the seasonal variations of emissions in the
nationwide emissions inventories. Techniques for calculating emissions of criteria pollutants on
a seasonal basis are given in U.S. Environmental Protection Agency (1999).
3-96
-------
Trends in nationwide, annual average concentrations of PM10 and precursor gases (SO2,
NO2, and VOC) during the 10 years from 1992 to 2001 are shown in Table 3-13. As can be seen
from Table 3-13, there have been substantial decreases in the ambient concentrations of PM10,
SO2, and NO2. Not enough data are available to define trends in VOC concentrations. There
also have been substantial decreases in the emissions of all the species shown in Table 3-13,
except for NO2, although its average ambient concentration has decreased by 11%. These entries
suggest that decreases in the average ambient concentration of PM10 could have been produced
by both decreases in emissions of primary PM10 and in the formation of secondary PM10. The
large reductions in ambient SO2 concentrations have resulted in reductions in sulfate formation
that would have been manifest in PM2 5 concentrations on the regional scale in the East and
Midwest, where sulfate has constituted a larger fraction of PM25 than in the West. Likewise,
reductions in NO2 concentrations would have had a more noticeable effect on PM2 5
concentrations in the West than in the East, because nitrate is a larger component of the aerosol
in the West.
TABLE 3-13. NATIONWIDE CHANGES IN AMBIENT CONCENTRATIONS AND
S OF PM10 AND GASEOUS PRECURSORS TO SE
PARTICULATE MATTER FROM 1992 TO 2001
EMISSIONS OF PM10 AND GASEOUS PRECURSORS TO SECONDARY
PM10
PM25
SO2
NOX
VOC
% Change 1992-2001
Ambient Concentration Emissions
-14% -13%*
— -10%*
-35%(SO2) -24%(SO2)
-11%(NO2) -3%(NOX)
— -8%
* Includes only primary PM.
Source: U.S. Environmental Protection Agency (2002a).
3-97
-------
Trends in aerosol components (i.e., nitrate, sulfate, carbon, etc.) are needed for a more
quantitative assessment of the effects of changes in emissions of precursors. Aerosol nitrate and
sulfate concentrations obtained at North Long Beach and Riverside, CA tracked downward
trends in NOX concentrations. SO2 and sulfate concentrations have both decreased; however, the
rate of decline of sulfate has been smaller than that of SO2, indicating the long range transport of
sulfate from outside the airshed may be an important source in addition to the oxidation of
locally generated SO2. There are a number of reasons why pollutant concentrations do not track
well estimated reductions in emissions. Some of these reasons are related to atmospheric effects
such as meteorological variability and secular changes in the rates of photochemical
transformations and deposition (U.S. Environmental Protection Agency, 2000c). Other reasons
are related to uncertainties in ambient measurements and in emissions inventories.
3.3.5 Uncertainties of Emissions Inventories
As described in the 1996 PM AQCD, it is difficult to quantitatively assign uncertainties to
entries in emissions inventories. Methods that can be used to verify or place constraints on
emissions inventories are sparse. In general, the overall uncertainty in the emissions of a given
pollutant includes contributions from all of the terms used to calculate emissions (i.e., activity
rates, emissions factors, and control device efficiencies). Additional uncertainties may arise
during the compilation of an emissions inventory due to missing sources and computational
errors. The variability of emissions can cause errors when annual average emissions are applied
to applications involving shorter time scales.
Activity rates for well-defined point sources (e.g., power plants) should have the smallest
uncertainty associated with their use, because emissions are monitored continuously in many
cases and accurate production records need to be kept. On the other hand, activity rates for a
number of very disperse fugitive sources are difficult to quantify. Emissions factors for easily
measured fuel components that are released quantitatively during combustion (e.g., CO2, SO2)
should be the most reliable. Emissions of components formed during combustion are more
difficult to characterize, as the emissions rates are dependent on factors specific to individual
combustion units and on combustion stage (i.e., smoldering or active). Although the AP-42
emissions factors (U.S. Environmental Protection Agency, 1995) contain extensive information
for a large number of source types, these data are very limited in the number of sources sampled.
3-98
-------
The efficiency of control devices is determined by their design, age, maintenance history, and
operating conditions. It is virtually impossible to assign uncertainties in control device
performance because of these factors. It should be noted that the largest uncertainties occur for
those devices that have the highest efficiencies (> 90%). This occurs because the efficiencies are
subtracted from one, and small errors in assigning efficiencies can lead to large errors in
emissions.
Ideally, an emissions inventory should include all major sources of a given pollutant. This
may be an easy task for major point sources. However, area sources of both primary PM and
precursors to secondary PM formation are more difficult to characterize than point sources; and,
thus, they require special emphasis when preparing emission inventories. Further research is
needed to better characterize the sources of pollutants to reduce this source of uncertainty.
Errors also can arise from the misreporting of data, and arithmetic errors can occur in the course
of compiling entries from thousands of individual sources. A quality assurance program is
required to check for outliers and arithmetic errors. Because of the variability in emissions rates,
there can be errors in the application of inventories developed on an annually averaged basis (as
are the inventories shown in Tables 3-11 and 3-12) to episodes occurring on much shorter time
scales. As an example, most modeling studies of air pollution episodes are carried out for
periods of a few days.
Uncertainties in annual emissions were estimated to range from 4 to 9% for SO2 and from
6 to 11% for NOX in the 1985 NAPAP inventories for the United States (Placet et al., 1991).
Uncertainties in these estimates increase as the emissions are disaggregated both spatially and
temporally. The uncertainties quoted above are minimum estimates and refer only to random
variability about the mean, assuming that the variability in emissions factors was adequately
characterized and that extrapolation of emissions factors to sources other than those for which
they were measured is valid. The estimates do not consider the effects of weather or variations
in operating and maintenance procedures.
Fugitive dust sources, as mentioned above, are extremely difficult to quantify; and stated
emission rates may represent only order-of-magnitude estimates. Although crustal dust
emissions constitute about 50% of the total primary PM25 inventory, they constitute less than
about 15% of the source strengths inferred from the receptor modeling studies shown in
3-99
-------
Table 3-9. However, it should be remembered that secondary components (sulfate, nitrate, and
some fractions of OC) often account for most of the mass of ambient PM25 samples.
Although mineral dust sources represent the major category in Table 3-11, their
contributions are distributed much more widely than are those from combustion sources.
Watson and Chow (2000) reexamined the methodology used to determine emissions of fugitive
dust. The standard methods use data obtained by particle monitors stacked at several elevations
from 1 to 2 m up to 7 to 10 m above the surface. However, small-scale turbulent motions and
variable winds characterize atmospheric flow patterns immediately adjacent to the surface in this
height range (Garratt, 1994). The depth of this turbulent layer is determined by surface
roughness elements, and there is a high probability of particles being entrained in turbulent
eddies and redepositing on the ground within a very short distance. In addition to the source-
sampling problem referred to above, it should be remembered that dust often is raised in remote
areas far removed from population centers. Precipitation or scavenging by cloud droplets and
dry deposition removes particles during transport from the source area. In addition, gravitational
settling can be an important loss mechanism for particles larger than a few micrometers in
aerodynamic diameter.
As rough estimates, uncertainties in emissions estimates could be as low as 10% for the
best characterized source categories; whereas emissions figures for windblown dust should be
regarded as order-of-magnitude estimates. The application of emissions inventories to the
estimation of source contributions at monitoring sites is also limited by the effects of local
topography and meteorology. For example, Pinto et al. (1998) found that the contribution of
power plants and residential space heating to PM2 5 concentrations in northwestern Bohemia are
comparable on the basis of CMB receptor modeling. However, according to the emissions
inventories, the contribution from power plants should have been roughly an order of magnitude
larger than that from residential space heating. The difference between the two methods can be
explained by noting that mixing of the emissions from the power plants downward to the surface
is inhibited by strong surface inversions that develop during the winter season in this area.
There have been few field studies designed to test emissions inventories observationally.
The most direct approach would be to use aircraft to obtain cross sections of pollutants upwind
and downwind of major urban areas. The computed mass flux through a cross section of the
urban plume can then be equated to emissions from the chosen city. This approach has been
3-100
-------
attempted on a few occasions, but results have been ambiguous because of contributions from
fugitive sources, variable wind flows, and logistic difficulties.
3.4 SUMMARY AND KEY CONCLUSIONS
The recently deployed PM2 5 FRM network has returned data for a large number of sites
across the United States. Annual mean PM2 5 concentrations range from ~5 |ig/m3 to -30 |ig/m3.
In the eastern United States, the data from 1999 to 2001 indicate that highest quarterly mean
concentrations and maximum concentrations most often occur during the summer. In the
western United States, highest quarterly mean values and maximum values occur mainly during
the winter at a number of sites, although there were exceptions to these general patterns. Sites
affected strongly by sources of primary PM are expected to show winter maxima. Sources
emitting precursors of secondary ammonium nitrate are also contributors to winter PM maxima.
These findings are generally consistent with those based on longer term data sets such as these
form the Metropolitan Acid Aerosol Characterization Study (MAAQS) in the eastern United
States and the California Air Resources Board (CARB) network of dichotomous samplers in
California. PM25 and PM10 concentrations in a number of urban areas have generally declined
over the past few decades, and they appear to have leveled off in the past few years.
Differences in annual mean PM2 5 concentrations between monitoring sites in urban areas
examined are typically less than 6 or 7 |ig/m3. However, on individual days, differences in 24-h
average PM2 5 concentrations can be much larger. Some sites in metropolitan areas are highly
correlated with each other but not with others, due to the presence of local sources, topographic
barriers, etc. Although PM2 5 concentrations at sites within a MSA can be highly correlated,
significant differences in their concentrations can occur on any given day. Consequently,
additional measures should be used to characterize the spatial variability of PM25 concentrations.
The degree of spatial uniformity in PM2 5 concentrations in urban areas varies across the country.
These factors should be considered in using data obtained by the PM2 5 FRM network to estimate
community-scale human exposure, and caution should be exercised in extrapolating conclusions
obtained in one urban area to another. PM2 5 to PM10 ratios were generally higher in the East
than in the West, and values for this ratio are consistent with those found in numerous earlier
studies presented in the 1996 PM AQCD.
3-101
-------
Data for PM10_2 5 are not as abundant as they are for PM2 5, and their interpretation is
complicated by the difference method used to determine their concentrations. The more
sporadic nature of sources of PM10_2 5 and its shorter atmospheric lifetime tend to result in lower
spatial correlations for PM10_2 5 than for PM2 5 concentrations. Errors in the measurement
of PM25 and PM10 also result in calculations of lower spatial correlations of PM10_25. Calculated
concentrations of PM10_25 are occasionally negative as reflected by PM25 to PM10 ratios greater
than one. Because analytical errors are generally larger for individual species than for total
mass, similar problems arise in their determination in PM10_2 5 samples by the difference
approach. Some, but not all of these problems could be resolved by the use of dichotomous
samplers that also provide a direct sample of PM10_25 for compositional analyses.
Estimates of concentrations of individual species in PM10_25 samples used in this chapter
were limited to those obtained by dichotomous samplers. Generally, concentrations of most
elements differ for PM2 5 and PM10_2 5. However, the available data suggest that concentrations of
many metals are of the same order of magnitude in both size fractions. This is in marked
contrast to the situation 20 years ago, when uncontrolled combustion sources were prevalent.
At that time, concentrations of many metals, especially Pb, were much higher than today in
fine-mode particles, and their concentrations were much higher in the fine mode than in the
coarse mode. No substantive conclusions about contemporary concentrations and composition
of ultrafine particles (< 0.1 jim Da) can be drawn for the nation as a whole because of a lack of
data.
Ambient PM contains both primary and secondary components. The results of ambient
monitoring studies and receptor modeling studies indicate that PM2 5 is dominated by secondary
components in the eastern United States. General statements about the origin of organic carbon
(OC) in ambient PM2 5 samples cannot yet be made; and, so, the contribution of secondary
components throughout the rest of the United States is still highly uncertain. Primary
constituents represent smaller but still important components of PM25. Crustal materials, which
are primary constituents, constitute the largest measured fraction of PM10_2 5 throughout the
United States. Data for the concentration of bioaerosols in both the PM25 and PM10_25 size
ranges are sparse. Data collected in several airsheds, including the Los Angeles Basin,
Bakersfield and Fresno, CA and Philadelphia, PA airsheds, suggest that secondary PM
3-102
-------
components are more uniformly distributed than are primary PM components. Compositional
data obtained at multiple sites in other urban areas are sparse.
Because of the complexity of the composition of ambient PM25 and PM10_25, sources
are best discussed in terms of individual constituents of both primary and secondary PM25
and PM10_25. Each of these constituents can have anthropogenic and natural sources as shown
in Table 3-8. The distinction between natural and anthropogenic sources is not always obvious.
Although windblown dust might seem to be the result of natural processes, highest emission
rates are associated with agricultural activities in areas that are susceptible to periodic drought.
Examples include the dust bowl region of the midwestern United States and the Sahel of Africa.
There is also ongoing debate about characterizing wildfires as either natural or anthropogenic.
Land management practices and other human actions affect the occurrence and scope of
wildfires. Similarly, prescribed burning can be viewed as anthropogenic or as a substitute for
wildfires that would otherwise eventually occur on the same land.
During the past decade, a significant amount of research has been carried out to improve
our understanding of the atmospheric chemistry of secondary organic PM (SOPM) formation.
Although additional sources of SOPM might still be identified, there appears to be a general
consensus that biogenic compounds (monoterpenes, sesquiterpenes) and aromatic compounds
(toluene, ethylbenzene) are the most significant SOPM precursors. A large number of
compounds have been detected in biogenic and aromatic SOPM, although the chemical
composition of these two categories has not been fully established, especially for aromatic
SOPM. Transformations that occur during the aging of particles are still inadequately
understood. There still exist large gaps in our current understanding of a number of key
processes related to the partitioning of semivolatile compounds between the gas phase and
ambient particles containing organic compounds, liquid water, and inorganic salts and acids.
In addition, there is a general lack of reliable analytical methods for measuring multifunctional
oxygenated compounds in the gas and aerosol phases.
The results of receptor modeling studies throughout the United States indicate that the
combustion of fossil and biomass fuels is the major source of measured ambient PM2 5. Fugitive
dust, found mainly in the PM10_25 range size, represents the largest source of measured ambient
PM10 in many locations in the western United States. Quoted uncertainties in the source
apportionment of constituents in ambient aerosol samples typically range from 10 to 50%. It is
3-103
-------
apparent that a relatively small number of source categories, compared to the total number of
chemical species that typically are measured in ambient monitoring-source receptor model
studies, are needed to account for the majority of the observed mass of PM in these studies.
The application of any of the source apportionment techniques is still limited by the
availability of source profile data as well as by the level of detail and the quality of ambient
measurement data. Whereas the chemical mass balance (CMB) approach relies directly on
source profile data, solutions from the positive matrix factorization (PMF) technique yield
profiles for the factors that contribute to PM. However, there is some rotational ambiguity
present in the solutions. Source profile data obtained by PMF must still be verified by
comparison with observational data and these data can be used in techniques such as PMF to
improve the solutions. Serious limitations still exist with regard to source profiles for organic
compounds. The complexity of reactions involving organic compounds in particles adds to the
difficulties of finding stable species that could be used as tracers.
As seen in Table 3-8, emissions of mineral dust, organic debris, and sea spray are
concentrated mainly in the coarse fraction of PM10 (> 2.5 jim Da). A small fraction of this
material is in the PM25 size range (< 2.5 jim Da). Still, PM25 concentrations of crustal material
can be appreciable, especially during dust events. It also should be remembered that from one-
third to one-half of the Saharan dust reaching the United States is in the PM25 size range.
Emissions from combustion sources (mobile and stationary sources and biomass burning) are
also predominantly in the PM2 5 size range.
A number of sources contribute to policy relevant background (PRB) concentrations. Data
obtained at relatively remote monitoring sites (RRMS) in the western United States could be
used to place reasonable upper limits on PRB concentrations. More definitive results for both
annual average and daily average concentrations could potentially be obtained from the
application of source-receptor models and/or the application of large-scale chemistry transport
models. Many areas in the East are affected by dust transported from northern Africa, and it has
recently become apparent that many areas, especially, but not limited to the Northwest, are
affected by dust transported from the deserts of Asia. In addition to crustal material, pollutants
and primary biological aerosol particles (PBPs) are also transported during intercontinental
transport events. Many areas are also affected by smoke from wildfires occurring within the
United States or in Canada, Mexico, or Central America. Storms, in which the winds can
3-104
-------
suspend material from the surface of the land or seas, also contribute soil, sea spray and PBP.
Contributions of primary PM from natural sources and sources outside Northern America as
given above are all episodic. Because concentrations of PBP are so poorly quantified, even
though they can constitute significant portions of the organic fraction of the atmospheric aerosol,
estimates of PRB concentrations will remain highly uncertain. Estimates of annually averaged
PRB concentrations or their range have not changed from the 1996 PM AQCD.
3-105
-------
REFERENCES
Altshuller, A. P. (1983) Review: natural volatile organic substances and their effect on air quality in the
United States. Atmos. Environ. 17: 2131-2165.
Ansari, A. S.; Pandis, S. N. (1998) Response of inorganic PM to precursor concentrations. Environ. Sci. Technol.
32: 2706-2714.
Atkinson, R. (1994) Gas-phase tropospheric chemistry of organic compounds. Washington, DC: American Chemical
Society. (J. Phys. Chem. Ref. Data Monograph no. 2).
Aumont, B.; Madronich, S.; Bey, I.; Tyndall, G. S. (2000) Contribution of secondary VOC to the composition of
aqueous atmospheric particles: a modeling approach. J. Atmos. Chem. 35: 59-75.
Bahadori, T.; Suh, H.; Koutrakis, P. (2000a) A cross-study analysis of ambient concentrations of fine mass. J. Air
Waste Manage. Assoc.: in press.
Bahadori, T.; Suh, H.; Koutrakis, P. (2000b) PM2 5 and PM10 concentrations in four metropolitan cities in eastern
United States. J. Air Waste Manage. Assoc.: in press.
Blando, J. D.; Porcja, R. J.; Li, T.-H.; Bowman, D.; Lioy, P. J.; Turpin, B. J. (1998) Secondary formation and the
Smoky Mountain organic aerosol: an examination of aerosol polarity and functional group composition
during SEAVS Environ. Sci. Technol. 32: 604-613.
Cabada, J. C.; Pandis Spyros, N.; Robinson, A. L. (2002) Sources of atmospheric carbonaceous paniculate matter in
Pittsburgh, Pennsylvania. J. Air Waste Manage. Assoc. 52: 732-741.
Cahill, T. A. (1996) [Data from IMPROVE data base.] Fort Collins, CO: Air Resource Specialists, Inc.
California Air Resources Board. (1995) [Database provided by JohnR. Holmes, Chief, Research Division.]
Sacramento, CA: California Air Resources Board.
Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some
reflections on controlling vehicle emissions. Science (Washington, DC) 261: 37-45.
Cass, G. R. ; Hughes, L. A.; Bhave, P.; Kleeman, M. J.; Allen, J. O.; Salmon, L. G. (2000) The chemical
composition of atmospheric ultrafine particles. Philos. Trans. R. Soc. London Ser. A 358: 2581-2592.
Castro, L. M.; Pio, C. A.; Harrison, R. M.; Smith, D. J. T. (1999) Carbonaceous aerosol in urban and rural European
atmospheres: estimation of secondary organic carbon concentrations. Atmos. Environ. 33: 2771-2781.
Chadwick, O. A.; Deny, L. A.; Vitousek, P. M.; Huebert, B. J.; Hedin, L. O. (1999) Changing sources of nutrients
during four million years of ecosystem development. Nature (London, UK) 397: 491-497.
Chow, J. C.; Watson, J. G. (2002) Review of PM2 5 and PM10 apportionment for fossil fuel combustion and other
sources by the chemical mass balance receptor model. Energy Fuelsl6: 222-260.
Chow, J. C.; Watson, J. G.; Frazier, C. A.; Egami, R. T.; Goodrich, A.; Ralph, C. (1988) Spatial and temporal source
contributions to PM10 andPM25 in Reno, NV. In: Mathai, C. V.; Stonefield, D. H., eds. PM-10:
implementation of standards, an APCA/EPA international specialty conference; February; San Francisco, CA.
Pittsburgh, PA: Air Pollution Control Association; pp. 438-457. (APCA transactions series TR-13).
Chow, J. C.; Watson, J. G.; Richards, L. W.; Haase, D. L.; McDade, C.; Dietrich, D. L.; Moon, D.; Sloane, C. (1991)
The 1989-90 Phoenix PM10 study. Volume II: source apportionment. Final report. Phoenix, AZ: Arizona
Department of Environmental Quality; Desert Research Institute document no. 8931.6F1.
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Frazier, C. A.; Hinsvark, B. A.; Prichett, L. C.; Neuroth, G. R.
(1992a) Wintertime PM10 and PM2 5 chemical compositions and source contributions in Tucson, Arizona.
In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source controls, an
A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste
Management Association; pp. 231-243. (A&WMA transactions series no. 22).
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W.
(1992b) PM10 source apportionment in California's San Joaquin Valley. Atmos. Environ. Part A
26: 3335-3354.
Chow, J. C.; Liu, C. S.; Cassmassi, J.; Watson, J. G.; Lu, Z.; Pritchett, L. C. (1992c) A neighborhood-scale study of
PM10 source contributions inRubidoux, California. Atmos. Environ. Part A 26: 693-706.
Chow, J. C.; Watson, J. G.; Edgerton, S. A.; Vega, E.; Ortiz, E. (2002) Spatial differences in outdoor PM10 mass and
aerosol composition in Mexico City. J. Air Waste Manage. Assoc. 52: 423-434.
Christoffersen, T. S.; Hjorth, J.; Horie, O.; Jensen, N. R.; Kotzias, D.; Molander, L. L.; Neeb, P.; Ruppert, L.;
Winterhalter, R.; Virkkula, A.; Wirtz, K.; Larsen, B. R. (1998) cis-Pinic acid, a possible precursor for organic
aerosol formation from ozonolysis of a-pinene. Atmos. Environ. 32: 1657-1661.
Chung, A.; Herner, J. D.; Kleeman, M. J. (2001) Detection of alkaline ultrafine atmospheric particles at Bakersfield,
California. Environ. Sci. Technol. 35: 2184-2190.
3-106
-------
Cocker, D. R., Ill; Clegg, S. L.; Flagan, R. C.; Seinfeld, J. H. (2001) The effect of water on gas-particle partitioning
of secondary organic aerosol. Part I: a-pinene/ozone system. Atmos. Environ. 35: 6049-6072.
Darlington, T. L.; Kahlbaum, D. F.; Heuss, J. M; Wolff, G. T. (1997) Analysis of PM10 trends in the United States
from 1988 through 1995. J. Air Waste Manage. Assoc. 47: 1070-1078.
Davis, A. P.; Shokouhian, M.; Ni, S. (2001) Loading estimates of lead, copper, cadmium, and zinc in urban runoff
from specific sources. Chemosphere 44: 997-1009.
Dolislager, L. J.; Motallebi, N. (1999) Characterization of paniculate matter in California. J. Air Waste Manage.
Assoc. 49: PM-45-56.
Duce, R. A. (1995) Sources, distributions, and fluxes of mineral aerosols and their relationship to climate.
In: Charlson, R. J.; Heintzenberg, J., eds. Aerosol forcing of climate: report of the Dahlem workshop on
aerosol forcing of climate; April 1994; Berlin, Federal Republic of Germany. Chichester, United Kingdom:
John Wiley & Sons, Ltd.; pp. 43-72.
Dzubay, T. G.; Stevens, R. K.; Gordon, G. E.; Olmez, I.; Sheffield, A. E.; Courtney, W. J. (1988) A composite
receptor method applied to Philadelphia aerosol. Environ. Sci. Technol. 22: 46-52.
Edney, E. O.; Driscoll, D. J.; Weathers, W. S.; Kleindienst, T. E.; Conver, T. S.; Mclver, C. D.; Li, W. (2001)
Formation of polyketones in irradiated toluene/propylene/NOx/air mixtures. Aerosol Sci. Technol.
35: 998-1008.
Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the
United States of America. Atmos. Environ. 28: 1009-1019.
Faust, B. C. (1994) Photochemistry of clouds, fogs, and aerosols. Environ. Sci. Technol. 28: 217A-222A.
Fitz-Simons, T. S.; Mathias, S.; Rizzo, M. (2000) Analyses of 1999 PM data for the PMNAAQS review. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
November 17. Available: http://www.epa.gov/oar/oaqps/pm25/analyses.html [2 April, 2002].
Forstner, H. J. L.; Flagan, R. C.; Seinfeld, J. H. (1997a) Molecular speciation of secondary organic aerosol from
photooxidation of the higher alkenes: 1-octene and 1-decene. Atmos. Environ. 31: 1953-1964.
Forstner, H. J. L.; Flagan, R. C.; Seinfeld, J. H. (1997b) Secondary organic aerosol from the photooxidation of
aromatic hydrocarbons: molecular composition. Environ. Sci. Technol. 31: 1345-1358.
Garratt, J. R. (1994) The atmospheric boundary layer. Cambridge, United Kingdom: Cambridge University Press.
(Cambridge atmospheric and space science series).
Gatz, D. F.; Prospero, J. M. (1996) A large silicon-aluminum aerosol plume in central Illinois: North African Desert
dust? Atmos. Environ. 30: 3789-3799.
Geron, C.; Rasmussen, R.; Arnts, R. R.; Guenther, A. (2000) A review and synthesis of monoterpene speciation
from forests in the United States. Atmos. Environ. 34: 1761-1781.
Geron, C. D.; Harley, P.; Guenther, A. (2001) Isoprene emission capacity for US tree species. Atmos. Environ.
35: 3341-3352.
Gillette, D. A.; Hanson, K. J. (1989) Spatial and temporal variability of dust production caused by wind erosion in
the United States. J. Geophys. Res. [Atmos.] 94: 2197-2206.
Gillies, J. A.; Gertler, A. W.; Lowenthal, D.; Abu-Allaban, M.; Jennison, B.; Goodrich, A. (2000) Enhanced PM25
source apportionment using chemical mass balance receptor modeling and scanning electron microscopy.
Aerosol Sci. Technol.: submitted.
Glasius, M.; Lahaniati, M.; Calogirou, A.; Di Bella, D.; Jensen, N. R.; Hjorth, J.; Kotzias, D.; Larsen, B. R. (2000)
Carboxylic acids in secondary aerosols from oxidation of cyclic monoterpenes by ozone. Environ. Sci.
Technol. 34: 1001-1010.
Glover, D. M.; Hopke, P. K.; Vermette, S. J.; Landsberger, S.; D'Auben, D. R. (1991) Source apportionment with
site specific source profiles. J. Air Waste Manage. Assoc. 41: 294-305.
Gone, J.-K.; Olmez, I.; Ames, M. R. (2000) Size distribution and probable sources of trace elements in submicron
atmospheric paniculate material. J. Radioanal. Nucl. Chem. 244: 133-139.
Graedel, T. E.; Goldberg, K. I. (1983) Kinetic studies of raindrop chemistry: 1. Inorganic and organic processes.
J. Geophys. Res. C: Oceans Atmos. 88: 10,865-10,882.
Gray, H. A.; Landry, B.; Liu, C. S.; Henry, R. C.; Cooper, J. A.; Sherman, J. R. (1988) Receptor modeling for PM10
source apportionment in the South Coast Air Basin of California. In: Mathai, C. V.; Stonefield, D. H., eds.
PM.10: implementation of standards, an APCA/EPA international specialty conference; February;
San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 399-418. (APCA transactions
series TR-13).
Griffin, R. J.; Cocker, D. R., Ill; Seinfeld, J. H. (1999) Incremental aerosol reactivity: application to aromatic and
biogenic hydrocarbons. Environ. Sci. Technol. 33: 2403-2408.
3-107
-------
Griffin, D. W.; Garrison, V. H.; Herman, J. R.; Shinn, E. A. (2001) African desert dust in the Caribbean atmosphere:
microbiology and public health. Aerobiologia 17: 203-213.
Grosjean, D. (1992) In situ organic aerosol formation during a smog episode: estimated production and chemical
functionality. Atmos. Environ. Part A 26: 953-963.
Grosjean, D.; Seinfeld, J. H. (1989) Parameterization of the formation potential of secondary organic aerosols.
Atmos. Environ. 23: 1733-1747.
Guenther, A.; Zimmerman, P.; Klinger, L.; Greenberg, J.; Ennis, C.; Davis, K.; Pollock, W.; Westberg, H.;
Allwine, G.; Geron, C. (1996) Estimates of regional natural volatile organic compound fluxes from enclosure
and ambient measurements. J. Geophys. Res. 101: 1345-1360.
Guenther, A.; Geron, C.; Pierce, T.; Lamb, B.; Harley, P.; Fall, R. (2000) Natural emissions of non-methane volatile
organic compounds, carbon monoxide, and oxides of nitrogen from North America. Atmos. Environ.
34: 2205-2230.
Haagen-Smit, A. J. (1952) Chemistry and physiology of Los Angeles smog. Ind. Eng. Chem. 44: 1342-1346.
Helmig, D.; Greenberg, J.; Guenther, A.; Zimmerman, P.; Geron, C. (1998) Volatile organic compounds and
isoprene oxidation products at a temperate deciduous forest site. J. Geophys. Res. 103(D17): 22,397-22,414.
Henry, R. C. (1997) History and fundamentals of multivariate air quality receptor models. Chemom. Intell. Lab.
Syst. 37: 37-42.
Hewitt, C. N. (2001) The atmospheric chemistry of sulphur and nitrogen in power station plumes. Atmos. Environ.
35: 1155-1170.
Hidy, G. M. (1994) Atmospheric sulfur and nitrogen oxides: eastern North American source-receptor relationships.
San Diego, CA: Academic Press, Inc.
Hildemann, L. M.; Rogge, W. F.; Cass, G. R.; Mazurek, M. A.; Simoneit, B. R. T. (1996) Contribution of primary
aerosol emissions from vegetation-derived sources to fine particle concentrations in Los Angeles. J. Geophys.
Res. [Atmos.] 101: 19,541-19,549.
Hoffmann, T.; Odum, J. R.; Bowman, F.; Collins, D.; Klockow, D.; Flagan, R. C.; Seinfeld, J. H. (1997) Formation
of organic aerosols from the oxidation of biogenic hydrocarbons. J. Atmos. Chem. 26: 189-222.
Hopke, P. K.; Wlaschin, W.; Landsberger, S.; Sweet, C.; Vermette, S. J. (1988) The source apportionment of PM10
in south Chicago. In: Mathai, C. V.; Stonefield, D. H., eds. PM-10: implementation of standards, an
APCA/EPA international specialty conference; February; San Francisco, CA. Pittsburgh, PA: Air Pollution
Control Association; pp. 484-494. (APCA transactions series TR-13).
Houck, J. E.; Rau, J. A.; Body, S.; Chow, J. C. (1992) Source apportionment: Pocatello, Idaho PM10 nonattainment
area. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source controls, an
A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste
Management Association; pp. 219-230. (A&WMA transactions series no. 22).
Huang, X.; Olmez, L; Aras, N. K.; Gordon, G. E. (1994) Emissions of trace elements from motor vehicles: potential
marker elements and source composition profile. Atmos. Environ. 28: 1385-1391.
Husar, R. B.; Tratt, D. M.; Schichtel, B. A.; Falke, S. R.; Li, F.; Jaffe, D.; Gasso, S.; Gill, T.; Laulainen, N. S.;
Lu, F.; Reheis, M. C.; Chun, Y.; Westphal, D.; Holben, B. N.; Gueymard, C.; McKendry, I.; Kuring, N.;
Feldman, G. C.; McClain, C.; Frouin, R. J.; Merrill, J.; DuBois, D.; Vignola, F.; Murayama, T.; Nickovic, S.;
Wilson, W. E.; Sassen, K.; Sugimoto, N.; Malm, W. C. (2001) Asian dust events of April 1998. J. Geophys.
Res. (Atmos.) 106: 18,317-18,330.
Inhalable Paniculate Network. (1985) [Inhalable Paniculate Network Database.] Research Triangle Park, NC:
U.S. Environmental Protection Agency, National Exposure Research Laboratory; Jose M. Sune, contact.
Izumi, K.; Fukuyama, T. (1990) Photochemical aerosol formation from aromatic hydrocarbons in the presence of
NOX. Atmos. Environ. Part A 24: 1433-1441.
Jang, M.; Kamens, R. M. (1999) Newly characterized products and composition of secondary aerosols from the
reaction of a-pinene with ozone. Atmos. Environ. 33: 459-474.
Jang, M.; Kamens, R. M. (200la) Atmospheric secondary aerosol formation by heterogeneous reactions of
aldehydes in the presence of a sulfuric acid aerosol catalyst. Environ. Sci. Technol. 35: 4758-4766.
Jang, M.; Kamens, R. M. (200 Ib) Characterization of secondary aerosol from the photooxidation of toluene in the
presence of NOxand 1-propene. Environ. Sci. Technol. 35: 3626-3639.
Kamens, R.; Jang, M.; Chien, C.-J.; Leach, K. (1999) Aerosol formation from the reaction of a-pinene and ozone
using a gas-phase kinetics-aerosol partitioning model. Environ. Sci. Technol. 33: 1430-1438.
Kao, A. S.; Friedlander, S. K. (1995) Frequency distributions of PM10 chemical components and their sources.
Environ. Sci. Technol. 29: 19-28.
Kavouras, I. G.; Mihalopoulos, N.; Stephanou, E. G. (1998) Formation of atmospheric particles from organic acids
produced by forests. Nature (London) 395: 683-686.
3-108
-------
Kavouras, I. G.; Mihalopoulus, N.; Stephanou, E. G. (1999a) Formation and gas/particle partitioning of monoterpene
photooxidation products over forests. Geophys. Res. Lett. 26: 55-58.
Kavouras, I. G.; Mihalopoulus, N.; Stephanou, E. G. (1999b) Secondary organic aerosol formation vs primary
organic aerosol emission: in situ evidence for the chemical coupling between monoterpene acidic
photooxidation products and new particle formation over forests. Environ. Sci. Technol. 33: 1028-1037.
Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
segregated mass concentration data and ultrafine particle number concentrations in urban areas. Atmos.
Environ. 33:2907-2913.
Khalil, M. A. K.; Rasmussen, R. A. (1992) Forest hydrocarbon emissions: relationships between fluxes and ambient
concentrations. J. Air Waste Manage. Assoc. 42: 810-813.
Kim, B. M.; Zeldin, M. D.; Liu, C. S. (1992) Source apportionment study for state implementation plan development
in the Coachella Valley. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate
source controls: an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ.
Pittsburgh, PA: Air & Waste Management Association; pp. 979-991. (A&WMA transactions series no. 22).
Kim, B. M.; Teffera, S.; Zeldin, M. D. (2000) Characterization of PM2 s and PM10 in the South Coast Air Basin of
southern California: part 2-temporal variations. J. Air Waste Manage. Assoc. 50: 2045-2059.
Kinney, P. L.; Aggarwal, M.; Northridge, M. E.; Janssen, N. A. H.; Shepard, P. (2000) Airborne concentrations of
PM2 5 and diesel exhaust particles on Harlem sidewalks: a community-based pilot study. Environ. Health
Perspect. 108: 213-218.
Koch, S.; Winterhalter, R.; Uherek, E.; Kolloff, A.; Neeb, P.; Moortgat, G. K. (2000) Formation of new particles in
the gas-phase ozonolysis of monoterpenes. Atmos. Environ. 34: 4031-4042.
Koutrakis, P. (1995) [PM25 andPM10 data for Philadelphia, 1992-1995]. Boston, MA: Harvard School of Public
Health.
Landis, M. S.; Norris, G. A.; Williams, R. W.; Weinstein, J. P. (2001) Personal exposures to PM2 5 mass and trace
elements in Baltimore, MD, USA. Atmos. Environ. 35: 6511-6524.
Larsen, B. R.; Di Bella, D.; Glasius, M.; Winterhalter, R.; Jensen, N. R.; Hjorth, J. (2001) Gas-phase OH oxidation
of monoterpenes: gaseous and paniculate products. J. Atmos. Chem. 38: 231-276.
Leach, K. B.; Kamens, R. M.; Strommen, M. R.; Jang, M. (1999) Partitioning of semivolatile organic compounds in
the presence of a secondary organic aerosol in a controlled atmosphere. J. Atmos. Chem. 33: 241-264.
Leaderer, B. P.; Naeher, L.; Jankun, T.; Balenger, K.; Holford, T. R.; Toth, C.; Sullivan, J.; Wolfson, J. M.;
Koutrakis, P. (1999) Indoor, outdoor, and regional summer and winter concentrations of PM10, PM2 5, SO42",
H+, NH4+, NO3", NH3, and nitrous acid in homes with and without kerosene space heaters. Environ. Health
Perspect. 107:223-231.
Lin, J. J.; Noll, K. E.; Holsen, T. M. (1994) Dry deposition velocities as a function of particle size in the ambient
atmosphere. Aerosol Sci. Technol. 20: 239-252.
Lipfert, F. W. (1998) Trends in airborne paniculate matter in the United States. Appl. Occup. Environ. Hyg.
13: 370-384.
Magliano, K. L.; Ranzieri, A.; Solomon, P.; Watson, J. (1998) Chemical mass balance modeling of data from the
1995 integrated monitoring study. Sacramento, CA: California Air Resources Board; California Regional
PM10/PM2 5 Air Quality Study.
Matthias-Maser, S. (1998) Primary biological aerosol particles: their significance, sources, sampling methods and
size distribution in the atmosphere. In: Harrison, R. M.; Van Grieken, R., eds. Atmospheric Particles.
New York, NY: John Wiley & Sons; pp. 349-368. (IUPAC series on Analytical and Physical Chemistry of
Environmental Systems, v. 5).
Mazurek, M.; Masonjones, M. C.; Masonjones, H. D.; Salmon, L. G.; Cass, G. R.; Hallock, K. A.; Leach, M. (1997)
Visibility-reducing organic aerosols in the vicinity of Grand Canyon National Park: properties observed by
high resolution gas chromatography. J. Geophys. Res. [Atmos.] 102: 3779-3793.
McMurry, P. H.; Woo, K. S.; Weber, R.; Chen, D.-R.; Pui, D. Y. H. (2000) Size distributions of 3 to 10 nm
atmospheric particles: implications for nucleation mechanisms. Philos. Trans. R. Soc. London Ser. A
358: 2625-2642.
Monn, C. (2001) Exposure assessment of air pollutants: a review on spatial heterogeneity and
indoor/outdoor/personal exposure to suspended paniculate matter, nitrogen dioxide and ozone. Atmos.
Environ. 35: 1-32.
Motallebi, N. (1999) Wintertime PM2 5 and PM10 source apportionment at Sacramento, California. J. Air Waste
Manage. Assoc. 49: PM-25-34.
NARSTO. (2002) Fine particle assessment (2nd review draft). Available: http://www.cgenv.com/Narsto/
(13 February 2002).
3-109
-------
Neas, L. M. (1996) [Data provided for 1986 from Harvard six cities data base.] Boston, MA: Harvard School of
Public Health.
Norris, G. (1998) Air pollution and the exacerbation of asthma in an arid, western, US city [dissertation]. Seattle,
WA: University of Washington. Available from: University Microfilms International, Ann Arbor, MI;
publication no. 9836228.
Norris, G.; Young-Pong, S. N.; Koenig, J. Q.; Larson, T. V.; Sheppard, L.; Stout, J. W. (1999) An association
between fine particles and asthma emergency department visits for children in Seattle. Environ. Health
Perspect. 107: 489-493.
Odum, J. R.; Hoffmann, T.; Bowman, F.; Collins, D.; Flagan, R. C.; Seinfeld, J. H. (1996) Gas/particle partitioning
and secondary organic aerosol yields. Environ. Sci. Technol. 30: 2580-2585.
Odum, J. R.; Jungkamp, T. P. W.; Griffin, R. J.; Flagan, R. C.; Seinfeld, J. H. (1997a) The atmospheric
aerosol-forming potential of whole gasoline vapor. Science (Washington, DC) 276: 96-99.
Odum, J. R.; Jungkamp, T. P. W.; Griffin, R. J.; Forstner, H. J. L.; Flagan, R. C.; Seinfeld, J. H. (1997b) Aromatics,
reformulated gasoline, and atmospheric organic aerosol formation. Environ. Sci. Technol. 31: 1890-1897.
Ondov, J. M. (1996) Paniculate tracers for source attribution: potential for application to California's San Joaquin
Valley. J. Aerosol Sci. 27(suppl. 1): S687-S688.
Ott, S.-T.; Ott, A.; Martin, D. W.; Young, J. A. (1991) Analysis of a trans-Atlantic Saharan dust outbreak based on
satellite and GATE data. Mon. Weather Rev. 119: 1832-1850.
Paatero, P.; Tapper, U. (1993) Analysis of different modes of factor analysis as least squares fit problems. Chemom.
Intelligent Lab. Syst. 18: 183-194.
Paatero, P.; Tapper, U. (1994) Positive matrix factorization: a non-negative factor model with optimal utilization of
error estimates of data values. Environmetrics 5: 111-126.
Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1991) Aerosol formation in the photooxidation of
isoprene and p-pinene. Atmos. Environ. Part A 25: 997-1008.
Pandis, S. N.; Harley, R. A.; Cass, G. R.; Seinfeld, J. H. (1992) Secondary organic aerosol formation and transport.
Atmos. Environ. Part A 26: 2269-2282.
Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Dutcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
Pinto, J. P.; Pleasant, M.; Kellogg, R.; Torok, S.; Stiles, D.; Willis, R. D.; Koutrakis, P.; Allen, G. A. (1995)
Chemical characterization of the ambient aerosol in Philadelphia. Presented at: Paniculate matter: health and
regulatory issues: an A&WMA international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA:
Air & Waste Management Association.
Pinto, J. P.; Stevens, R. K.; Willis, R. D.; Kellogg, R.; Mamane, Y.; Novak, J.; Santroch, J.; Benes, L; Lenicek, J.;
Bures, V. (1998) Czech air quality monitoring and receptor modeling study. Environ. Sci. Technol.
32: 843-854.
Pio, C.; Alves, C.; Duarte, A. (2001) Organic components of aerosols in a forested area of central Greece. Atmos.
Environ. 35: 389-401.
Placet, M.; Battye, R. E.; Fehsenfeld, F. C.; Bassett, G. W. (1991) Emissions involved in acidic deposition
processes. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume I: emissions,
atmospheric processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment
Program. (State of science and technology report no. 1).
Poore, M. W. (2000) Oxalic acid in PM2 5 paniculate matter in California. J. Air Waste Manage. Assoc.
50: 1874-1875.
Prospero, J. M. (1996) The atmospheric transport of particles to the ocean. In: Ittekkot, V.; Honjo, S.; Depetris, P. J.,
eds. Particle flux in the ocean. New York, NY: John Wiley & Sons, Ltd.; pp. 19-52. (SCOPE report no. 57).
Prospero, J. M. (1999) Long-range transport of mineral dust in the global atmosphere: impact of African dust on the
environment of the southeastern United States. Proc. Natl. Acad. Sci. U.S.A. 96: 3396-3403.
Prospero, J. M.; Nees, R. T.; Uematsu, M. (1987) Deposition rate of paniculate and dissolved aluminum derived
from Saharan dust in precipitation at Miami, Florida. J. Geophys. Res. [Atmos.] 92: 14,723-14,731.
Prospero, J. M.; Olmez, I.; Ames, M. (2001) Al and Fe in PM25 and PM10 suspended particles in south-central
Florida: the impact of the long range transport of African mineral dust. J. Water Air Soil Pollut. 125: 291-317.
Pun, B. K.; Seigneur, C.; Grosjean, D.; Saxena, P. (2000) Gas-phase formation of water-soluble organic compounds
in the atmosphere: a retrosynthetic analysis. J. Atmos. Chem. 35: 199-223.
Pun, B. K.; Wu, S.-Y.; Seigneur, C. (2002) Contribution of biogenic emissions to the formation of ozone and
paniculate matter in the eastern United States. Environ. Sci. Technol. 36: 3586-3596.
Ramadan, Z.; Song, X.-H.; Hopke, P. K. (2000) Identification of sources of Phoenix aerosol by positive matrix
factorization. J. Air Waste Manage. Assoc. 50: 1308-1320.
3-110
-------
Rizzo, M; Pinto, J. P. (2001) Initial characterization of fine paniculate matter (PM25) collected by the National
Federal Reference Monitoring Network. Presented at: 94th annual conference & exhibition of the Air &Waste
Management Association; June, Orlando, FL. Pittsburgh, PA: Air & Waste Management Association.
Rodes, C. E.; Evans, E. G. (1985) Preliminary assessment of 10 um paniculate sampling at eight locations in the
United States. Atmos. Environ. 19: 293-303.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993) Sources of fine organic
aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci.
Technol. 27:636-651.
Ryan, W. M.; Badgett-West, C. R.; Holtz, D. R.; Peters, T. A.; Cooper, J. A.; Ono, D. (1988) Reconciliation of
receptor and dispersion modeling impacts of PM10 in Hayden, Arizona. In: Mathai, C. V.; Stonefield, D. H.,
eds. PM-10: implementation of standards, an APCA/EPA international specialty conference; February;
San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 419-429. (APCA transactions
series TR-13).
Salma, L; Maenhaut, W.; Zaray, G. (2002) Comparative study of elemental mass size distributions in urban
atmospheric aerosol. J. Aerosol Sci. 33: 339-356.
Sarwar, G.; Corsi, R.; Kumura, Y.; Allen, D.; Weschler, C. J. (2002) Hydroxyl radicals in indoor environments.
Atmos. Environ. 36: 3973-3988.
Savoie, D. L.; Prospero, J. M. (1980) Water-soluble potassium, calcium, and magnesium in the aerosols over the
tropical North Atlantic. J. Geophys. Res. 85: 385-392.
Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem. 24: 57-109.
Schauer, J. J.; Cass, G. R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants
using organic compounds as tracers. Environ. Sci. Technol. 34: 1821-1832.
Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurik, M. A.; Cass, G. R. (1996) Source apportionment of
airborne paniculate matter using organic compounds as tracers. Atmos. Environ. 30: 3837-3855.
Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles? J. Air
Waste Manage. Assoc. 46: 927-939.
Shinn, E. A.; Smith, G. W.; Prospero, J. M.; Betzer, P.; Hayes, M. L.; Garrison, V.; Barber, R. T. (2000) African
dust and the demise of Caribbean coral reefs. Geophys. Res. Lett. 27: 3029-3032.
Skidmore, L. W.; Chow, J. C.; Tucker, T. T. (1992) PM10 air quality assessment for the Jefferson County, Ohio air
quality control region. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source
controls, an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ. Pittsburgh, PA:
Air and Waste Management Association; pp. 1016-1031. (A&WMA transactions series no. 22).
Smith, G. T.; Ives, L. D.; Nagelkerken, I. A.; Ritchie, K. B. (1996) Caribbean sea fan mortalities. Nature (London)
383: 487.
Song, X.-H.; Polissar, A. V.; Hopke, P. K. (2001) Sources of fine particle composition in the northeastern US.
Atmos. Environ. 35: 5277-5286.
Spengler, J. D.; Briggs, S. L. K.; Ozkaynak, H. (1986) Relationship between TSP measurements and
size-fractionated particle mass measurements in six cities participating in the Harvard Air Pollution Health
Study. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards; EPA cooperative agreement no. CR-812667-01-0; December 5.
Spengler, J. D.; Koutrakis, P.; Dockery, D. W.; Raizenne, M.; Speizer, F. E. (1996) Health effects of acid aerosols
on North American children: air pollution exposures. Environ. Health Perspect. 104: 492-499.
Strader, R.; Lurmann, F.; Pandis, S. N. (1999) Evaluation of secondary organic aerosol formation in winter. Atmos.
Environ. 33:4849-4863.
Suarez, A. E.; Caffrey, P. F.; Ondov, J. M.; Thaung, K. C. (1996) Size distribution of Ir-tagged soot from diesel
sanitation trucks. J. Aerosol Sci. 27(suppl. 1): S697-S698.
Tanner, R. L.; Parkhurst, W. J. (2000) Chemical composition of fine particles in the Tennessee Valley region. J. Air
Waste Manage. Assoc. 50: 1299-1307.
Taylor, C. A., Jr.; Stover, C. A.; Westerdahl, F. D. (1998) Speciated fine particle (<2.5 um aerodynamic diameter)
and vapor-phase acid concentrations in southern California. Presented at: 91st annual meeting & exhibition of
the Air & Waste Management Association; June; San Diego, CA. Pittsburgh, PA: Air & Waste Management
Association; paper no. WA7401.
Taylor, P. E.; Flagan, R. C.; Valenta, R.; Glovsky, M. M. (2002) Release of allergens as respirable aerosols: a link
between grass pollen and asthma. J. Allergy Clin. Immunol. 109: 51-56.
Thanukos, L. C.; Miller, T.; Mathai, C. V.; Reinholt, D.; Bennett, J. (1992) Intercomparison of PM10 samplers and
source apportionment of ambient PM10 concentrations at Rillito, Arizona. In: Chow, J. C.; Ono, D. M., eds.
3-111
-------
PM10 standards and nontraditional particulate source controls, an A&WMA/EPA international specialty
conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association;
pp. 244-261. (A&WMA transactions series no. 22).
Tobias, H. I; Ziemann, P. J. (2000) Thermal desorption mass spectrometric analysis of organic aerosol formed from
reactions of 1-tetradecene and O3 in the presence of alcohols and carboxylic acids. Environ. Sci. Technol.
34:2105-2115.
Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin: a descriptive
analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
Turpin, B. J.; Lim, H.-J. (2001) Species contributions to PM2 5 mass concentrations: revisiting common assumptions
for estimating organic mass. Aerosol Sci. Technol. 35: 602-610.
Turpin, B. J.; Huntzicker, J. J.; Larson, S. M; Cass, G. R. (1991) Los Angeles summer midday particulate carbon:
primary and secondary aerosol. Environ. Sci. Technol. 25: 1788-1793.
U.S. Environmental Protection Agency. (1995) Compilation of air pollutant emission factors. Volume I: stationary
point and area sources. 5th ed. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
report no. AP-42-ED-5. Available from: NTIS, Springfield, VA; PB95-196028.
U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
U.S. Environmental Protection Agency. (1999) Getting started: emissions inventory methods for PM-2.5. Research
Triangle Park, NC: Office of Air Quality Planning and Standards, Emission Factor and Inventory Group; EIIP
document series - volume IX. Available: http://www.epa.gov/ttn/chief/eiip/techreport/volume09/index.html
[7 May 2003].
U.S. Environmental Protection Agency. (2000a) Aerometric Information Retrieval System (AIRS). Research
Triangle Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/airsaqs
[7 May 2003].
U.S. Environmental Protection Agency. (2000b) Air quality criteria for carbon monoxide. Research Triangle Park,
NC: National Center for Environmental Assessment; report no. EPA/600/P-99/00 IF. Available:
http://www.epa.gov/ncea/pdfs/coaqcd.pdf (7 May 2003).
U.S. Environmental Protection Agency. (2000c) National air pollutant emission trends, 1900-1998. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-00-002. Available:
http://www.epa.gov/ttn/chief/trends/trends98/ [7 May 2003].
U.S. Environmental Protection Agency. (2001) National air quality and emissions trends report, 1999. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA 454/R-01-004. Available:
http://www.epa.gov/oar/aqtrnd99/ [7 May 2003].
U.S. Environmental Protection Agency. (2002a) Latest findings on national air quality: 2001 status and trends.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards; report no. EPA-454/K-02-001.
U.S. Environmental Protection Agency. (2002b) Aerometric Information Retrieval System (AIRS). Research
Triangle Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/airsaqs
[2 June 2003].
Vermette, S. J.; Williams, A. L.; Landsberger, S. (1992) PM10 source apportionment using local surface dust
profiles: examples from Chicago. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional
particulate source controls, an A&WMA/EPA international specialty conference, v. I; January; Scottsdale,
AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 262-278. (A&WMA transactions series
no. 22).
Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of submicron
particle exposure. Environ. Health Perspect. 108: 1139-1145.
Watson, J. G.; Chow, J. C. (2000) Reconciling urban fugitive dust emissions inventory and ambient source
contribution estimates: summary of current knowledge and needed research. Reno, NV: Desert Research
Institute; DRI document no. 6110.4F.
Watson, J. G.; Chow, J. C. (2001) Estimating middle-, neighborhood-, and urban-scale contributions to elemental
carbon in Mexico City with a rapid response aethalometer. J. Air Waste Manage. Assoc. 51: 1522-1528.
Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
(1990a) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
3-112
-------
Watson, J. G.; Chow, J. C.; Pritchett, L. C.; Houck, J. A.; Ragazzi, R. A.; Burns, S. (1990b) Chemical source
profiles for paniculate motor vehicle exhaust under cold and high altitude operating conditions. Sci. Total
Environ. 93: 183-190.
Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994)
Chemical mass balance source apportionment of PM10 during the Southern California Air Quality Study.
Aerosol Sci. Technol. 21: 1-36.
West, J. J.; Ansari, A. S.; Pandis, S. N. (1999) Marginal PM25: nonlinear aerosol mass response to sulfate reductions
in the eastern United States. J. Air Waste Manage. Assoc. 49: 1415-1424.
White, W. H. (1998) Statistical considerations in the interpretation of size-resolved paniculate mass data. J. Air
Waste Manage. Assoc. 48: 454-458.
Wilson, W. E.; Stockburger, L. (1990) Diurnal variations in aerosol composition and concentration. In: Masuda, S.;
Takahashi, K., eds. Aerosols: science, industry, health and environment: proceedings of the third international
aerosol conference, v. 2; September; Kyoto, Japan. Oxford, United Kingdom: Pergamon Press; pp. 962-965.
Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
Winer, A. M.; Arey, J.; Atkinson, R.; Aschmann, S. M.; Long, W. D.; Morrison, C. L.; Olszyk, D. M. (1992)
Emission rates of organics from vegetation in California's Central Valley. Atmos. Environ. Part A
26: 2647-2659.
Winterhalter, R.; Neeb, P.; Grossmann, D.; Kolloff, A.; Horie, O.; Moortgat, G. (2000) Products and mechanisms of
the gas phase reaction of ozone with p-pinene. J. Atmos. Chem. 35: 165-197.
Wongphatarakul, V.; Friedlander, S. K.; Pinto, J. P. (1998) A comparative study of PM25 ambient aerosol chemical
databases. Environ. Sci. Technol. 32: 3926-3934.
Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; McMurry, P. H. (2001) Measurement of Atlanta aerosol size distributions:
observations of ultrafine particle events. Aerosol Sci. Technol. 34: 75-87.
Wotawa, G.; Trainer, M. (2000) The influence of Canadian forest fires on pollutant concentrations in the United
States. Science (Washington, DC) 288: 324-328.
Yu, J.; Cocker, D. R., Ill; Griffin, R. J.; Flagan, R. C.; Seinfeld, J. H. (1998) Gas-phase ozone oxidation of
monoterpenes: gaseous and paniculate products. J. Atmos. Chem. 34: 207-258.
Yu, J.; Griffin, R. J.; Cocker, D. R., Ill; Flagan, R. C.; Seinfeld, J. H. (1999) Observation of gaseous and paniculate
products of monoterpene oxidation in forest atmospheres. Geophys. Res. Lett. 26: 1145-1148.
Zhang, S.-H.; Shaw, M.; Seinfeld, J. H.; Flagan, R. C. (1992) Photochemical aerosol formation from a-pinene and
P-pinene. J. Geophys. Res. [Atmos.] 97: 20,717-20,729.
Zweidinger, R.; Purdue, L.; Fitzgerald, K.; Carmichael, L.; Kellog, R. (1998) NERL PM research monitoring
platforms: Baltimore, Fresno and Phoenix, data report. Research Triangle Park, NC: U.S. Environmental
Protection Agency, National Exposure Research Laboratory; report no. EPA/600/R-98/138. Available from:
NTIS, Springfield, VA; PB99-105660.
3-113
-------
APPENDIX 3A
Spatial and Temporal Variability of the Nationwide AIRS
PM2 5 and PM10_2 5 Data Sets
Aspects of the spatial and temporal variability of 24-h average PM25 concentrations for
1999, 2000, and 2001 in a number of metropolitan statistical areas (MS As) across the United
States are presented in this Appendix. PM2 5 data for multiple sites in 27 urban areas have been
obtained from the AIRS data base and analyzed for their seasonal variations, for their spatial
correlations, and for their spatial uniformity. A number of aspects of the spatial and temporal
variability of the PM25 data set from 1999 were presented in Rizzo and Pinto (2001) based in
part on analyses given in Fitz-Simons et al. (2000).
Information about seasonal and spatial variability in PM2 5 concentrations within 27 MSAs
across the United States are provided in the accompanying figures (Figures 3A-1 to 3A-27).
Underneath the value for r, the 90th percentile (P90) values of the absolute difference in PM2 5
concentrations (in |ig/m3) and the coefficient of divergence (COD) are provided in parentheses.
Beneath these two measures of spatial variability, the number of observations used in the
calculations of the statistics in part c of each figure is given.
Quality assured measurements from four monitoring sites for at least fifteen days during
each calendar quarter for 1999, 2000, and 2001 (preferably) or for 2000 and 2001 at a minimum
in a given MSA were required for their inclusion in the analyses presented in this appendix. The
Columbia, SC and Baton Rouge, LA MSAs, which had only three sites meeting this criterion,
are exceptions. Typically, at least 200 measurements were available for each monitoring site
chosen. Monitoring sites were chosen without consideration of the land use type used to
characterize their locations.
Because of changes in monitoring strategies, funding levels etc., there were year to year
changes in monitoring sites meeting the above criteria in a number of MSAs. Data for the MSAs
in Philadelphia, PA, Norfolk, VA, Pittsburgh, PA, Detroit, MI, Chicago, IL, Louisville, KY,
St. Louis, MO, and the Dallas, TX have been analyzed only for 2000 and 2001 because of a lack
of consistent coverage in 1999.
5A-1
-------
Information about seasonal and spatial variability in PM10_2 5 concentrations within
17 MS As across the United States are provided in the accompanying figures (Figures 3A-28 to
3A-44). Underneath the value for r, the 90th percentile values of the absolute difference in PM25
concentrations (in |ig/m3) and the CODs are given in parentheses. Beneath these two measures
of spatial variability, the number of observations used in the calculations of the statistics in part c
of each figure is given. In order to maximize coverage, data were calculated for a number of
sampling periods. Only Milwaukee, WI, and Salt Lake City, UT, had enough data for a 3 year
average (1999 to 2001). Tampa, FL; Cleveland, OH; Steubenville, OH; Baton Rouge, LA;
Portland, OR; and Riverside, CA had data for a 2 year average (2000 to 2001), as did Chicago,
IL and Pittsburgh, PA (1999 to 2000). Other MS As had only one year of data (for 2000 or
2001).
The COD was defined mathematically and used earlier in Chapter 3 as a measure of the
degree of similarity between two data sets. A COD of zero implies that values in both data sets
are identical, and a COD of one indicates that two data sets are completely different. Values
of P90 provide a measure in absolute terms of differences in concentrations between sites, and
CODs provide a relative measure of these differences. The maximum number of days of
coincident data from paired sites were used to calculate correlation coefficients, values for P90,
and CODs. The correlation coefficients were also calculated by using only concurrent
measurements obtained at all of the monitoring sites within urban areas meeting the above
selection criteria. The correlation coefficients that were calculated differed only in the third
significant figure between the two methods. Metrics used above for characterizing differences
between separated monitors are applied to collocated monitors in Table 3A-1.
Figures 3 A-28 to 3 A-44 summarize information about the spatial and temporal variability
of 24-h average PM10_25 concentrations. Data are shown for a subset of MS As included in the
analyses for PM25. Not all MSAs could be included, because of a lack of data for some.
A schematic map showing locations of sampling sites within each MSA is given in part a, at the
top of each figure. Also included in the map are major highways and a distance scale. A key
giving the AIRS site ID numbers (#'s) is shown alongside each map. Box plots showing lowest,
lower quartile, median, upper quartile and highest PM2 5 concentrations for each calendar quarter
are shown in part b of each figure. AIRS site ID #'s, annual mean concentrations, the number of
observations, and the standard deviation of the data are shown above the box plots. Finally,
5A-2
-------
in part c of each figure, statistics characterizing the spatial variability in PM25 concentrations are
given. For each site-pair, the Pearson correlation coefficient (r) is provided. Underneath each
value for r, the 90th percentile of the absolute difference in PM10_2 5 concentrations, the COD, and
number of observations is given. In some cases, because of negative concentration values, the
COD may not be calculated. Dashes are shown for such cases.
It should be noted that the vertical axes for each MSA are drawn to different scales.
Therefore, care should be taken in attempting to compare values in different MSAs. For PM2 5,
the maxima on the y-axis vary by up to a factor of four, from 40 |ig/m3 to 150 |ig/m3. The use of
a single range of values would result in compression of the figures and loss of detail in a number
of MSAs which have a lower range of values. Comparisons between MSAs can more easily be
made by inspection of Figures 3-4a,b and 3-6a,b.
REFERENCES
Fitz-Simons, T. S.; Mathias, S.; Rizzo, M. (2000) Analyses of 1999 PM data for the PMNAAQS review. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
November 17. Available: http://www.epa.gov/oar/oaqps/pm25/analyses.html [2 April, 2002].
Rizzo, M.; Pinto, J. P. (2001) Initial characterization of fine paniculate matter (PM25) collected by the National
Federal Reference Monitoring Network. Presented at: 94th annual conference & exhibition of the Air &
Waste Management Association; June, Orlando, FL. Pittsburgh, PA: Air & Waste Management Association.
JA-2
-------
TABLE 3A-1. PERFORMANCE METRICS FOR PM2 5 FROM
COLLOCATED SAMPLERS (Mean concentrations, the number of samples, the
standard deviation at each site, and the Pearson correlation coefficient and the
coefficient of divergence for the site pair are shown; concentrations are given in ug/m3).
Columbia, SC
MEAN
N
SD
r
COD
Dallas, TX
MEAN
N
SD
r
COD
Detroit, MI
MEAN
N
SD
r
COD
Grand Rapids, MI
MEAN
N
SD
r
COD
Louisville, KY
MEAN
N
SD
r
COD
Steubenville, OH
Mean
Obs
SD
r
COD
Washington DC
Mean
Obs
SD
r
COD
Sampler 1
45-079-0019
15.3
226
6.5
Sampler 1
48-113-0069
12.7
687
5.7
Sampler 1
26-163-0001
16.5
637
9.2
Sampler 1
26-081-0020
14
1050
8.6
Sampler 1
18-043-1004
15.8
196
8.2
Sampler 1
54-029-0011
16.5
342
10.2
Sampler 1
51805
16.8
600
9.7
0.995
0.022
0.996
0.032
0.986
0.041
0.992
0.059
0.997
0.027
0.985
0.087
0.953
0.15
Sampler 2
45-079-0019
15.5
211
6.5
Sampler 2
48-113-0069
13.1
116
6
Sampler 2
26-163-0001
16.2
111
8.8
Sampler 2
26-081-0020
14.3
181
8.3
Sampler 2
18-043-1004
16
104
7.8
Sampler 2
54-029-0011
16.6
325
10.3
Sampler 2
51805
17.5
132
11.1
5A-4
-------
Philadelphia, PA MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
34-007-1007
34-015-5001
42-017-0012
42-045-0002
42-091-0013
42-101-0004
42-101-0136
B
Mean
Obs
SD
in
cs
14.8
197
9.2
14.6
208
8.6
14.1
217
8.5
16.0
230
8.6
14.2
221
8.9
15.7
610
9.2
15.8
616
9.3
1234 1234 1234
1234
Quarter
1234 1234 1234
c.
Site
A
B
C
D
E
F
G
1 0.91 0.93
(6.3,0.14) (5.2,0.15)
170 167
1 0.84
(7.5,0.19)
176
1
0.87
(6.9, 0.19)
183
0.88
(7.4, 0.18)
194
0.85
(7.5. 0.16)
198
1
0.88
(5.0,0.16)
176
0.83
(7.1,0.18)
184
0.88
(4.7,0.13)
200
0.87
(6.1.0.15)
208
1
0.94
(4.6,0.15)
163
0.88
(7.4,0.17)
169
0.94
(4.9,0.11)
177
0.94
(5.1,0.11)
187
0.90
(4.6,0.11)
181
1
0.93
(5.1,0.14)
186
0.85
(6.9, 0.18)
173
0.89
(5.3, 0.13)
180
0.88
(4.0, 0.12)
193
0.87
(4.5,0.11)
185
0.96
(3.3, 0.08)
550
1
Figure 3A-1. Philadelphia, PA-NJ MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM2 5 concentrations (ug/m3) for
2000-2001; (c) Intersite correlation statistics for each data pair given as the
correlation coefficient, (P90, COD), and number of measurements.
5A-5
-------
Washington, DC MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
11-001-0041
11-001-0043
51-013-0020
51-059-0030
51-107-1005
54-003-0003
B
Mean
Obs
SD
100-
co
E 75-
^2.
iq
1 -
25-
0-
16.7 15.5 14.7 14.1 13.8 16.1
859 926 337 848 324 314
9.8 8.7 8.0 7.4 8.2 9.3
I
!
! I
! In
n ' ' ' I '
!lltl||l||lt|l
1234 1234 1234 1234 1234 1234
Quarter
C. Site
A
D
A
B
C
D
E
F
1 0.83 0.91
(5,1,0.18) (5.7,0.15)
784 273
1 0.84
(5.1,0.16)
297
1
0.82
(7,4,0.17)
691
0.87
(6.2,0.16)
731
0.96
(3.5, 0.08)
276
1
0.84
(7,7, 0,19)
267
0.79
(7.5, 0.20)
284
0.94
(4.9, 0.11)
312
0,91
(4.5, 0.11)
266
1
0.77
(8,8, 0,20)
262
0,77
(7.6,0.19)
281
0.85
(8.3.0.16)
289
0,82
(9.1.0.18)
246
0.87
(8.5.0,17)
284
1
Figure 3A-2. Washington, DC MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
5A-6
-------
Norfolk, VA MSA
Site A
SiteB
SiteC
SiteD
SiteE
AIRS Site ID
51-550-0012
51-650-0004
51-700-0013
51-710-0024
51-810-0008
b.
B
Mean
Obs
SD
in
evi
40-
30-
20-
10-
13.6
598
6.9
13.5
208
7.2
12.6
222
6.7
13.7
227
7.0
c.
Site
1234 1234
B
1234
Quarter
13.0
232
7.0
1234 1234
D
A
B
C
D
E
1 0.94 0.91
(3.7,0.09) (5.0,0.11)
167 180
1 0.98
(2.6, 0.09)
191
1
0.96
(3.0, 0.08)
182
0.96
(3.1,0.07)
194
0.96
(3.6, 0.09)
206
1
0.94
(3.5, 0.08)
184
0.93
(4.0, 0.09)
198
0.92
(4.3,0.11)
212
0.93
(3.6,0-10)
217
1
Figure 3A-3. Norfolk, VA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
5A-7
-------
Columbia, SC MSA
Site A
SiteB
SiteC
B
Mean
Obs
SD
60—1
50 —
to
30 —
20 —
10 —
15.7
343
6.9
14.7
337
6.6
C. Site
B
15.5
341
6.8
1234 1234 1234
Quarter
A
B
C
1 0.93
(3.3, 0,08)
316
1
0.95
(3.2, 0.0?)
319
0.97
(2.8, 0.06)
313
1
AIRS Site ID
45-063-0008
45-079-0007
45-079-0019
Figure 3A-4. Columbia, SC MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
5A-8
-------
Atlanta, GA MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
13-063-0091
13-089-0002
13-089-2001
13-121-0032
13-121-0039
13-121-1001
13-223-0003
b.
n
Mean
Obs
SD
125—
100—
75—
50—
25—
0—
A B
19.2 18.:
316 88(
9.4 9.7
W II
l!
1
1
1
1
1
c
9.f
18
O.C
I
,
I
1
£
!!
D
8.7
98
i.8
|
E F (
21.2 18.4 1
317 304 3
10.6 9.3 1
I ((I) |,
3
6.8
23
1.7
J
1234 1234 1234 1234 1234 1234
Quarter
C. Site
A
B
C
D
E
F
1 0.77 0.87
(8.2, 0.16) (6.4.0 12)
256 273
1 0.73
(6.4,0.15)
750
1
0.82
(6.7.0.16)
259
0.67
(7.1,0.18)
739
0.75
(5.3, 0.13)
767
1
0.78
(7.5.0.15)
267
0.81
(9.3,0,17)
260
0.87
(7.1.0.12)
273
0.82
(8.7,0.15)
261
1
o.as
{7.2, 0,161
260
0,77
(9.0, 0,18)
250
0.82
(7.9. 0 17)
260
0.83
(7.1.016)
254
0.75
(10,8,0.18)
259
1
0.59
(11,2,0,21)
272
0,63
(10,9,0.22)
268
0.62
(10.3,0.20)
273
0.59
(10,0,0,21)
268
0.59
(13.2,0.24)
277
O.S9
(8.6, 0.19)
260
Figure 3A-5. Atlanta, GA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
5A-9
-------
Birmingham, AL MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
01-073-0023
01-073-1005
01-073-2003
01-073-2006
01-073-5002
b.
B
Mean
Obs
SD
n
~
60-
40-
20-
21.6
1056
11.1
16.7
360
8.4
20.0
1046
10.3
17.6
346
8.5
16.6
356
8.2
1234
Quarter
c.
Site
A
B
D
A
B
C
D
E
1 0,80 0.86
(14.5,0.20) (9.8,0.15)
356 1011
1 0.79
(10.2.0.18)
348
1
0.79
(13.7,0.20)
342
0.86
(7.8,0.16)
334
0.78
(9.9,0.18)
335
1
0.80
(15.2. 0.21)
353
0.86
(6.7, 0.15)
345
0.78
(10.3.0.18)
344
0.86
(7.6, 0.15)
329
1
Figure 3A-6. Birmingham, AL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3 A-10
-------
Tampa, FL MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
12-057-0030
12-057-1075
12-103-0018
12-103-1008
i
50
B
in
oi
Mean
Obs
SD
100 —
80 —
40 —
20 —
0 —
12.7
987
5.7
II
11
|
1
1
1
1
11
2.3
00.
5.0
?
j
1
1
e
I
1.8
D24
5.0
J
11.1
337
6.0
JlL
||TT!
!!
1234 1234 1234
Quarter
c.
Site A
B
D
A
B
C
n
1 0.79
(3.6, 0,10)
919
1
0.87
(4.0, 0.11 ')
920
0.70
(4.6,0.14)
939
1
0.87
(4.3,0.12)
308
0.71
(5.0,0.13)
314
0.82
(3.1,0.10)
325
1
Figure 3A-7. Tampa, FL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3 A-11
-------
Cleveland, OH MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
39-035-0013
39-035-0038
39-035-0060
39-035-0065
39-035-0066
39-035-1002
39-085-1001
39-093-2003
Mean
Obs
SD
B
H
CO
"
18.3
368
9.6
20.2
931
11.5
18.4
353
17.5
340
9.0
14.7
332
8.4
15.0
351
8.2
14.0
342
8.4
15.2
298
8.8
II
1234 1234
1234 1234
Quarter
1234 1234
c.
Site
A
B
C
D
E
F
G
H
1 0.91 0.96 0.94
(7.1,0.13) (3.3,0.12) (5.4,0.10)
320 322 314
1 0.92 0.89
(6.9,0.14) (9.4,0.15)
306 296
1 0.93
(5.1,0.14)
309
1
0.92
(7.2,0.18)
300
0.85
(13.1,0.21)
280
0.90
(8.15,0.19)
300
096
(4.7, 0.14)
295
1
0.88
(9.0,0.18)
308
0.84
(12.9. 0,21)
304
0.87
(8.6, 0.20)
310
0-91
(5.4, 0.16)
310
0.88
(4.9,0.16)
304
1
0.89
(10.7,0.21)
308
0.84
(14.3,0.23)
294
0.88
(10.8,0.22)
307
0-90
(7.8, 0.20)
306
0.91
(5.8, 0.15)
295
0.89
(6.0, 0.18)
303
1
0.92
{8.1,0-17)
265
0.892
(11.2, 0.18)
256
0.90
(8.9,0.18)
256
0-91
(7.7,0.18)
264
0.91
(5.8,0.13)
247
0.87
(5.3,0.18)
261
0.90)
{6.6,0.15)
275
1
Figure 3A-8. Cleveland, OH MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-12
-------
Pittsburgh, PA MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
Site I
SiteJ
SiteK
42-003-0008
42-003-0021
42-003-0064
42-003-0116
42-003-1008
42-003-1301
42-007-0014
42-125-0005
42-125-0200
42-125-5001
42-129-0008
b.
in 50-
rg
16.4
596
8.5
|||
15.4
207
9.3
ll
1234
ll
1234
22.0
665
14.5
I
j
15.5
201
8.7
II
ll
1234 1234
16.1
201
9.5
,
I
16.9
197
9.8
I
'!
!!
16.6 15.6 15.5 13.8 1
201 226 227 656 S
9.7 8.5 8.2 7.7
I
I
i
• 1 11 \ \
II li
6.0
16
$.8
J
1234 1234 1234 1234 1234 1234 1234
Quarter
c.
Site
A
B
C
D
E
F
G
H
I
J
K
1 0.96 0 79 0 97 0 94
(2.6,0.07) (15.9,0.19) (3.0,0.07) (4.7,0.09)
171 572 168 167
1 0.77 0.95 0.92
(18.0,0.21) (3.4,0.08) (6.1,0.10)
193 180 173
0.93
(5.2,0.10)
173
0.91
(7.0,0.13)
179
0.94
(5.9, 0.09)
164
0.93
(6.1,0.10)
132
0.92
(5.1,0.11)
183
0.93
(4.2, 0.09)
149
0.93
(5.1,0.10)
185
0.94
(4.8, 0.09)
151
0.91
(6.6,0.14)
531
0.91
(6.4,0.13)
189
0.87
(5.4,0.11)
174
0.87
(4.9,0.13)
144
1 0.80 0.81 0.86 0.72 0.74 0.67 0.69 0.66
(16.5,0.20) (17.1,0.20) (12.6,0.15) (19.1,0.21) (18.1,0.22) (20.3,0.22) (21.0,0.26) (18.1,0.23)
188 188 186 182 203 203 601 194
1 0.92
(5.8,0.11)
169
1
0.91
(6.0,0.12)
172
0.91
(6.2,0.13)
165
1
0.94
(5.7,0.11)
131
0.94
(5.3,0.11)
126
0.91
(7.1,0.13)
126
1
0.94
(4.9,0.10)
143
0.90
(7.0,0.13)
147
0.89
(6.7,0.14)
144
0.89
(8.0,0.13)
186
1
0.94
(4.9,0.10)
143
0.89
(7.0,0.14)
147
0.86
(7.6,0.16)
144
0.89
(8.1,0.13)
184
0.94
(4.7, 0.09)
219
1
0.91
(6.6,0.14)
181
0.89
(8.7,0.16)
190
0.86
(10.3,0.19)
177
0.90
(8.3,0.16)
179
0.89
(4.7,0.14)
201
0.94
(5.6,0.12)
202
1
0.87
(5.5,0.13)
142
0.83
(6.6,0.15)
137
0.86
(7.2,0.15)
139
0.81
(9.5,0.16)
175
0.88
(5.7,0.12)
200
0.88
(6.3,0.12)
202
0.84
(7.3,0.16)
195
1
Figure 3A-9. Pittsburgh, PA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (jig/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3 A-13
-------
Steubenville, OH MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
39-081-0016
39-081-1001
54-009-0005
54-029-0011
54-029-1004
50km
b.
B
Mean
Obs
SD
n
~
20-
18.9
323
10.1
18.5
856
9.0
17.3
332
8.4
16.5
342
10.2
17.3
325
9.8
1234 1234
1234
Quarter
C. Site
1234 1234
D
A
B
C
D
E
1 0.85 O.B8
(9.6,0.16) (8.0,0.16)
269 294
1 0.87
(6.3, 0.16)
267
1
0.86
(7.7, 0.16)
302
0.84
(8.6, 0.17)
277
0.90
(7.4, 0.15)
320
1
0.85
(7.9,0.18)
290
0.79
(10, 0.20)
266
0.89
(7.1.0.16)
295
0.93
(6.2,0.16)
307
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
(ug/m3) for 2000-2001; (c) Intersite correlation statistics for each data pair
given as the correlation coefficient, (P90, COD), and number of
measurements.
3A-14
-------
Detroit, Ml MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
Site I
SiteJ
26-099-0009
26-115-0005
26-125-0001
26-147-0005
26-163-0001
26-163-0015
26-163-0016
26-163-0025
26-163-0033
26-163-0036
b.
Mean 13.5
Obs 228
SD 8.6
15.2 15.3 14.1
228 186 214
9.0 9.4 9.2
16.5
637
9.2
18.2
233
10.0
15.7
653
9.4
14.6
228
9.2
19.9
223
10.6
17.9
216
10.7
1234 1234 1234 1234
1234 1234 1234 1234 1234 1234
Quarter
c.
Site
A
B
C
D
E
F
G
H
1 0.89 0.95 0.95
(7.7,0.16) (5.0,0.13) (5.4,0.14)
210 172 203
1 0.91 0.83
(6.5,0.14) (9.7,0.20)
169 196
1 0.92
(6.4,0.15)
161
1
0.94
(6.7,0.16)
202
0.95
(5.6,0.10)
200
0.96
(5.0,0.12)
166
0.88
(8.6,0.19)
190
1
0.92
(9.5,0.19)
216
0.91
(7.5,0.15)
214
0.93
(7.3,0.16)
174
0.89
(10.1,0.21)
203
0.96
(4.6, 0.09)
210
1
0.95
(6.8,0.14)
202
0.91
(6.8,0.12)
202
0.96
(4.3,0.10)
168
0.92
(7.5,0.17)
192
0.94
(4.6,0.10)
580
0.97
(4.8,0.11)
210
1
0.93 0.89
(5.0,0.13) (12.8,0.23)
203 205
0.91
(6.9,0.13)
201
0.94
(5.0,0.11)
164
0.86
(11.9,0.19)
202
0.91
(11.7,0.19)
166
0.90
(9.6,0.18)
200
0.90
(7.7,0.13)
194
0.88
(8.9,0.17)
162
0.90 0.84 0.82
(6.5, 0.1 5) (1 3.8, 0.25) (1 1 .8, 0.22)
190 193 190
0.94 0.89
(5.9,0.13) (10.8,0.15)
200 196
0.90
(9.1,0.17)
207
0.90
(8.7,0.13)
210
0.94 0.92
(6.4,0.12) (10.4,0.16)
204 201
1
0.86
0.93
(5.0,0.10)
192
0.91
(6.6,0.12)
204
0.91
(7.1,0.12)
192
0.87
0.90
(7.7,0.13)
193
1
Figure 3A-11. Detroit MI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3 A-15
-------
a.
Grand Rapids, Ml MSA
Site A
SiteB
SiteC
SiteD
AIRS Site ID
26-005-0003
26-081-0020
26-121-0040
26-139-0005
b.
B
Mean
Obs
SD
50—1
40 —
E
TO
_a
in
pj
20 —
10 —
0—I
12.2
1036
8.4
14.0
1050
B.6
12.1
378
8.3
13.3
351
8.5
34 12
Quarter
c.
Site
A
B
C
D
A B
1 0.94
(5.7,0.15)
998
1
C
0.93
(4.4,0.13)
357
0.93
(6.1,0.16)
359
1
D
0.95
(4.4,0.12)
337
0.98
(3.1,0.10)
332
0.94
(5.2,0.13)
324
1
Figure 3A-12. Grand Rapids, MI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3 A-16
-------
Milwaukee, Wl MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
55-079-0010
55-079-0026
55-079-0043
55-079-0051
55-079-0059
55-079-0099
55-133-0027
55-133-0034
i
50
b.
B
H
Mean
Obs
SD
80-
1 60-
U)
•? «-
Q.
20-
0-
1
j
4.
05
B.£
"I
D
5
'I
I
1
1
!
3.
01
3.5
l|
2
2
!
14.4
326
9.0
I
!
f
13.1
332
8.4
!
!
14.1
342
8.2
!!!
!
14.1
356
9.1
I
|!
|l
1
8
a
I
'I
3.8
11
.3
'I
'I
13.1
349
8.3
!IT
1234 1234 1234 1234 1234
Quarter
1234 1234 1234
C. Site
A
B
C
D
E
F
G
H
A B C D
1 0.91 0.84 0.86
(3.3,0.11) (4.0,0.12) (3.0,0.11)
974 319 321
1 0.95 0.95
(3.7,0.12) (3.9,0.12)
311 315
1 0.94
(4.4, 0.13)
285
1
E
0.97
(2.9,0.10)
332
0.96
(3.1,0.11)
324
0.96
(3.8, 0.09)
284
0.96
(3.9,0.12)
306
1
F
0.87
(3.4, 0.09)
345
0.97
(3.5,0,10)
337
0.97
(3.4, 0.09)
296
096
(4.5,0.12)
327
0.97
(3.5, 0.09)
318
1
G
0.91
(3.3,0.10)
782
0.96
(3.6,0.12)
749
0.94
(4.0,0.12)
315
0.97
(3.5,0.10)
319
0.96
(3.4, 0.09)
325
0.97
(3.7,0.10)
335
1
H
0.82
(3.4,0.11)
338
0.91
(3.6,0.13)
330
0.89
(5.3,0.14)
293
0.92
(3.6,0.11)
318
0.93
(3.6,0.12)
317
0.92
(5.0,0.12)
335
0.93
(2.8,0.10)
330
1
Figure 3A-13. Milwaukee, WI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (jig/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3 A- 17
-------
Chicago, IL MSA
a.
c.
"^^^s
^
0
Mean
Obs
so
60-
s
a 40-
IO
Q.
20-
n
~ } AIRS Site ID
\\ Site A 17-031-0022
^^^^oMSHsjrTX SiteB 17-031-0052
"^qp %\ SiteC 17-031-0057
— 3t3lilk SiteD 17-031-0076
L»33=JR5>I SiteF 17-031-2001
JP&^ If 1 ojtp Q -17 n?1 T101
i p \ SiteH 17-031-4201
Jt — I Site 17-043-4002
^-J Site J 17-089-0003
50 100km SiteK 17-197-1002
ABCDEFGH JK
17.5 18.8 16.7 16.3 20.6 17.0 16.6 14.5 15.4 14.7 16.0
226 671 226 220 224 222 233 635 226 223 225
8.5 9.6 8.4 8.2 10.0 8.1 8.0 8.5 8.1 8.3 8.3
i i i i i
llllllllllllilll
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Quarter
Site ABCDEFGH JK
A 1 0.89 0.90 0.94 0.89 0.92 0.90 0.89 0.89 089 0.85
(7.0, 0,14) (5.3,0.12) (4,4, 0.10) (8.9, 0,15 (5,0. 0,11) (5.8, 0.12 (8.1,0,18) (7.0,0.16) (7.6,0.18) (6.6,0,16)
210 205 ' 202 211 201 208 196 197 178 173
R 1 0.90 091 0.90 0,88 0.89 0.80 0,87 090 0.81
(6.2,0.12) (6,6,0.13) (7.9,0,13) (7.9.0.13) 7.6,0.13) (9.8,0.19) 10.2,0.17) (9.3,0.18) (10.1,0,17
210 208 212 209 217 584 212 205 20B
p 1 0,93 0.92 0.92 0.90 0.92 0.90 0,91 0.89
(39,0.09) (7.7,014) (5.1,0.10) (5.2,0.10) (6.2,0.14) (6.3,0.12) (6.2.013) (68,012)
201 204 197 206 199 189 175 173
n 1 0.92 096 0.95 0.93 0,94 094 0.93
(9.0, 0.14) (3.5. 0.07) (3.5, 0.07) (8.5,0.14) (5.0,0.10) (5.5,0.13) (5.0,0.10)
201 198 205 191 188 181 172
C 1 092 0.90 0.91 0.90 092 0.87
(9.1,0,14 (9.6,0,14) (10,8, 0,221(10.7, 0,18)111.3, 0.20)(10.8. 0,18)
199 206 195 192 177 174
p 1 0.93 0.91 0.94 0,92 0.90
4.5,0.09) (7,0,0.16) (4.9,0.11) (6.7,014) (58,012)
204 190 188 172 167
Q 1 0.81 094 092 0-89
(6.7,0.15 4.8,010) (60014) (5.5.012)
200 193 178 173
H 1 0.92 095 0.85
(6.5,0.13) (5.0,0. 1) (6.6.015)
198 195 197
1 095 0.94
(4.2, 0.10) (3.9. 0 09)
168 164
J 1 093
(55,011)
153
Figure 3A-14. Chicago, IL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3 A-18
-------
a.
Gary, IN MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
18-089-0006
18-089-0022
18-089-1016
18-127-0024
B
Mean
Obs
SD
80—1
60 —
in
cs
40 —
20 —
0—1
15.7
907
8.5
17.6
927
10.1
16.2
965
8.3
14.0
312
8.0
1234 1234 1234 1234
Quarter
C. Site
B
D
A
B
C
D
1 0.65
(9.1,0.19)
823
1
0.93
(4.2,0.13)
822
0.59
(11.3,0,20)
841
1
0.91
(5.0,0.17)
277
0.56
(10.9, 0.23)
277
0.92
(6.4,0.17)
281
1
Figure 3A-15. Gary, IN MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations (ug/m3) for 2000-2001;
(c) Intersite correlation statistics for each data pair given as the correlation
coefficient, (P90, COD), and number of measurements.
3 A-19
-------
Louisville, KY MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
18-019-0005
18-043-1004
21-029-0006
21-111-0044
21-111-0048
B
Mean
Obs
SD
80-
co
"Bl
20-
O-l
17.4
331
8.7
15.7
296
8.3
16.1
323
7.8
17.2
1011
8.6
17.0
315
8.5
1234 1234
1234
Quarter
1234 1234
c.
Site A
B
D
A
B
C
D
E
1 0,88 0.90
(5.8,0.13) (6.0,0.12)
279 289
1 0.87
(6.0,0.15)
260
1
0.91
(3 .9, 0.10)
315
0.89
(5.9. 0.14)
282
0.93
(5-4, 0.12)
307
1
0.91
(3.8, 0.08)
273
0,91
(5.3,0.13)
250
0.90
(5.6,0.12)
256
0.93
(4.2,0.11)
302
1
Figure 3A-16. Louisville, KY MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM2 5 concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-20
-------
St. Louis, MO MSA
AIRS Site ID
2% LM\ Site A
s^s====5==— 1 Site C
X ^ -^T^s. .-ml*"- C!l« Pt
,f\ ~p~~r- Site D
X. Site E
13 \ Site F
X^J Site G
Site H
i i i
0 50 100km Site I
SiteJ
SiteK
17-119-0023
17-119-1007
17-119-2009
17-119-3007
29-099-0012
29-183-1002
29-189-2003
29-189-5001
29-510-0085
29-510-0086
29-510-0087
b.
Mean 2O2
60-i :—
15.8
7A9
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Quarter
c.
Site
A 1
B
C
D
E
F
G
H
'
J
K
0.81 0.73 0.74 0.67
(9.8,0.15) (12.2,0.20) (11.9,0.20) (14.6,0.22)
207 195 207 218
1 0.85 0.86 0.84
(6.8,0.12) (6.7,0.12) (7.3,0.14)
195 210 217
1 0.86 0.86
(4.8,0.13) (5.9,0.12)
196 209
1 0.83
(5.8,0.14)
222
1
0.71 0.68 0.69 0.69
(14.0,0.23) (14.6,0.23) (15.2,0.25) (14.1,0.21)
215 214 217 221
0.84
(7.8,0.16)
216
0.94
(4.0,0.11)
206
0.86
(5.7,0.15)
221
0.90
(5.3,0.11)
233
1
0.79
(7.5,0.16)
215
0.84
(6.3,0.13)
207
0.81
(6.1,0.15)
219
0.90
(4.0,0.11)
231
0.87
(4.8,0.12)
230
1
0.84
(8.3,0.17)
217
0.90
(5.7,0.13)
208
0.86
(6.1,0.15)
222
0.93
(4.1,0.11)
234
0.94
(3.8, 0.09)
233
0.92
(2.8,0.10)
232
1
0.86
(6.7,0.13)
226
0.91
(5.2,0.11)
228
0.86
(5.3,0.14)
229
0.95
(3.7, 0.08)
234
0.93
(5.1,0.11)
233
0.90
(3.7,0.11)
231
0.95
(4.3,0.10)
234
1
0.69
(15.0,0.23)
207
0.85
(7.6,0.15)
211
0.91
(5.4,0.12)
214
0.87
(5.5,0.14)
216
0.95
(3.8, 0.09)
218
0.93
(4.7,0.10)
218
0.90
(3.1,0.10)
215
0.95
(2.9, 0.09)
218
0.98
(2.8, 0.07)
671
1
0.68
(13.9,0.20)
217
0.85
(6.5,0.12)
221
0.89
(5.0,0.11)
224
0.84
(5.5,0.14)
225
0.93
(4.3, 0.09)
230
0.91
(5.4,0.12)
229
0.87
(5.3,0.12)
227
0.94
(4.9,0.12)
230
0.97
(2.3, 0.06)
691
0.96
(3.3, 0.08)
652
1
Figure 3A-17. St. Louis, MO MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-21
-------
Baton Rouge, LA MSA
AIRS Site ID
Site A
SiteB
SiteC
22-033-0002
22-033-0009
22-121-0001
b.
B
Mean
Obs
SD
50-
M
30 —
20 —
10 —
14.5
331
7.1
14.5
1067
6.6
14.1
1031
6.6
1234
1234
Quarter
1234
C. Site
B
A
B
1 0.93
(2.7, 0.08)
326
1
0.93
(2.9, 0.09)
318
0.97
(2.5, 0.07)
1006
Figure 3A-18. Baton Rouge, LA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-22
-------
Kansas City, MO MSA
B
Mean
Obs
SD
50-
£
ro
40-
30-
20-
10-
0-
11.7
330
6.0
11.4
336
6.0
11.6
341
6.2
12.9
1033
6.4
12.4
988
6.2
AIRS Site ID
Site A
SiteB
SiteC
Site D
SiteE
SiteF
20-091-0008
20-091-0009
29-047-0005
29-047-0026
29-047-0041
29-095-2002
13.8
354
6.7
1234 1234 1234 1234 1234 1234
Quarter
C.
Site A
B
D
A
B
D
E
F
0.94 0.92
(1.9,0.09) (4.2,0.12)
320 304
1
0.90
(4.0 0.13)
312
1
0.90 0.93 0.89
(4.3,0.13) (3.8,0.10) (5.9,0.15)
318 296 320
0.89 0.90 0.87
(4.1,0.14) (3.6,0.12) (6.2,0.17)
322 ' 300 326
0.96 0.96 0.90
(3.1 0.09) (2.8,0.10) (6.5,0.16)
327 300 329
1 0.95 0.95
(2.9,0.09) (4.0,0,11)
940 338
1
0.94
(5.1,0.12)
314
1
Figure 3A-19. Kansas City, KS-MO MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-23
-------
a.
Dallas, TX MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
48-085-0005
48-113-0020
48-113-0035
48-113-0050
48-113-0057
48-113-0069
48-113-0087
B
Obs
SD
"5)
10 20-
11.5
234
5.6
12.4
677
5.7
12.9
222
5.8
13.3
644
5.8
13.7
215
6.1
12.7
687
5.7
11.7
216
5.4
1234 1234
1234
Quarter
1234 1234 1234
c.
Site
A
B
C
D
E
F
G
1 0.92 0.94 0.94
(3.5.0.11) (3,6,0.11) (4.3,0.13)
220 204 213
1 0.95 0.94
(3.2, 0.08) (3.3, 0.09)
212 803
1 0.97
(2.0, 0.06)
203
1
0.89
(6.3, 0.15)
195
0.92
14.1, 0.11)
205
0.93
(3.9, 0.09)
191
0.94
(2.7, 0.08)
199
1
0.94
(3.7,0.10)
218
0.95
(2.5, 0.07)
635
097
(1.9,0.06)
207
0.98
(2.2, 0.06)
808
0.95
(3.1,0.08)
198
1
0,94
(3.1, 0.10)
189
087
(2,3, 0.071
207
094
(3.8, 0.10)
185
0.94
(4.2, 0.12)
198
091
(5,5, 0.141
182
0,66
(3,0, 0.091
198
1
Figure 3A-20. Dallas, TX MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-24
-------
Boise, ID MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
16-001-0011
16-001-0017
16-027-0004
16-027-0005
T
100k
B
Mean
Obs
SD
80 —
60 —
CO
E
TO
10 40 —
CM
V
Q.
20 —
0 —
!
9.7
602
9.0
''
|
|
8.7
318
7.6
j
I
T I
9.6
358
9.1
y
10.3
328
9.4
J
1234 1234 1234
Quarter
c.
Site
A
B
D
A
B
C
D
1 0.91
(4.3, 0.16)
304
1
0.95
(3.8, 0.13)
344
0.85
(6.0, 0.19)
313
1
0.92
(5.1,0.16)
315
0.78
(8.8, 0.23)
317
0.96
(3.9, 0.12)
322
1
Figure 3A-21. Boise, ID MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations (ug/m3) for 2000-2001;
(c) Intersite correlation statistics for each data pair given as the correlation
coefficient, (P90, COD), and number of measurements.
3A-25
-------
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
B
Mean
Obs
SD
80-
«""
E
"oi 6°-
3.
10
CNJ
Q. 40-
20-
0-
,
1
'
,
'!
i.O
Si
3.2
••''
,
•\
1
|
-
,
f.
2.
(49
2.;
. j
1
'1
1
3
1
ill
'ITi
j.6
46
1.7
:•
1 1
i
f
1
1
1
9
1
,
H
.3
J3
.7
1
i i
|
!1
11.8
333
11.4
I I
'Til
I
1
1
1
1
1
t
;
,
|!
i.8
28
'.4
!!
!!
1234 1234 1234 1234
Quarter
c.
Site
B
D
1234 1234
A
B
C
D
E
F
1 0.89 0.90
(8.8,0.22) (9.9,0.27)
334 331
1 0.89
(7.0,0,19)
327
1
0.94
(6.9,0.19)
322
0.92
(4,9,0.17)
316
0.93
(7-1,0.19)
314
1
0.90
(8.2, 0.22)
315
0.89
(6.1,0.17)
313
0.96
(4.8, 0.17)
308
0.94
(5.3, 0.16)
300
1
0.94
(4.4, 0.15)
306
0.88
(9.3, 0.20)
307
0.86
(11.4,0.24)
302
0.92
(8.7, 0.18)
297
0.89
(9.6, 0.20)
288
1
Figure 3A-22. Salt Lake City, UT MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-26
-------
a.
Seattle, WA MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
53-033-0017
53-033-0021
53-033-0057
53-033-0080
53-061-1007
b.
B
Mean
Obs
SD
—. 40-
£
"3)
10 30-
20-
10-
0—1
5.8
316
3.5
10.9
1057
7.3
11.9
1054
7.0
8.9
808
5.2
11.4
357
8.6
C. Site
1234 1234
B
1234
Quarter
1234 1234
D
A
B
C
D
E
1 0.31 0.28
(15.0. 0.39) (15.7,0.43)
303 304
1 0.96
(3.6.0.14)
1021
1
0,45
(8.8, 0.32)
291
0.92
(6.8,0.16)
774
0.91
(7.5, 0.20)
779
1
0.37
(17.9, 0.39)
298
0.81
(6.2,0.17)
344
0.79
(8.2, 0.20)
344
0.75
(8.5, 0,20!
327
1
Figure 3A-23. Seattle, WA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-27
-------
Portland, OR MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
41-009-0004
41-051-0080
41-051-0244
41-067-0111
B
in
c-j
Mean
Obs
SD
125-
100-
75-
50-
25-
0-
6.3
457
4.4
I
4J'
j
9.1 8.7 7.3
1042 1037 503
7.2 5.4 5.9
j
l|.
It U
1234 1234 1234 1234
Quarter
C.
Site
B
D
A
B
C
D
1 0.79
(6.5, 0.24)
433
1
0.90
(5.1,0.20)
429
0.77
(4.1,0.14)
986
1
0.81
(4.5,0.19)
427
0.89
(4.3, 0.17)
477
0.83
(4.5,0.19)
472
1
Figure 3A-24. Portland, OR MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-28
-------
a.
Los Angeles, CA MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
06-037-0002
06-037-1103
06-037-1201
06-037-1301
06-037-2005
06-037-4002
B
Mean
Obs
SD
100-
50 —
25 —
0—I
20.9
641
13.3
22.5
656
13.5
18.3
217
11.7
23.7
237
13.4
20.1
220
11.9
20.3
621
12.4
nmpi
1234 1234
c.
Site
A
B
1234 1234
Quarter
C D
1234 1234
A
B
C
D
E
F
1 0.87 0.76
(10.7,0.18) (14,6,0.23)
581 208
1 0.86
(12.8, 0,20)
205
1
0,68
(17.9, 0,25)
229
0.89
(10.1, 0,12)
222
0.76
(18.1, 0.24)
212
1
0.95
(6.2,0,14)
212
0.93
(7.1,0,11)
207
0.85
(12.1,0.18)
197
0.78
(13.2,0.18)
214
1
0.80
(18,1,0.26)
553
0.80
(13.6, 0.17)
563
0.66
(18.2, 0.24)
197
0.95
(8.1, 0.11)
216
0.62
(15, 0.20)
204
1
Figure 3A-25. Los Angeles-Long Beach, CA MSA. (a) Locations of sampling sites by AIRS
ID#; (b) Quarterly distribution of 24-h average VM2S concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-29
-------
Riverside, CA MSA
Site A
SiteB
SiteC
SiteD
SiteE
AIRS Site ID
06-065-1003
06-065-8001
06-071-0025
06-071-2002
06-071-9004
B
Mean
Otis
SO
40-
26.9
327
16.1
30.0
766
18.0
25.4
320
14.6
25.0
347
14.9
1234 1234
1234
Quarter
c.
Site A
B
D
25.7
307
16.4
1234 1234
A
B
C
D
E
1 0.94 0.83
(6.6, 0.10) (14.3,0.21)
294 ' 289
1 0.81
(17.8, 0.23)
290
1
0.93
(10.6,0.13)
306
0.93
(13.3,0.14)
313
0,86
(11,8,0.20)
302
1
0.90
(10.5.0.13)
278
0.91
(11.9.0.13)
275
0,78
(16.9.0.22)
269
0.94
(8.9,0.11)
290
1
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
(ug/m3) for 2000-2001; (c) Intersite correlation statistics for each data pair
given as the correlation coefficient, (P90, COD), and number of
measurements.
3A-30
-------
San Diego, CA MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
06-073-0001
06-073-0003
06-073-1002
06-073-1007
50
100km
b.
B
Mean
Obs
SD
80—|
60 —
"5)
0—1
14.6
313
7.2
16.5
938
8.1
17.0
886
9.2
16.8
879
9.3
III!
I
1234
1234 1234
Quarter
1234
c.
Site A
B
D
A
B
C
D
1 0.76 0.73
(10.0,0.16) (10.0,0.19)
270 253
1 0.85
(6.3,0.13)
773
1
0.83
(7.6,0.16)
255
0.78
(9.7,0.18)
769
0.73
(11.0,0.20)
728
1
Figure 3A-27. San Diego, CA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM2 5 concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation statistics for each data pair given
as the correlation coefficient, (P90, COD), and number of measurements.
3A-31
-------
Columbia, SC MSA
Site A
Site B
b.
B
Mean
Obs
SD
40 H
30-
n
i
-
10-
0-
-10—1
7.4
56
4.3
9.6
53
6.2
1234 1234
Quarter
C.
Site
B
A
B
1 0.70
(8.0, 0.37)
49
AIRS Site ID
45-063-0005
45-079-0019
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 (ug/m3)
for 2000; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-32
-------
Tampa, FL MSA
AIRS Site ID
Site A
SiteB
\
50
A
B
Mean
Obs
SD
50-1
40-
30-
in
cvi
o
20-
10-
0—I
11.3
112
6.6
10.1
104
5.2
1234 1234
Quarter
c.
Site
A
B
B
0.81
(5.3,0.17)
95
1
12-057-0030
12-103-0018
Figure 3A-29. Tampa, FL MSA (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-33
-------
Cleveland, OH MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
39-035-0013
39-035-0038
39-035-0045
39-035-0060
39-035-0065
39-085-1001
A
D
Mean
Obs
SO
150-
100-
1
3.
"0 50-
oj
o
Q_
0-
-50-
26.4
>16
17.4
!
.
1
I'
11
j
234
18.6
614
11.6
16.8
II
I:
1
i
1234 1
112
9.0
ii
i i
i
"
i
r
234
21.3
113
16.4
I
!ii
16.B
111
9.5
I
|ll
7.2
109
4.9
TTTf
1234 1234 1234
Quarter
c.
Site
D
A
B
C
D
E
F
1 0.67 0.67
(23,2. 0 26) (28.5, 0.28)
182 95
1 0.65
(16.1,0.22)
90
1
0.73
(17,9, 0,22)
97
0.73
(11.7, 0,18)
93
0.86
(18,1,0,23)
94
1
0.62
(27,0, 0.31)
98
0,69
(13,9, 0,62)
90
0.71
(10,6,0,31)
102
0.74
(15.4, 0.38)
93
1
0.41
(40,0, 0,60)
94
0.44
(24.9, 0,53)
89
0.4B
(19,9, 0 50)
99
0.31
(28,0, 0.59)
94
0.22
(20.4, 0,55)
99
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 (ug/m3)
for 2000-2001; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-34
-------
Steubenville, OH MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
39-081-0016
39-081-1001
54-009-0005
54-029-1004
50km
B
D
Mean
Obs
SD
100-1
TO
75-
50-
25-
-25-I
12.7
107
8.6
14.3
190
10.9
10.2
211
6.3
13.0
114
10.4
1234 1234
Quarter
c.
Site
B
D
A
B
C
D
1 0.64 0.69
(10.9,0.77) (11.6,--)
83 91
1 0.54
(14.7, -)
166
1
0.68
(11.3,-)
100
0.48
(18.5, -)
88
0.69
(12.8, -)
97
1
Figure 3A-31. Steubenville, OH MSA (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-35
-------
Detroit, Ml MSA
AIRS Site ID
Site A
SiteB
SiteC
26-163-0001
26-163-0015
26-163-0025
b.
B
Mean
Obs
SD
80-
60-
n
E
o ^~
_3.
>o
tvi
o
S"
°- 20-
0-
-20-
11.5
56
10.3
'
I
'.|
19.4 7.3
58 55
15.6 7.6
,
I
jj
i I
ill
1234 1234 1234
Quarter
C. Site
B
A
B
C
1 0.58
(29.4, -)
53
1
0.54
(15.7. 0.54)
50
0.39
(34.9, 0.79)
51
1
Figure 3A-32. Detroit, MI MSA (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2000; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-36
-------
Milwaukee, Wl MSA
Site A
SiteB
AIRS Site ID
55-079-0059
55-133-0027
b.
c.
Mean
Obs
SD
son
—. 30-
n
~
20-
10-
0-
-10-1
Site
7.9
160
5.7
B
9.1
175
7.4
1234 1234
Quarter
A
B
B
0.65
(9.2, 0.53)
150
1
Figure 3A-33. Milwaukee, WI MSA (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 1999-2001; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-37
-------
Chicago, IL MSA
AIRS Site ID
Site A
SiteB
SiteC
17-031-1016
17-031-2001
17-031-3301
50
100km
b.
A
B
Mean
Obs
SD
100 H
75-
co
0-
-20 -I
14.7
163
19.0
12.8
109
10.7
16.1
112
12.2
1234
1234
Quarter
234
c. Site
A
B
A
B
C
1 0.68
(20.0, --)
97
1
0.53
{24,6, -)
101
0.82
(11,1,0.40)
103
1
Figure 3A-34. Chicago, IL MSA (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 1999-2000; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-38
-------
Gary, IN MSA
a.
b.
A
Mean
Obs
SD
30 n
n
~
20-
10-
10
rsi
o
Q. o-
-10-
-20-I
3.9
54
5.6
AIRS Site ID
Site A
SiteB
SiteC
18-089-0006
18-089-1016
18-127-0024
B
5.1
55
6.4
c.
Site
3.4
56
3.9
1234 1234 1234
Quarter
ABC
A
B
C
1 0.79
(7.8, -5
49
1
0.63
(6.3, -)
49
0.60
(8.0, 0.83}
50
1
Figure 3A-35. Gary, IN MSA (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM10_25 concentrations (ug/m3) for 2001;
(c) Intersite correlation coefficients, (P90, COD), and number of
observations.
3A-39
-------
Louisville, KY MSA
Site A
SiteB
i
50
A
B
c.
Mean
Obs
SD
30-
20-
co
E
"3)
«> 10-
CM
O
Q.
0-
-10-
9.1 7.6
55 51
5.0 4.0
'
I
I
J
I '
1234 1234
Site
A
R
Quarter
A B
1 0.65
(5.5,
0.48)
46
1
AIRS Site ID
18-019-0005
21-029-0006
Figure 3A-36. Louisville, KY MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2001; (c) Inter site correlation coefficients, (P90, COD), and number
of observations.
3A-40
-------
St. Louis, MO MSA
AIRS
ID
Site A
SiteB
SiteC
17-119-0023
17-119-3007
17-163-0010
b.
B
Mean
Obs
SD
100-1
75-
1O
cvi
o
25-
0-
-25—1
22.5
105
17.4
12.1
57
13.7
15.5
52
14.2
1234
1234
Quarter
1234
c.
Site
B
A
B
C
1 0.70
(27,2, -)
51
1
0.73
(26.2, 0.76)
47
0.82
(13,0,0,91)
50
1
Figure 3A-37. St. Louis, MO MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2000; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-41
-------
Baton Rouge, LA MSA
Site A
SiteB
\
50
A
B
Mean
Obs
SD
60 n
40-
it)
t
-------
Dallas, TX MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
48-113-0020
48-113-0035
48-113-0050
48-113-0057
b.
B
D
Mean
Obs
SD
50 -t
E
ro
Q.
40-
30-
20-
10-
O-l
11.2
60
5.4
12.9
55
6.7
14.5
56
6.4
19.1
55
10.5
1234 1234 1234 1234
Quarter
Site
B
D
A
B
C
D
1 0,79
(4.5, 0.17)
54
1
0.71
(9.3, 0.22)
55
0.69
(7.8,0.18)
50
1
0.66
(16.5, 0.32)
54
0.60
(13.2, 0.30)
50
0.69
(13.5, 0.24)
50
1
Figure 3A-39. Dallas, TX MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2001; (c) Inter site correlation coefficients, (P90, COD), and number
of observations.
3A-43
-------
Salt Lake City, UT MSA
AIRS Site ID
Site A
SiteB
SiteC
49-035-0003
49-035-0012
49-035-3006
b.
B
Mean
Obs
SD
150—1
100-
0-
-50—1
14.8
315
9.8
27.5
327
18.3
15.3
910
9.7
T
1234 1234
Quarter
1234
c.
Site
A
B
C
1 0.72
(28.7, -)
283
1
0.74
(9.8, -)
264
0.70
(27.6, 0.47)
274
1
Figure 3A-40. Salt Lake City, UT MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 1999-2001; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-44
-------
Portland, OR MSA
A
Mean
Obs
SD
40 H
30-
"5)
20-
10-
0-
-10-1
5.7
320
3.5
Site A
SiteB
B
6.7
113
4.3
1234 1234
Quarter
c.
Site
B
A
B
0.69
(5.1,-)
107
1
AIRS Site ID
41-051-0080
53-011-0013
Figure 3A-41. Portland, OR MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-45
-------
Los Angeles, CA MSA
a.
Site A
SiteB
SiteC
SiteD
b.
A
B
Mean
Obs
SD
60 H
E
ro
0.
50-
40-
30-
20-
10-
O-l
24.1
56
11.7
15.3
56
6.7
21.4
57
8.7
C. Site
B
AIRS Site ID
06-037-0002
06-037-1002
06-037-1103
06-037-4002
D
16.1
53
6.6
1234 1234 1234 1234
Quarter
D
A
B
C
D
1 0.82 0.63
(19.0,0.24) (15.5,0.18)
49 49
1 0.74
(11.5,0.21)
49
1
0.58
(17.3,0.27)
45
0.54
(11.5,0.25)
47
0.57
(12.5, 0.22)
45
1
Figure 3A-42. Los Angeles, CA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2001; (c) Inter site correlation coefficients, (P90, COD), and number
of observations.
3A-46
-------
Riverside, CA MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
06-065-2002
06-065-8001
06-071-0025
06-071-2002
0 50 100km
D.
A
B
D
C.
Obs
SD
500-
400-
n
ra
tN
° 200-I
100-
0-
100
210
46.3
32.1
200
20.4
25.5
108
15.4
26.3
112
15.4
1234 1234 1234 1234
Quarter
Site A
B
D
A
B
C
D
1 0.38 0.32
(42.6, 0.38) (36.6, 0.39)
184 102
1 0.62
(25.9, 0.32)
104
1
0.45
(39.0, 0.38)
107
0.79
(18.2, 0.33)
108
0.80
(13,3, 0.28)
98
1
Figure 3A-43. Riverside, CA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2000-2001; (c) Intersite correlation coefficients, (P90, COD), and number
of observations.
3A-47
-------
San Diego, CA MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
06-073-0001
06-073-0003
06-073-1006
06-073-1002
50
100km
b.
c.
A
B
D
Mean
Obs
SD
50-
40-
30-
iq 20-
rj
0-
10-
13.3
55
4.0
19.4
51
6.2
11.6
57
5.1
12.9
56
7.7
1234 1234 1234 1234
Quarter
Site
B
D
A
B
C
D
1 0.38 0.01
(14.4, 0.25) (9.5, 0.34)
46 51
1 0.65
(12,6, 0.28)
46
1
0,09
(11.7,0.32)
50
0.63
(14.7, 0.34)
46
0.70
(8.3, 0.45)
51
1
Figure 3A-44. San Diego, CA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations (ug/m3)
for 2001; (c) Inter site correlation coefficients, (P90, COD), and number
of observations.
3A-48
-------
APPENDIX 3B
Aerosol Composition Data from the Speciation Network
The U.S. Environmental Protection Agency (EPA), working with state and local air quality
agencies, began implementing an ambient air monitoring network in 1999/2000 to provide a
consistent data set for the characterization and evaluation of trends in PM components (chemical
species). The network was designed to include about 52 core trends sites across the United
States and to provide a stable ongoing national perspective. In 1999, an initial thirteen sites were
installed and operated to be used as a model for the deployment of the more comprehensive
network consisting of the 52 core trends sites and roughly 200 additional locally relevant sites.
Data from the initial 13 sites were presented in the Third External Review Draft of this
document, released by the EPA in 2001 for public comment and CAS AC review. These sites
were designed and operated to evaluate the suitability of various aerosol sampling devices for
obtaining PM2 5 composition data, and data from those sites were summarized in Appendix 3B of
the 2001 Third External Review Draft. Three types of collocated aerosol sampling devices were
used to collect the data from February through July 2000, and data obtained from the three
sampling devices were shown for each site. A complete description of the data, techniques used
to analyze the filters, and the results of the evaluation of the performance of the sampling
devices (including a number of caveats regarding the data) can be found in Coutant and Stetzer
(2001) and the analyses of data in Coutant et al. (2001).
More recent measurements of ambient levels of PM25 constituents are presented in the
tables provided below. Table 3B-1 presents the locations and general sampling schedules for
13 monitoring sites included for various reasons and representing a cross section of the country.
Many of these sites were chosen because they are located in (or close to) MS As in which risk
assessments are to be performed. Sacramento, CA and Riverside, CA, for example, are located
near San Francisco and Los Angeles, respectively. The data reported here are from the period
October 2001 to September 2002. For this time period, a total of 51 sites (including both
"Trends" and "non-Trends" sites) across the country have complete data (as defined by 50% of
3B-1
-------
observations available for every quarter for each of the major chemical species: sulfate,
ammonium, nitrate, elemental carbon, organic carbon, and the five trace elements that go into
the calculation of the crustal contribution to PM25 (Al, Fe, Ca, Ti, and Si).
Summary statistics for concentrations of PM25 and chemical species are given for each of
the 13 sites in Table 3B-2. The number of samples (n) and the AIRS site code are given above
each table. Entries in the tables provide the mean, minimum, and maximum component
concentrations, and minimum detection limits for each component. Minimum detection levels
(MDL) differ among the various sampling methods; these limits were estimated by Research
Triangle Institute (RTI) in July of 2001 and are subject to review, revision, and reinterpretation.
Anions and cations (ammonium, nitrate, sodium, potassium, and sulfate) were determined by ion
chromatography; carbonaceous species were determined by the thermal optical transmittance
method (NIOSH method); and trace elements (aluminum through zirconium) were determined
by X-ray fluorescence spectrometry. The sulfate (calculated) entry is based on the XRF
determination of sulfur. In general, relatively good agreement is found between the
reconstructed mass and the PM2 5 concentration measured by the collocated FRM monitor at
each site. However, there are exceptions at several locations, as can be seen from inspection of
Table 3B-2. Somewhat different sampling trains are used in the different sampling systems.
These differences can result in differences in performance metrics and in differences in the
entities that are measured. All samplers use a denuder in front of the filter for species to be
analyzed by ion chromatography. In the RAAS and SASS samplers the denuder is followed by a
nylon filter, whereas in the MASS samplers (Chicago, Houston, Seattle) a teflon filter is
followed by a nylon filter. Particulate nitrate is collected on the teflon filter and is referred to as
nonvolatile nitrate. However, there may be volatilization of nitrate containing compound from
the teflon filter which is then adsorbed on the nylon filter. Nitrate extracted from the nylon filter
is referred to as volatile nitrate.
Organic carbon (OC) concentrations are multiplied by a factor of 1.4 when calculating
mass to account for the presence of H, N, and O in organic compounds on all samplers.
Carbonate carbon has never been detected in any of the samples. Field blank corrections that
could be applied to elemental carbon (EC) and OC concentrations are shown in Table 3B-3 for
different samplers. Blank corrections for OC and EC shown in Table 3B-3 were applied to
concentrations shown in Table 3B-2. However, subtracting blank corrections from OC
concentrations results in negative values in several cases. A possible cause for results such as
3B-2
-------
these could have been that not enough blank samples were obtained to fully characterize the
blank levels. Although the concentrations of 47 elements could be obtained by X-ray
fluorescence spectrometry, the concentrations of many of these elements are beneath MDLs and
are not shown. The same elements shown in Appendix 6 A of the 1996 PM AQCD are shown
here. The usual practice of denoting table entry values below MDL by (—) is followed here.
Missing data for PM2 5 are also indicated by (—). Often, environmental data below MDL are
represented by 0.5 x MDL for various purposes such as risk assessments and for mass closure.
Unfortunately, speciation data are not available for the PM10_2 5 size fraction.
REFERENCES
Coutant, B.; Stetzer, S. (2001) Evaluation of PM25 speciation sampler performance and related sample collection and
stability issues: final report. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards; report no. EPA-454/R-01-008. Available:
http://www.epa.gov/ttn/amtic/pmspec.html [5 April, 2002].
Coutant, B.; Zhang, X.; Pivetz, T. (2001) Summary statistics and data displays for the speciation minitrends study:
final report. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards; contract no. 68-D-98-030.
3B-2
-------
TABLE 3B-1. PM2 s SPECIATION SAMPLERS BY LOCATION: SITES SELECTED
FOR SUMMARY IN 2004 PM AQCD
Sampler
Type
SASS
SASS
RAAS
SASS
MASS
RAAS
MASS
SASS
SASS
SASS
MASS
SASS
SASS
Location_Name
Burlington
Philadelphia
(AMS Laboratory)
Atlanta
(South DeKalb)
Detroit (Allen Park)
Chicago (Com ED)
St. Louis (Blair Street)
Houston (Deer Park)
Minneapolis (Philips)
Boulder
(Commerce City)
Phoenix (Supersite)
Seattle (Beacon Hill)
Sacramento
(Del Paso Manor)
Riverside-Rubidoux
State
VT
PA
GA
MI
IL
MO
TX
MN
CO
AZ
WA
CA
CA
Began
Operation
12/14/2000
9/10/2001
3/2/2001
12/14/2000
5/22/2001
2/9/2000
2/9/2000
12/14/2000
2/18/2001
2/21/2000
2/9/2000
2/9/2000
5/13/2001
AIRS_Code
000500070012
000421010004
000130890002
000261630001
000170310076
000295100085
000482011039
000270530963
000080010006
000040139997
000530330080
000060670006
000060658001
POC
5
7
5
5
5
6
37778
5
5
7
6
5
37746
Reason for
Schedule Inclusion*
l-in-3
l-in-3
l-in-3
l-in-3
Alt l-in-3
l-in-3
l-in-3
Alt l-in-3
Alt l-in-3
l-in-3
l-in-3
l-in-3
l-in-3
3
3
2
1
3
1
3
2
3
1
3
2
1
* Reason for Inclusion:
1 = Completeness of geographic coverage
2 = PM10 risk assessment city
3 = PM2 5 and PM10 risk assessment city
3B-4
-------
TABLE 3B-2a. BURLINGTON, VT SUMMARY DATA
(October 2001 to September 2002).
All concentrations are given in iig/m3; n = 201; AIRS Site Code: 500070012
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
10.0
10.9
2.6
2.8
1.1
0.22
0.035
1.3
—
—
0.26
—
2.5
0.020
—
—
0.002
—
0.035
—
0.002
0.002
0.041
—
—
0.002
—
0.002
—
0.041
—
—
0.069
0.10
—
0.88
—
0.004
—
0.008
Max Min
61.0 1.0
63.0 1.6
17.7 0.001
18.8 0.17
6.8 —
2.1 —
0.38 —
8.8 —
— —
— —
0.84 —
— —
31.4 -0.43
0.28 —
0.005 —
0.12 —
0.007 —
0.014 —
0.19 —
0.12 —
0.060 —
0.014 0.001
0.390 0.007
0.016 —
0.19 —
0.007 —
0.012 —
0.022 0.001
0.092 —
0.41 —
0.003 —
0.005 —
0.54 —
0.40 —
0.004 —
5.9 —
0.033 —
0.023 —
0.006 —
0.057 0.007
MDL
—
—
—
0.012
0.017
0.030
0.014
0.008
—
—
0.15
0.15
0.15
0.011
0.002
0.059
0.002
0.011
0.003
0.006
0.002
0.001
0.002
0.006
0.018
0.002
0.005
0.001
0.006
0.003
0.002
0.002
0.008
0.051
0.003
0.007
0.018
0.002
0.002
0.002
3B-5
-------
TABLE 3B-2b. PHILADELPHIA, PA SUMMARY DATA (October 2001 to
September 2002). All concentrations are given in ug/m3; n = 262;
AIRS Site Code: 421010004
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
14.2
16.0
4.3
4.4
2.0
0.21
0.042
2.1
0.84
0.61
0.66
—
3.1
0.019
—
—
0.004
—
0.037
0.011
0.002
0.004
0.084
0.005
—
0.002
—
0.006
—
0.053
—
0.002
0.086
0.087
—
1.45
—
0.006
0.004
0.015
Max Min
86.8 2.1
239.0 0.0
29.0 0.004
30.5 0.020
11.4 —
1.6 —
0.83 —
8.7 0.050
4.0 0.030
5.6 0.020
2.3 —
— —
50.5 -1.13
0.54 —
0.006 —
0.12 —
0.013 —
0.017 —
0.18 —
0.68 —
0.018 —
0.025 0.003
0.53 —
0.025 —
0.30 —
0.009 —
0.012 —
0.13 0.001
0.039 —
1.1 —
0.003 —
0.005 —
1.1 —
0.63 —
0.016 —
9.6 —
0.035 —
0.049 —
0.035 —
0.095 0.007
MDL
—
—
—
0.012
0.017
0.030
0.014
0.008
—
—
0.15
0.15
0.15
0.011
0.002
0.059
0.002
0.011
0.003
0.006
0.002
0.001
0.002
0.005
0.018
0.002
0.005
0.001
0.006
0.003
0.002
0.002
0.008
0.051
0.003
0.007
0.018
0.002
0.002
0.001
3B-6
-------
TABLE 3B-2c. ATLANTA, GA SUMMARY DATA (October 2001 to
All concentrations are given in ug/m3; n = 183; AIRS Site Code:
September 2002).
130890002
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Gold
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
—
16.3
4.8
4.8
1.3
0.27
0.044
0.70
—
—
0.90
—
4.3
0.028
0.001
—
0.003
—
0.037
0.003
—
0.002
—
0.084
0.003
0.008
0.002
—
—
—
0.060
0.005
0.001
0.11
0.056
0.001
1.6
—
0.006
0.001
0.008
Max
—
40.0
14.5
15.2
4.2
1.7
0.31
3.5
—
—
3.5
—
11.2
0.65
0.014
0.054
0.009
0.006
0.28
0.049
0.003
0.011
0.004
0.47
0.008
0.16
0.011
0.003
0.002
0.017
0.36
0.002
0.004
1.4
0.31
0.005
4.8
0.019
0.043
0.003
0.034
Min
—
4.7
0.27
0.88
—
—
—
0.16
—
—
—
—
0.8
—
—
—
—
—
0.006
—
—
—
—
0.013
—
—
—
—
—
—
0.015
—
0.000
0.010
—
0.001
0.089
—
—
—
0.001
MDL
—
—
—
0.011
0.015
0.028
0.013
0.008
—
—
0.13
0.13
0.13
0.004
0.001
0.024
0.001
0.004
0.001
0.002
0.001
0.001
0.002
0.001
0.002
0.007
0.001
0.002
0.001
0.003
0.001
0.001
0.001
0.003
0.021
0.001
0.003
0.007
0.001
0.001
0.001
3B-7
-------
TABLE 3B-2d. DETROIT, MI SUMMARY DATA (October 2001 to September 2002).
All concentrations are given in iig/m3; n = 189; AIRS Site Code: 261630001
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
16.6
18.0
4.4
4.6
2.2
0.27
0.061
3.1
—
—
0.68
—
3.2
0.025
0.002
—
0.003
—
0.069
0.017
0.002
0.006
0.12
0.006
0.020
0.004
—
0.002
—
0.078
—
0.002
0.11
0.098
—
1.5
—
0.007
0.002
0.025
Max
53.1
52.9
22.0
24.4
9.3
1.7
1.2
15.2
—
—
3.7
—
11.2
0.53
0.010
0.11
0.011
0.014
0.33
0.65
0.033
0.043
0.60
0.034
0.30
0.025
0.013
0.022
0.028
1.3
0.003
0.011
0.84
0.53
0.021
7.3
0.032
0.079
0.012
0.19
Min
2.6
3.2
0.010
0.040
0.009
0.015
0.007
0.004
—
—
—
—
-0.46
—
—
—
—
—
—
—
—
0.001
0.003
—
—
—
—
0.001
—
—
—
—
—
—
—
—
—
—
—
0.001
MDL
—
0.012
0.017
0.030
0.014
0.008
—
—
0.15
0.15
0.15
0.011
0.002
0.059
0.002
0.011
0.003
0.006
0.002
0.001
0.002
0.005
0.018
0.002
0.005
0.001
0.006
0.003
0.002
0.002
0.008
0.051
0.003
0.007
0.018
0.002
0.002
0.001
3B-8
-------
TABLE 3B-2e. CHICAGO, IL SUMMARY DATA (October 2001 to September 2002).
All concentrations are given in iig/m3; n = 139; AIRS Site Code: 170310076
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
15.7
15.0
4.3
4.2
1.9
0.052
0.068
2.0
0.71
1.3
0.61
—
2.8
0.03
0.001
—
0.003
—
0.058
0.017
0.001
0.004
0.091
0.006
0.017
0.003
—
0.001
0.002
0.085
—
0.001
0.11
0.048
0.001
1.4
—
0.004
0.001
0.023
Max
42.7
44.5
22.8
21.2
8.8
0.26
3.1
10.2
3.6
9.9
1.6
—
7.4
1.1
0.006
0.22
0.011
0.007
0.55
0.67
0.005
0.056
0.53
0.040
0.69
0.014
0.006
0.007
0.015
3.1
0.001
0.004
2.2
0.58
0.061
7.6
0.023
0.070
0.002
0.093
Min
3.6
3.4
0.64
0.56
0.17
—
—
0.12
0.04
0.03
—
—
0.31
—
0.001
—
—
—
0.10
—
—
—
0.014
—
—
—
—
—
—
0.005
—
0.001
0.010
—
0.001
0.21
—
—
—
0.001
MDL
—
—
—
0.005
0.007
0.012
0.006
0.003
0.003
0.003
0.059
0.059
0.059
0.004
0.001
0.024
0.001
0.004
0.001
0.002
0.001
0.001
0.001
0.002
0.007
0.001
0.002
0.001
0.003
0.001
0.001
0.001
0.003
0.021
0.001
0.003
0.007
0.001
0.001
0.001
3B-9
-------
TABLE 3B-2f. ST. LOUIS, MO SUMMARY DATA (October 2001 to September 2002).
All concentrations are given in ^g/m3; n = 324; AIRS Site Code: 295100085
Parameter
PM2 5 (FRM Mass)
PMi5_ (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
15.6
16.4
4.3
4.3
1.9
0.19
0.069
2.3
—
—
0.72
—
3.5
0.044
0.002
0.024
0.004
—
0.13
0.027
0.002
0.018
0.24
0.014
0.011
0.014
—
0.002
0.004
0.099
—
0.001
0.17
0.045
0.001
1.4
0.008
0.008
0.001
0.033
Max
52.1
53.2
22.5
20.6
7.2
1.6
5.9
14.0
—
—
2.7
—
11.5
2.5
0.011
0.38
0.16
0.010
1.1
0.57
0.034
0.72
1.3
0.09
0.79
0.13
0.007
0.04
0.095
5.6
0.002
0.006
4.0
0.43
0.12
7.5
0.028
0.14
0.011
0.48
Min
3.8
2.8
0.61
0.64
0.14
0.014
0.007
0.20
—
—
—
—
—
—
0.001
—
—
—
0.010
—
—
—
0.012
—
—
—
—
—
—
0.016
—
—
0.011
—
0.001
0.20
—
—
—
0.002
MDL
—
—
—
0.011
0.015
0.028
0.013
0.008
—
—
0.13
0.13
0.13
0.004
0.001
0.024
0.001
0.004
0.001
0.002
0.001
0.001
0.001
0.002
0.007
0.001
0.002
0.001
0.003
0.001
0.001
0.001
0.003
0.021
0.001
0.003
0.007
0.001
0.001
0.001
3B-10
-------
TABLE 3B-2g. HOUSTON,
All concentrations are
TX SUMMARY DATA (October 2001 to
given in ug/m3; n = 229; AIRS Site Code:
September 2002).
482011039
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
12.4
11.0
3.4
3.5
1.2
0.22
0.051
0.68
0.37
0.32
0.30
—
2.2
0.092
0.001
—
0.003
—
0.055
0.050
0.001
0.003
0.073
0.002
0.020
0.002
0.002
0.002
0.002
0.072
—
0.001
0.23
0.17
0.001
1.1
—
0.007
0.003
0.006
Max
21.9
48.0
24.5
26.0
8.8
1.1
0.38
5.8
3.5
4.2
1.1
—
12.1
1.6
0.006
0.059
0.022
0.008
0.44
0.70
0.009
0.024
0.83
0.007
0.38
0.013
0.022
0.047
0.021
0.64
0.002
0.001
2.8
1.1
0.009
8.2
0.019
0.091
0.011
0.030
Min
5.9
0.0
0.004
0.030
—
—
—
0.040
0.010
0.020
—
—
—
—
0.001
—
—
—
0.001
—
—
—
—
—
—
—
—
—
—
0.001
—
0.001
—
—
0.001
—
—
—
—
—
MDL
—
—
—
0.005
0.007
0.012
0.006
0.003
0.003
0.003
0.059
0.059
0.059
0.004
0.001
0.024
0.001
0.004
0.001
0.002
0.001
0.001
0.001
0.002
0.007
0.001
0.002
0.001
0.003
0.001
0.001
0.001
0.003
0.021
0.001
0.003
0.007
0.001
0.001
0.001
3B-11
-------
TABLE 3B-2h. MINNEAPOLIS, MN SUMMARY DATA (October 2001 to
September 2002). All concentrations are given in ug/m3; n = 163;
AIRS Site Code: 270530963
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
10.3
12.4
2.3
2.4
1.2
0.27
0.059
2.3
—
—
0.39
—
2.6
0.028
0.002
—
0.002
—
0.071
0.008
0.002
0.003
0.065
0.005
0.018
0.002
—
0.002
—
0.068
—
—
0.12
0.093
—
0.78
—
0.006
0.002
0.008
Max Min
33.4 1.8
37.7 2.2
8.6 0.12
8.9 0.24
6.0 —
3.7 —
1.7 —
16.4 0.098
— —
— —
1.8 —
— —
9.9 —
0.67 —
0.012 —
0.15 —
0.008 —
0.028 —
0.46 —
0.17 —
0.016 —
0.045 0.001
0.37 0.016
0.066 —
0.52 —
0.009 —
0.011 —
0.014 0.001
0.11 —
1.8 0.005
0.002 —
0.004 —
1.3 —
0.69 —
0.026 —
2.9 0.040
0.034 —
0.043 —
0.021 —
0.039 0.001
MDL
—
—
—
0.012
0.017
0.030
0.014
0.008
—
—
0.15
0.15
0.15
0.011
0.002
0.059
0.002
0.011
0.003
0.006
0.002
0.001
0.002
0.005
0.018
0.002
0.005
0.001
0.006
0.003
0.002
0.002
0.008
0.051
0.003
0.007
0.018
0.002
0.002
0.001
3B-12
-------
TABLE 3B-21. BOULDER, CO SUMMARY DATA (October 2001 to
All concentrations are given in ug/m3; n = 161; AIRS Site Code:
September 2002).
080010006
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
9.5
11.3
1.5
1.6
0.79
0.24
0.032
1.5
—
—
1.0
—
3.2
0.092
—
—
0.002
—
0.12
0.019
0.002
0.004
0.13
0.005
0.020
0.003
—
0.001
0.006
0.063
—
—
0.28
0.075
—
0.49
—
0.010
—
0.023
Max
26.9
28.2
5.3
5.9
4.3
1.7
0.29
7.9
—
—
4.4
—
13.8
0.41
0.005
0.12
0.011
0.016
0.55
0.66
0.009
0.013
0.47
0.036
0.15
0.015
0.010
0.010
0.096
0.33
0.003
0.004
0.97
0.28
0.005
1.8
0.033
0.031
0.004
0.53
Min
2.1
3.7
0.31
0.29
0.009
0.015
0.007
0.14
—
—
0.15
—
0.28
—
—
—
—
—
0.008
—
—
0.001
0.016
—
—
—
—
0.001
—
—
—
—
0.027
—
—
0.10
—
—
—
0.001
MDL
—
—
—
0.012
0.017
0.030
0.014
0.008
—
—
0.15
0.15
0.15
0.011
0.002
0.059
0.002
0.011
0.003
0.006
0.002
0.001
0.002
0.005
0.018
0.002
0.005
0.001
0.006
0.003
0.002
0.002
0.008
0.051
0.003
0.007
0.018
0.002
0.002
0.001
3B-13
-------
TABLE 3B-2J. PHOENIX, AZ SUMMARY DATA (October 2001 to
All concentrations are given in ug/m3; n = 275; AIRS Site Code:
September 2002).
040139997
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
9.9
12.0
1.2
1.3
0.5
0.28
0.062
1.1
—
—
0.75
—
4.0
0.12
0.002
—
0.003
—
0.15
0.047
0.002
0.006
0.17
—
0.025
0.004
—
0.003
0.006
0.11
—
—
0.36
0.10
0.003
0.39
—
0.013
—
0.009
Max
32.1
158.0
2.9
3.8
2.6
1.9
0.64
6.9
—
—
5.0
—
21.0
0.87
0.009
0.13
0.009
0.024
0.82
0.40
0.021
0.038
0.95
0.020
0.38
0.027
0.011
0.13
0.13
0.69
0.003
0.005
2.2
0.52
0.10
0.98
0.038
0.058
0.005
0.080
Min MDL
3.3 —
0.8 —
0.001 —
— 0.012
— 0.017
— 0.030
— 0.014
— 0.008
— —
— —
— 0.15
— 0.15
-0.66 0.15
— 0.011
— 0.002
— 0.059
— 0.002
— 0.011
0.005 0.003
— 0.006
— 0.002
0.001 0.001
0.003 0.002
— 0.005
— 0.018
— 0.002
— 0.005
— 0.001
— 0.006
— 0.003
— 0.002
— 0.002
— 0.008
— 0.051
— 0.003
— 0.007
— 0.018
— 0.002
— 0.002
0.001 0.001
3B-14
-------
TABLE 3B-2k. SEATTLE, WA SUMMARY DATA (October 2001 to
All concentrations are given in ug/m3; n = 314; AIRS Site Code:
September 2002).
530330080
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
8.2
8.0
1.4
1.3
0.47
0.16
0.045
0.67
0.25
0.41
0.60
—
2.6
0.015
0.001
—
0.002
—
0.029
0.055
0.002
0.003
0.053
0.004
0.014
0.003
0.002
0.002
—
0.055
—
0.001
0.049
0.15
0.001
0.45
0.007
0.003
0.003
0.009
Max
29.1
28.5
4.9
4.6
2.0
0.76
2.3
2.9
1.6
2.5
2.7
—
11.0
0.24
0.007
0.15
0.014
0.007
0.20
0.84
0.016
0.045
0.29
0.078
0.19
0.024
0.011
0.020
0.028
2.3
0.002
0.008
0.52
0.84
0.044
1.6
0.030
0.026
0.029
0.048
Min
1.9
1.5
0.057
0.14
—
0.012
—
0.12
0.010
0.020
—
—
—
—
0.001
—
—
—
0.001
—
—
—
0.001
—
—
—
—
—
—
0.007
—
0.001
—
—
0.001
0.019
—
—
—
—
MDL
—
—
—
0.005
0.007
0.012
0.006
0.003
0.003
0.003
0.059
0.059
0.059
0.004
0.001
0.024
0.001
0.004
0.001
0.002
0.001
0.001
0.001
0.002
0.007
0.001
0.002
0.001
0.003
0.001
0.001
0.001
0.003
0.021
0.001
0.003
0.007
0.001
0.001
0.001
3B-15
-------
TABLE 3B-21. SACRAMENTO, CA SUMMARY DATA (October 2001 to September
2002). All concentrations are given in ng/m3; n = 265; AIRS Site Code: 060670006
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Silver
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
9.4
15.0
1.2
1.3
0.74
0.38
0.077
2.3
—
—
0.66
—
5.2
0.040
0.002
—
0.002
—
0.043
0.052
0.002
0.006
0.079
—
0.021
0.002
—
0.010
0.006
0.10
—
—
0.12
—
0.19
—
0.41
—
—
—
0.005
Max
78.0
120.0
5.5
4.1
7.4
3.5
3.6
23.4
—
—
8.4
—
54.1
0.52
0.006
0.26
0.013
0.018
0.56
1.8
0.043
0.16
0.64
0.044
0.88
0.013
0.015
0.61
0.11
3.3
0.003
0.004
1.1
0.015
1.7
0.077
1.8
0.035
0.052
0.006
0.11
Min
3.0
3.0
0.14
0.06
0.009
—
—
—
—
—
—
—
0.18
—
—
—
—
—
0.005
—
—
0.001
0.005
—
—
—
—
0.001
—
—
—
—
0.011
—
—
—
0.045
0.009
—
—
—
MDL
—
—
—
0.012
0.017
0.030
0.014
0.008
—
—
0.15
0.15
0.15
0.011
0.002
0.059
0.002
0.011
0.003
0.006
0.002
0.001
0.002
0.005
0.018
0.002
0.005
0.001
0.006
0.003
0.002
0.002
0.008
0.010
0.010
0.051
0.003
0.007
0.018
0.002
0.002
3B-16
-------
TABLE 3B-2m. RIVERSIDE-RUBIDOUX, CA SUMMARY DATA (October 2001 to
September 2002). All concentrations are given in ug/m3; n = 161;
AIRS Site Code: 060658001
Parameter
PM2 5 (FRM Mass)
PM2 5 (Reconstructed Mass)
Sulfate (Calculated)
Sulfate (by I.C.)
Ammonium (by I.C.)
Sodium Ion (by I.C.)
Potassium (by I.C.)
Nitrate
Volatile Nitrate
Nonvolatile Nitrate
Elemental Carbon
Carbonate Carbon
Organic Carbon
Aluminum
Arsenic
Barium
Bromine
Cadmium
Calcium
Chlorine
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorous
Potassium
Rubidium
Selenium
Silicon
Silver
Sodium
Strontium
Sulfur
Tin
Titanium
Vanadium
Zinc
Mean
28.6
30.5
3.6
3.7
4.8
0.49
0.075
12.3
—
—
1.2
—
6.0
0.057
0.002
—
0.006
—
0.17
0.072
0.003
0.006
0.17
0.006
0.032
0.004
—
0.002
0.007
0.11
—
—
0.20
—
0.20
—
1.2
0.014
0.012
0.006
0.023
Max
78.2
79.2
9.6
10.0
16.8
1.8
1.3
40.3
—
—
4.3
—
13.0
0.32
0.006
0.17
0.016
0.017
1.0
0.71
0.024
0.037
0.58
0.026
0.31
0.015
0.024
0.014
0.067
1.31
0.003
—
0.88
0.015
0.94
0.026
3.2
0.081
0.041
0.017
0.21
Min
2.6
3.1
0.23
0.28
—
—
—
0.18
—
—
0.12
—
0.38
—
—
—
—
—
0.012
—
—
0.001
0.019
—
—
—
—
0.006
—
0.017
—
—
0.023
—
—
—
0.078
—
—
—
0.001
MDL
—
—
—
0.012
0.017
0.030
0.014
0.008
—
—
0.15
0.15
0.15
0.011
0.002
0.059
0.002
0.011
0.003
0.006
0.002
0.001
0.002
0.005
0.018
0.002
0.005
0.001
0.006
0.003
0.002
0.006
0.008
0.010
0.051
0.003
0.007
0.018
0.002
0.002
0.001
3B-17
-------
TABLE 3B-3. BLANK CORRECTIONS FOR ELEMENTAL CARBON (EC),
ORGANIC CARBON (OC), AND TOTAL CARBON (TC)
IN THE SPECIATION NETWORK
Sampler
Type
URG MASS
R&P2300
Anderson
RAAS
R & P 2025
MetOne SASS
Elemental
EC Mass
(figC/filter)
0.63
3.21
0.97
1.67
1.03
Carbon
EC Cone
OigC/m3)
0.03
0.22
0.09
0.07
0.11
Organic
OC Mass
(figC/filter)
7.08
12.93
12.54
18.42
13.75
Carbon
OC Cone
OigC/m3)
0.29
0.9
1.19
0.77
1.42
Total
TC Mass
(figC/filter)
7.71
16.13
13.51
19.91
14.78
Carbon
TC Cone
OigC/m3)
0.32
1.12
1.29
0.83
1.53
3B-18
-------
APPENDIX 3C
Organic Composition of Particulate Matter
Although organic compounds typically constitute approximately 10 to 70% of the total dry
fine particle mass in the atmosphere, organic PM concentrations, composition, and formation
mechanisms are poorly understood. This is because organic particulate matter is an aggregate of
hundreds of individual compounds spanning a wide range of chemical and thermodynamic
properties (Saxena and Hildemann, 1996). The presence of multiphase or "semivolatile"
compounds complicates collection of organic particulate matter. Furthermore, no single
analytical technique currently is capable of analyzing the entire range of compounds present.
Rigorous analytical methods frequently identify only 10 to 20% of the organic mass on the
molecular level (Rogge et al., 1993). The data shown in Appendix 3C are meant to complement
the data given for the inorganic components of particles in Appendix 6 A of the 1996 PM AQCD
(U.S. Environmental Protection Agency, 1996). Table 3C-1 lists a number of recent urban
and some rural measurements of particulate organic and elemental carbon in jig of carbon/m3
(jig C/m3). Emphasis is placed on measurements published after 1995. The analysis method and
artifact correction procedure, if any, are indicated. Table 3C-2 presents information on recent
(post-1990) studies concerning concentrations (in ng C/m3) of particulate organic compounds
found at selected U.S. sites.
3C-1
-------
TABLE 3C-1. PARTICULATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in jig C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
Reference
URBANPM25
Offenberg and Baker
(2000)
Allen etal. (1999)
Pedersen et al. (1999)
OJ
O
^ IMPROVE (2000)
Lewtasetal. (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
OC Mean
(Max)
(0.8-8.4)3
5.8
4.0
2.8
3.3
2.7
3.4
1.8
8
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
12 h
3h
10 min
24 h
24 h
23 h
6h
24 h
2h
2-6 h
Notes
PM12; Imp; TOT
PMM; Imp; TOT
PM25; DQQ; TORb
Aeth
PM20; Q; TOT
PM25; QQ; TOR
PM25; DQA; EGAC
PMLO; Q; Th
PM2l; Q; TOR
PM25; Q+TQ; TOTd
-------
TABLE 3C-1 (cont'd). PARTICULATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in jig C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
Reference
Location
Dates
OC Mean
(Max)
EC Mean
(Max)
TC Mean
(Max)
Avg.
Time
Notes
RURAL PM2 5
Klinedinst and Currie
(1999)
Andrews et al. (2000)
Malm and Gebhart
(1996)
IMPROVE (2000)
O
i
OJ
Heggetal. (1997)
Welby, CO
Brighton, CO
Look Rock, Smoky Mountains,
TN
Tahoma Wood, WA
Three Sisters Wilderness, OR
Rocky Mountains, CO
Brigantine, NJ
Acadia, MA
Jefferson: James River Face
Wilderness, VA
Glacier, MT
150 km East of Mid-Atlantic
Coast
(0.02-4 km altitude)
Dec 1996-Jan
1997
July-Aug 1995
June-Aug 1990
1994-1998
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
3.3(8.1)
1.9(3.6)
0.4
0.1
0.2
0.7 (2.2)
0.2
0.2
0.5
0.2
0.7
0.4
6 h PM25; Q; TOR
12 h (day) PM21;QQ;TORe
PM21;Q+TQ;TORd
PML8; Imp; TMO
12 h PM25; QQ; TORf
24 h
PM25; QQ; TOR
July 1996
2.9 (5.4)
PM10; QQ; EGA6
Cui et al. (1 997) Meadview, AZ
Chow et al. ( 1 996) Point Reyes, CA
Altamont Pass, CA
Pacheco Pass, CA
Crows Landing, CA
Academy, CA
Button- Willow, CA
Edison, CA
Caliente, CA
Sequoia, CA
Yosemite, CA
Aug6-15, 1992
July-Aug 1990 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)
3 12 h PM25; VDQA; EGA0
0.4(0.6) 5-7 h PM25;Q + TQ;TOR8
2.6 (3.9)
1.0(1.3)
1.8(2.5)
1.4 (2.4)
1.9 (2.7)
2.9(4.1)
3.3 (4.4)
1.6(3.0)
1.9(3.5)
-------
TABLE 3C-1 (cont'd). PARTICULATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in jig C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
Reference
Location
Dates
OCMean EC Mean TC Mean Avg.
(Max) (Max) (Max) Time
Notes
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 PM2 5; 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 24 h PM10; Q; TOR
4.6
3.5
3.4
2.1
3.1
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.
GSum of adsorbent and filter after correction for inlet losses and denuder efficiency.
dCorrected 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
Rogge et al. (1993)°
Jan-Dec 1982
(annual average)
PM^
n-Alkanes
n-tricosane
n-tetracosane
n-pentacosane
n-hexacosane
n-heptacosane
n-octacosane
n-nonacosane
n-triacontane
— . n-hentriacontane
(]/i 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
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
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
Schauer and Cass (2000)
Dec 26-28, 1995
(pollution episode)
™15
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
Bakersfield,
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
Khwaja (1995) Allen et al. (1997)
Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
no precut no precut PMj , (urban)
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
-------
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)°
Jan-Dec 1982
(annual average)
PM2.,
n-Alkanoic Acids
(cont'd)
n-docosanoic acid
n-tricosanoic acid
n-tetracosanoic acid
n-pentacosanoic acid
n-hexacosanoic acid
n-heptacosanoic acid
n-octacosanoic acid
n-nonacosanoic acid
n-triacontanoic acid
Toted n-alcanoic acids
n-Alkenoic Acids
n-9-hexadecenoic acid
n-9-octadecenoic acid
n-9, 1 2-octadecane-dienoic
Los Angeles,
CA
8.7
2.0
11.8
1.3
5.6
0.49
2.7
0.33
1.0
292.6
24.8
Pasadena,
CA
9.9
2.5
16.5
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
PM25 no precut no precut PMj , (urban)
Fresno,
CA
160
32.1
205
15.4
174
2.56
21.3
1.46
4.32
979.3
18.8
27.1
13.6
Bakersfield, Niwot Ridge, Schenectady, Kenmore Square,
CA CO NY Boston, MA Los Angeles Basin, CA
43.1
9.71
78.0
6.59
81.3
2.38
9.65
2.11
5.79
352.7
3.96
3.96
1.83
acid
Total n-alkenoic acids
24.8
26.0
9.75
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(14.4)
5.7 9.5 19.4 3.01 29.01 (62.8)
23.58(71.2)
6.01 (16.4)
6.50(25.8)
9.62 (30.7)
12.47(113.6)
17.45 (49.3)
24.09 (88.9)
1.84(11.7)
5.7 9.5 19.4 3.01 133.8
-------
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)" Schauer and Cass (2000)
Jan-Dec 1982 Dec 26-28, 1995
(annual average) (pollution episode)
PM^ PM2.5
Los Angeles, Pasadena, Fresno, Bakersfield,
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 15.0 24.0 8.80
(methylbutanedioic)
glutaricacid 32.3 28.3 21.3 10.5
(pentanedioic)
methylglutaric acid 19.3 16.6
(methylpentanedioic)
hydroxybutanedioic acid 14.3 16.0
adipicacid 14.1 14.1 3.39 3.07
(hexanedioic)
pimelic acid (heptanedioic) 2.22 1.03
suberic acid 3.4 4.1 4.41 13.4
(octanedioic)
axelaic acid (nonanedioic) 29.0 22.8 19.9 8.22
Total aliphatic 230.3 213.8 77.4 45.0
dicarboxylic acids
Khwaja (1995) Allen et al. (1997)
Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
no precut no precut PMj , (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
p
oo
Ketocarboxylic Acids
pyruvic acid (C3)
glyoxylic acid (C2)
Total ketocarboxylic acids
Diterpenoid/Resin Acids
dehydroabietic acid
abietic acid
13-isopropyl-5a-podocarpa-
6,8,1 l,13-tetraen-16-oic
acid
8,15-pimaradien-18-oic acid
pimaric acid
isopimaric acid
7-oxodehydroabietic acid
abieta-6,8,11,13, 15-pentaen-
18-oic acid
abieta-8,ll,13,15-tetraen-
18-oic acid
sandaracopimaric acid
Total diterpenold acids
Aromatic Polycarboxylic
Acids
1,2-benzene-dicarboxylic
acid (phthalic acid)
1,3-benzene-dicarboxylic
acid
Rogge et al. (1993)°
Jan-Dec 1982
(annual average)
PM^
Los Angeles, Pasadena,
CA CA
23.6 22.6
0.63 1.2
0.44 0.57
2.3 4.8
1.3 2.3
3.4 4.1
1.6 2.2
33.3 37.6
60.0 55.7
3.4 2.9
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
PM25 no precut no precut PMj , (urban)
Fresno, Bakersfield, Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CA CA CO NY Boston, MA CA
98.5
30.4
0.48
9.97
127
6.68
11.8
2.62
8.91
296.4
9.16
3.41
59(103)
44 (68)
103
8.01
0.784
0.03
0.735
7.95
1.43
2.43
0.251
0.525
22.15
6.78
1.98
-------
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
Aromatic Polycarboxylic
Acids (cont'd)
1,4-benzene-dicarboxylic
acid
benzene tricarboxylic acids
4-methyl-l,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
Polycyclic Aromatic
Hydrocarbons
retene
fluoranthene
acephenanthrylene
pyrene
Cj-202 MW PAH
C2-202 MW PAH
benz [a] anthracene
cyclopenta[crf|pyrene
benzo[gfe]-fluoranthene
Cj-226 MW PAH
chrysene/triphenyline
Rogge et al. (1993)'
Jan-Dec 1982
(annual average)
Los Angeles, Pasadena,
CA CA
2.8 1.5
27.8 28.8
0.52 0.84
20.6 17.2
0.74 0.80
115.9 107.7
0.07 0.06
0.15 0.13
0.26 0.17
0.29 0.25
0.23 0.41
0.39 0.30
0.61 0.43
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban)
PM25 no precut no precut PMj ,
Fresno,
CA
5.16
14.4
32.1
6.02
2.52
0.834
3.28
11.7
13.8
1.90
6.05
10.1
7.70
Bakersfield, Niwot Ridge, Schenectady, Kenmore Square,
CA CO NY Boston, MA
4.48
8.77
22.0
0.563
0.553
0.302
0.564
3.80
2.49
0.496
1.25
1.48
1.50
Fraser et al. (1998)
Sept 8-9, 1993
(urban)
Los Angeles Basin,
CA
0.07 (0.26)
0.02 (0.05)
0.07 (0.26)
0.07(0.36)
0.03 (0.32)
0.15(1.09)
0.14(1.02)
0.20 (0.97)
0.14(0.97)
0.34(1.62)
-------
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
o
I
O
Rogge et al. (1993)° Schauer and Cass (2000)
Jan-Dec 1982 Dec 26-28, 1995 Veltkamp et al. (1996)
(annual average) (pollution episode) July 24-Aug 4, 1989
PM2 j PM25 no precut
Los Angeles,
CA
Polycyclic Aromatic
Hydrocarbons (cont'd)
Cj-228 MW PAH
C2-228 MW PAH
benz[e]acephen-anthrylene
benzo[£]fluoranthene 1.15
benzo[fe]fluoranthene 1.23
benzo[/]fluoranthene
benzo[e]pyrene 0.97
benzo[a]pyrene 0.42
perylene
methyl-substituted 252 MW
PAH
indeno[7,2,_?-crf]-pyrene 0.37
indeno[7,2,_?-crf]-fluoranthene 1.05
benzo[gfa]perylene 4.47
anthanthrene
coronene
Total polycyclic aromatic 11.66
hydrocarbons
Pasadena, Fresno,
CA CA
17.6
1.20 8.69
0.85 10.7
3.62
0.93 7.20
0.44 8.23
1.50
0.42 6.84
1.09 1.36
4.43 9.75
0.180
11.10 139.57
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
Khwaja (1995) Allen et al. (1997)
October 1991 Summer 1994 Fraser et al. (1998)
(semiurban) (urban) Sept 8-9, 1993
no precut PMj , (urban)
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
Oxygenated PAHs/
Polycyclic Aromatic
Ketones/Quinones
1,4-naphthoquinone
1 -acenaphthenone
9-fluorenone
1,8-naphthalic anhydride
phenanthrenequinone
phenalen-9-one
anthracene-9,10-dione
methylanthracene-9,10-dione
1 lH-benzo[a]fluoren-l 1-one
0.26
2.07
1.77
0.43
1.03
0.29(1.04)
0.41 (1.65)
0.53 (2.23)
0.36(1.14)
0.09 (0.24)
-------
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)°
Jan-Dec 1982
(annual average)
Los Angeles, Pasadena,
CA CA
Oxygenated PAHs/
Polycyclic Aromatic
Ketones/Quinones (cont'd)
7H-benzo [c]fluoren-7-one
1 1 H-benzo [fejfluoren- 1 1 -one
1 H-phenalen- 1 -one
benzanthrone
5, 12-naphthacene-quinone
W 7H-benz[rfe]-anthracen-7-one 0.81 0.84
o
,' . benz[fife]anthracene-7-dione
benz[a]anthracene-7,12-dione 0.21 0.25
cyclopenta[rfe/]phen-anthrone
benzo[crf|pyren-6-one 0.80 1.24
6H-benzo [crf|pyrene-6-one
benzo[a]pyrene-6,12-dione
Total polycyclic aromatic 1.82 2.33
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban)
PM25 no precut no precut PMj ,
Fresno, Bakersfield, Niwot Ridge, Schenectady, Kenmore Square,
CA CA CO NY Boston, MA
0.37
0.85
7.96 0.588
1.18
0.32
7.80 1.48
1.34
0.096
15.76 2.07 9.72
Fraser et al. (1998)
Sept 8-9, 1993
(urban)
Los Angeles Basin,
CA
0.20(1.00)
0.09(0.31)
0.05(0.14)
0.54 (2.47)
2.56
ketones/quinones
Steroids
cholesterol
nd
1.9
Substituted Phenols
p-benzenediol
m-benzenediol
hydro xybenzaldehydes
Total substituted phenols
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
O
Rogge et al. (1993)° 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 j PM25 no precut no precut PMj , (urban)
Guaiacol and Substituted
Guaiacols
guaiacol
4-methylguaiacol
trans-isoeugenol
vanillin
acetovanillone
guaiacyl acetone
coniferyl aldehyde
Toted 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
Los Angeles, Pasadena, Fresno,
CA CA 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
Bakersfield, 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
Rogge et al. (1993)°
Jan-Dec 1982
(annual average)
PM2.,
Los Angeles, Pasadena,
CA CA
Other Compounds
divanillyl
divanillyl methane
vanillylmethylguaiacol
Total other
N-Containing Compounds
3-methoxypyridine 0.86 1.4
isoquinoline 1.1 1.1
1-methoxypyridine 0.27 0.24
l,2-dimethoxy-4-nitro-benzene 1.8 3.9
dihydroxynitrobenzene
Total N-containing 4.03 6.64
compounds
Total Quantified Organic 789 764
Compound Mass
Total Organic Compound
Mass
Percent of Organic Mass 8-15% (a) 8-15% (a)
Quantified
Percent of Organic Mass 45-60% (a) 45-60% (a)
Extractable and Elutable
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997) Fraser et al.
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM25 no precut no precut PMj , (urban)
Fresno, Bakersfield, Niwot Ridge, Schenectady, Kenmore Square, Los Angeles
CA CA CO NY Boston, MA Basin, CA
19.4 3.18
2.39 nd
3.24 0.568
25.0 3.75
1.62(10.52)
1.62
11410 2075 267 487 10 8
55700 18700
20% 11% <3%
30% 21%
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.
-------
REFERENCES
Allen, J. O.; Dookeran, N. M.; Taghizadeh, K.; Lafleur, A. L.; Smith, K. A.; Sarofim, A. F. (1997) Measurement of
oxygenated polycyclic aromatic hydrocarbons associated with a size-segregated urban aerosol. Environ. Sci.
Technol. 31:2064-2070.
Allen, G. A.; Lawrence, J.; Koutrakis, P. (1999) Field validation of a semi-continuous method for aerosol black
carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
southwestern PA. Atmos. Environ. 33: 817-823.
Andrews, E.; Saxena, P.; Musarra, S.; Hildemann, L .M.; Koutrakis, P.; McMurry, P. H.; Olmez, L; White, W. H.
(2000) Concentration and composition of atmospheric aerosols from the 1995 SEAVS experiment and a
review of the closure between chemical and gravimetric measurements. J. Air Waste Manage. Assoc.
50: 648-664.
Chow, J. C.; Watson, J. G.; Lu, Z.; Lowenthal, D. H.; Frazier, C. A.; Solomon, P. A.; Thuillier, R. H.; Magliano, K.
(1996) Descriptive analysis of PM25 and PM10 at regionally representative locations during
SJVAQS/AUSPEX. In: Parrish, D.; Trainer, M.; Rao, S. T.; Solomon, P. A., eds. A&WMA international
specialty conference on regional photochemical measurements and modeling, part 2; November 1993;
San Diego, CA. Atmos. Environ. 30: 2079-2112.
Christoforou, C. S.; Salmon, L. G.; Hannigan, M. P.; Solomon, P. A.; Cass, G. R. (2000) Trends in fine particle
concentration and chemical composition in southern California. J. Air Waste Manage. Assoc. 50: 43-53.
Cui, W.; Machir, J.; Lewis , L.; Eatough, D. J.; Eatough, N. L. (1997) Fine paniculate organic material at Meadview
during the project MOHAVE summer intensive study. J. Air Waste Manage. Assoc. 47: 357-369.
Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T.; Rasmussen, R. A. (1998) Air quality model evaluation data for
organics. 5. C6-C22 nonpolarand semipolar aromatic compounds. Environ. Sci. Technol. 32: 1760-1770.
Gertler, A. W.; Lowenthal, D. A.; Coulombe, W. G. (1995) PM10 source apportionment study in Bullhead City,
Arizona. J. Air Waste Manage. Assoc. 45: 75-82.
Hegg, D. A.; Livingston, J.; Hobbs, P. V.; Novakov, T.; Russell, P. (1997) Chemical apportionment of aerosol
column optical depth off the mid-Atlantic coast of the United States. J. Geophys. Res. 102: 25,293-25,303.
IMPROVE: interagency monitoring of protected visual environments [database]. (2000) [Data on paniculate organic
and elemental carbon concentrations after 1995]. Fort Collins, CO: National Park Service (NFS); Cooperative
Institute for Research in the Atmosphere (CIRA). Available at: http://vista.cira.colostate.edu/improve/
[2001, January 26].
Khwaja, H. (1995) Atmospheric concentrations of carboxylic acids and related compounds at a semiurban site.
Atmos. Environ. 29: 127-139.
Klinedinst, D. B.; Currie, L. A. (1999) Direct quantification of PM2 5 fossil and biomass carbon within the northern
front range air quality study's domain. Environ. Sci. Technol. 33: 4146-4154.
Lewtas, J.; Pang, Y.; Booth, D.; Reimer, S.; Eatough, D. J.; Gundel, L. A. (2001) Comparison of sampling methods
for semi-volatile organic carbon associated with PM25. Aerosol. Sci. Technol. 34: 9-22.
Lioy, P. J.; Daisey, J. M. (1987) Toxic air pollution: a comprehensive study of non-criteria air pollutants. Chelsea,
MI: Lewis Publishers.
Malm, W. C.; Day, D. E. (2000) Optical properties of aerosols at Grand Canyon National Park. Atmos. Environ.
1 /I • T5T5 11 O1
34: 3373-3391.
Malm, W. C.; Gebhart, K. A. (1996) Source apportionment of organic and light-absorbing carbon using receptor
modeling techniques. Atmos. Environ. 30: 843-855.
Offenberg, J. H.; Baker, J. E. (2000) Aerosol size distributions of elemental and organic carbon in urban and
over-water atmospheres. Atmos. Environ. 34: 1509-1517.
Omar, A. H.; Biegalski, S.; Larson, S. M.; Landsberger, S. (1999) Paniculate contributions to light extinction and
local forcing at a rural Illinois site. Atmos. Environ. 33: 2637-2646.
Pedersen, D. U.; Durant, J. L.; Penman, B. W.; Crespi, C. L.; Hemond, H. F.; Lafleur, A. L.; Cass, G. R. (1999)
Seasonal and spatial variations in human cell mutagenicity of respirable airborne particles in the northeastern
United States. Environ. Sci. Technol. 33: 4407-4415.
Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993) Quantification of urban
organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos. Environ.
Part A 27: 1309-1330.
Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem. 24: 57-109.
3C-14
-------
Schauer, J. I; Cass, G. R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants
using organic compounds as tracers. Environ. Sci. Technol. 34: 1821-1832.
Turpin, B. I; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
Veltkamp, P. R.; Hansen, K. J.; Barkley, R. M; Sievers, R. E. (1996) Principal component analysis of summertime
organic aerosols at Niwot Ridge, Colorado. J. Geophys. Res. [Atmos.] 101: 19,495-19,504.
3C-15
-------
APPENDIX 3D
Composition of Particulate Matter Source Emissions
This appendix includes discussions of the elemental composition of emissions from various
source categories discussed in Table 3-8. Discussions in this appendix incorporate material
dealing with the inorganic components of source emissions from Chapter 5 of the 1996 PM
AQCD (U.S. Environmental Protection Agency, 1996), provides updates to that material, and
adds material describing the composition of organic components in source emissions. Primary
emphasis is placed in the discussions on the composition of PM25 sources.
Soil and Fugitive Dust
The compositions of soils and average crustal material are shown in Table 3D-1 (adapted
from Warneck, 1988). Two entries are shown as representations of average crustal material.
Differences from the mean soil composition shown can result from local geology and climate.
Major elements in both soil and crustal profiles are Si, Al, and Fe, which are found in the form
of various minerals. In addition, organic matter constitutes a few percent, on average, of soils.
In general, the soil profile is similar to the crustal profiles, except for the depletion of soluble
elements such as Ca, Mg, Na, and K. It should be noted that the composition of soils from
specific locations can vary considerably from these global averages, especially for elements like
Ca, Mg, Na, and K.
Fugitive dust emissions arise from paved and unpaved roads, building construction and
demolition, parking lots, mining operations, storage piles, feed lots, grain handling, and
agricultural tilling in addition to wind erosion. Figure 3D-1 shows examples of size distributions
in dust from paved and unpaved roads, agricultural soil, sand and gravel, and alkaline lake bed
sediments, which were measured in a laboratory resuspension chamber as part of a study in
California (Chow et al., 1994). This figure shows substantial variation in particle size among
some of these fugitive dust sources. The PMX 0 abundance (6.9%) in the total suspended PM
(TSP) from alkaline lake bed dust is twice its abundance in paved and unpaved road dust.
Approximately 10% of the TSP is in the PM25 fraction and -50% of TSP is in the PM10 fraction.
3D-1
-------
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
13700
6,300
6,300
13,600
4,600
850
200
100
8
Abundances (ppm)
Crustal Rock
(2)
277,200
81,300
50,000
36,300
20,900
28,300
25,900
4,400
950
100
135
25
(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).
100
Paved
Road Dust
Un paved
Road Dust
Agricultural
Soil
Soil/Gravel
EZ3<1.0um I I <2.5 urn
Source
I < 10 urn E3 TSP
Alkaline
Lake Bed
Figure 3D-1. Size distribution of particles generated in a laboratory resuspension
chamber.
Source: Chow etal. (1994).
3D-2
-------
The sand/gravel dust sample shows that 65% of the mass is in particles larger than the PM10
fraction. The PM25 fraction of TSP is -30 to 40% higher in alkaline lake beds and sand/gravel
than in the other soil types. The tests were performed after sieving and with a short (< 1 min)
waiting period prior to sampling. It is expected that the fraction of PMLO and PM2 5 would
increase with distance from a fugitive dust emitter, as the larger particles deposit to the surface
faster than do the smaller particles.
The size distribution of samples of paved road dust obtained from a source characterization
study in California is shown in Figure 3D-2. As might be expected, most of the emissions are in
the coarse size mode. The chemical composition of paved road dust obtained in Denver, CO
during the winter of 1987-1988 is shown in Figure 3D-3. The chemical composition of paved
road dust consists of a complex mixture of PM from a wide variety of sources. Hopke et al.
(1980) found that the inorganic composition of urban roadway dust in samples from Urbana, IL
could be described in terms of contributions from natural soil, automobile exhaust, rust, tire
wear, and salt. Automobile contributions arose from: exhaust emissions enriched in Pb; Fe as
rust; tire wear particles enriched in Zn; brake linings enriched in Cr, Ba, and Mn; and cement
particles derived from roadways by abrasion. In addition to organic compounds from
combustion and secondary sources, road dust also contains biological material such as pollen
and 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
sample collected at several sites in the Central Valley of California in 1995. In both cases, road
dust samples were resuspended in the laboratory. Samples were drawn through a PM2 0 cyclone
upstream of the collection substrate to remove particles with aerodynamic diameters > 2.0 jim.
It is unclear if these samples are representative of road dust in other locations of the United
States. Table 3D-2 summarizes the organic compounds measured in these road dust samples.
-------
99.2%
100-1
80-
0. 60'
C/3
O
+•>
a
Q_ 40-
20-
52.3%
92.8%
82.7%
(<2.5M)
81.6%
^
r / J
95.8
34.9%
Road and Agriculteral Residential Diesel
Soil Dust Burning Wood Truck
Combustion Exhaust
Crude Oil Construction
Combustion Dust
>10p
2.5p- 10u
Source
1M-2.5M
Figure 3D-2. Size distribution of California source emissions, 1986.
Source: Houcketal. (1989, 1990).
3D-4
-------
Chemical Species
Figure 3D-3. Chemical abundances for PM2 5 emissions from paved road dust 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).
TABLE 3D-2. SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
PRESENT IN FINE PARTICLE ROAD DUST SAMPLE
Source
Pasadena Road Dust
(Roggeetal, 1993a)
San Joaquin Valley
Road Dust (Schauer
and Cass, 2000)
Contribution to Dominant Contributors to
Compound Class Particulate 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
C17, 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
3D-5
-------
Stationary Sources
The elemental composition of primary PM emitted in the fine fraction from a variety of
power plants and industries in the Philadelphia area is shown in Table 3D-3 as a representative
example of emissions from stationary fossil-fuel combustion sources (Olmez et al., 1988).
Entries for the coal fired power plant show that Si and Al, followed by sulfate, are the major
primary constituents produced by coal combustion, whereas fractional abundances of elemental
carbon (EC) were much lower and organic carbon (OC) species were not detected. Sulfate is the
major particulate constituent released by the oil-fired power plants examined in this study; and,
again, EC and OC are not among the major species emitted. Olmez et al. (1988) also compared
their results to a number of similar studies and concluded that their data should have much wider
applicability to receptor model studies in other areas with some of the same source types. The
high temperature of combustion in power plants results in the almost complete oxidation of the
carbon in the fuel to CO2 and very small amounts of CO. Combustion conditions in smaller
boilers and furnaces allow the emission of unburned carbon and sulfur in more reduced forms
such as thiophenes and inorganic sulfides. A number of trace elements are greatly enriched over
crustal abundances in different fuels, such as Se in coal and V, Zn, and Ni in oil. In fact, the
higher V content of the fuel oil than in coal could help account for the higher sulfate seen in the
profiles from the oil-fired power plant compared to the coal-fired power plant, because V at
combustion temperatures found in power plants is known to catalyze the oxidation of reduced
sulfur species. During combustion at lower temperatures, the emission of reduced sulfur species
also occurs. For example, Huffman et al. (2000) identified sulfur species emitted by the
combustion of several residual fuel oils (RFO) in a fire tube package boiler that is meant to
simulate conditions in small institutional and industrial boilers. They found that sulfur was
emitted not only as sulfate (26 to 84%), but as thiophenes (13 to 39%) with smaller amounts of
sulfides and elemental S. They also found that Ni, V, Fe, Cu, Zn, and Pb are present mainly as
sulfates in emissions. Linak et al. (2000) found, when burning RFO, that the fire tube package
boiler produced particles with a bimodal size distribution in which -0.2% of the mass was
associated with particles < 0.1 jim AD, with the rest of the mass lying between 0.5 and 100 jim.
Miller et al. (1998) found that larger particles consisted mainly of cenospheric carbon; whereas
trace metals and sulfates were found concentrated in the smaller particles in a fire tube package
boiler. In contrast, when RFO was burning in a refractory-lined combustor that is meant to
3D-6
-------
TABLE 3D-3. COMPOSITION OF FINE PARTICLES RELEASED BY VARIOUS STATIONARY SOURCES
IN THE PHILADELPHIA AREA
Species
(Units)
C-v (%)
C-e (%)
NH4 (%)
Na (%)
Al (%)
Si (%)
P (%)
0 S (%)
"^ SO4 (%)
Cl (%)
K(%)
Ca (%)
Sc (ppm)
Ti (%)
V (ppm)
Cr (ppm)
Mn (ppm)
Fe (%)
Co (ppm)
Ni (ppm)
Eddystone
Coal-Fired
Power Plant
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
11,500 ±3000
235 ±10
380 ± 40
1.6 ±0.2
790 ± 150
15,000 ± 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
19,000 ± 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
Species
(units)
Cu (ppm)
Zn (%)
As (ppm)
Se (ppm)
Br (ppm)
Rb (ppm)
Sr (ppm)
w Zr (ppm)
oo Mo (ppm)
Ag (ppm)
Cd (ppm)
In (ppm)
Sn (ppm)
Sb (ppm)
Cs (ppm)
Ba (ppm)
La (ppm)
Ce (ppm)
Nd (ppm)
Sm (ppm)
Eddystone
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
6 100 ±300
ND
ND
19 ±2
ND
ND
ND
Fluid Cat.
N Cracker N
2 14 ± 8 9
1 0.0026 ± 0.0007 3
1 ND
1 15 ±1 3
2 5.6 ±1.8 9
1 ND
36 ±6 9
130 ±50 2
ND
ND
ND
ND
2 ND
1 7.7 ±1.5 3
ND
290 ±90 2
1 3300 ±500 3
2700 ± 400 3
1800 ± 250 3
170 ±20 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
Species
(units)
Eu (ppm)
Gd (ppm)
Tb(ppm)
Yb (ppm)
Lu (ppm)
Hf(ppm)
^ Ta (ppm)
^ W (ppm)
Au (ppm)
Pb (%)
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
N Eddystone N
3 ND
ND
3 ND
1 ND
ND
3 0.39 ±0.07 1
ND
1 60 ± 5 2
0.054 ±0.017 2
9 1.8 ±0.6 9
3 1.9 ±0.5 2
6 93.5 ±2.5 6
Power Plants
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
a Omitted 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.
lppm=10-4%
Source: Adapted from Olmez et al. (1988).
-------
simulate combustion conditions in a large utility residual oil fired boiler, Linak et al. (2000)
found that particles were distributed essentially unimodally, with a mean diameter of about
0.1 |im.
Apart from emissions in the combustion of fossil fuels, trace elements are emitted as the
result of various industrial processes such as steel and iron manufacturing and nonferrous metal
production (e.g., for Pb, Cu, Ni, Zn, and Cd). As may be expected, emissions factors for the
various trace elements are highly source-specific (Nriagu and Pacyna, 1988). Inspection of
Table 3D-3 reveals that the emissions from the catalytic cracker and the oil-fired power plant are
greatly enriched in rare-earth elements (e.g., La) compared to other sources.
Emissions from municipal waste incinerators are heavily enriched in Cl, arising mainly
from the combustion of plastics and metals that form volatile chlorides. The metals can
originate from cans or other metallic objects, and some metals such as Zn and Cd are also used
as additives in plastics or rubber. Many elements such as S, Cl, Zn, Br, Ag, Cd, Sn, In, Sb, and
Pb are enormously enriched compared to their crustal abundances. A comparison of the trace
elemental composition of incinerator emissions in Philadelphia, PA (shown in Table 3D-3) with
those in Washington, DC and Chicago, IL (Olmez et al., 1988) shows agreement for most
constituents to within less than a factor of two.
Very limited data exist for characterizing the chemical composition of organic compounds
present in particulate emissions from industrial-scale stationary fuel combustion. Oros and
Simoneit (2000) have reported on the abundance and distribution of organic constituents in coal
smokes burned under laboratory conditions. Their work provides the basis for further
investigation addressing the emissions of coal fired boilers.
Rogge et al. (1997a) measured the composition of the organic constituents in the PM
emissions from a 50 billion kj/h boiler that was operating at 60% capacity and was burning
Number 2 distillate fuel oil. The fine carbon PM emissions from this boiler across five tests
were composed of an average of 14% OC and 86% EC (Hildemann et al., 1991). Significant
variability in the distribution of organic compounds present in the emissions from two separate
tests was observed. Most of the identified organic mass consisted of n-alkanonic acids, aromatic
acids, n-alkanes, PAH, oxygenated PAH, and chlorinated compounds. It is unclear if these
emissions are representative of typical fuel oil combustion units in the United States. Rogge
3D-10
-------
et al. (1997b) measured the composition of hot asphalt roofing tar pots, and Rogge et al. (1993b)
measured the composition of emissions from home appliances that use natural gas.
Motor Vehicles
Particulate emissions in exhaust from gasoline- and diesel-powered vehicles have changed
significantly over the past 25 years (Sawyer and Johnson, 1995; Cadle et al., 1999). These
changes have resulted from reformulation of fuels, the wide application of exhaust-gas treatment
in gasoline-powered motor vehicles, and changes in engine design and operation. Because of
these evolving tailpipe emissions, along with the wide variability of emissions between vehicles
of the same class (Hildemann et al., 1991; Cadle et al., 1997; Sagebiel et al., 1997; Yanowitz
et al., 2000), well-defined average emissions profiles for the major classes of motor vehicles
have not been established. Two sampling strategies have been employed to obtain motor vehicle
emissions profiles: (1) the measurement of exhaust emissions from vehicles operating on
dynamometers and (2) the measurement of integrated emissions of motor vehicles driving
through roadway tunnels. Dynamometer testing can be used to measure vehicle emissions
operating over an integrated driving cycle and allows the measurement of emissions from
individual vehicles. However, dynamometer testing requires considerable resources and usually
precludes testing a very large number of vehicles. In contrast, a large number of vehicles can be
readily sampled in tunnels; however, vehicles driving through tunnels operate over limited
driving conditions, and the measurements represent contributions from a large number of vehicle
types. As a result, except in a few cases, tunnel tests have not been effective at developing
chemically speciated PM emissions profiles for individual motor vehicle classes. Rather, several
studies have measured the contribution of both OC and EC to the PM emissions from different
classes of motor vehicles operating on chassis dynamometers.
The principal PM components emitted by diesel- and gasoline-fueled vehicles are OC and
EC as shown in Tables 3D-4a and 4b. As can be seen, the variability among entries for an
individual fuel type is large and overlaps that found between different fuel types. On average,
the abundance of EC is larger than that of OC in the exhaust of diesel vehicles, whereas OC is
the dominant species in the exhaust of gasoline-fueled vehicles. Per vehicle mile, total carbon
emissions from light and heavy duty diesel vehicles can range from 1 to 2 orders of magnitude
higher than those from gasoline vehicles.
3D-11
-------
TABLE 3D-4a. ORGANIC AND ELEMENTAL CARBON FRACTIONS OF DIESEL
AND GASOLINE ENGINE PARTICIPATE MATTER EXHAUST
Engine Type
Heavy-duty diesela
Heavy-duty diesel (SPECIATE)b
Light-duty diesel °
Light-duty diesel (SPECIATE)b
Gasoline (hot stabilized) a
Gasoline ("smoker" and "high emitter" ) a>0
Gasoline (cold start) a
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%
aFujita et al. (1998) and Watson et al. (1998).
b SPECIATE database (U.S. Environmental Protection Agency, 1999).
cNorbecketal. (1998).
Source: U.S. Environmental Protection Agency (2002).
As might be expected, most of the PM emitted by motor vehicles is in the PM25 size range.
Particles in diesel exhaust are typically trimodal (consisting of a nuclei mode, an accumulation
mode, and a coarse mode) and are log-normal in form (Kittelson, 1998). More than 90% of the
total number of particles are in the nuclei mode, which contains only about 1 to 20% of the
particle mass with a mass median diameter of about 0.02 jim, whereas the accumulation mode
(with a mass median diameter of about 0.25 jim) contains most of the mass with a smaller
fraction (5 to 20%) contained in the coarse mode. Kerminin et al. (1997), Bagley et al. (1998),
and Kleeman et al. (2000) also have shown that gasoline- and diesel-fueled vehicles produce
particles that are mostly < 2.0 jim in diameter. Cadle et al. (1999) found that 91% of PM emitted
by in-use gasoline vehicles in the Denver area was in the PM2 5 size range, which increased to
97% for "smokers" (i.e., light-duty vehicles with visible smoke emitted from their tailpipes) and
98% for light-duty diesels. Durbin et al. (1999) found that about 92% of the PM was < 2.5 |im
for smokers and diesels. The mass median diameter of the PM emitted by the gasoline vehicles
sampled by Cadle et al. (1999) was -0.12 jim and increased to -0.18 jim for smokers and
diesels. Corresponding average emissions rates of PM25 found by Cadle et al. (1999) were
552 mg/mile for diesels; 222 mg/mile for gasoline smokers; and 38 mg/mile for other gasoline
3D-12
-------
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
195 a
7
15
7
6
195 a
2
6
2
OC % of
Total Carbon
70
91
76
69
89
40
50 b
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.
a A total of 195 light-duty vehicles were tested that include both gasoline- and diesel-powered vehicles.
b Fraction of PM consisting of OC 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 OC comprised 39% of the PM carbon.
0 Driving cycle comprised of multiple idle, steady acceleration, constant speed, deceleration steps (see reference
for more details).
vehicles. The values for gasoline smokers and for diesels appear to be somewhat lower than
those given in Table 3D-5, whereas the value for other gasoline vehicles falls in the range given
for low and medium gasoline vehicle emissions.
Examples of data for the trace elemental composition of the emissions from a number of
vehicle classes obtained December 1997 in Colorado, as part of the North Frontal Range Air
3D-13
-------
TABLE 3D-5. EMISSION RATES (mg/mi) FOR CONSTITUENTS OF PARTICIPATE
MATTER FROM GASOLINE AND DIESEL VEHICLES
Gasoline Vehicles
TC
OC
EC
NO3
so42-
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 ±0.1 11
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 ±0.1 14
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).
Quality Study (NFRAQS), are shown in Table 3D-5. As can be seen from Table 3D-5,
emissions of total carbon (TC), which is equal to the sum of organic carbon (OC) and elemental
carbon (EC), from gasoline vehicles are highly variable. Gillies and Gertler (2000) point out that
there is greater variability in the concentrations of trace elements and ionic species than for OC
3D-14
-------
and EC among different source profiles, e.g., SPECIATE (U.S. Environmental Protection
Agency, 1999; Lawson and Smith, 1998; Norbeck et al., 1998). They suggest that this may arise
because emissions of trace elements are not related only to the combustion process, but also
to their abundances in different fuels and lubricants and to wear and tear during vehicle
operation. Emissions from gasoline smokers are comparable to those from light-duty diesel
vehicles. Thus, older, poorly maintained gasoline vehicles could be significant sources of PM25
(Sagebiel et al., 1997; Lawson and Smith, 1998), in addition to being significant sources of
gaseous pollutants (e.g., Calvert et al., 1993). Durbin et al. (1999) point out that although
smokers constitute only 1.1 to 1.7% of the light-duty fleet in the South Coast Air Quality
Management District in California, they contribute roughly 20% of the total PM emissions from
the light-duty fleet. In general, motor vehicles that are high emitters of hydrocarbons and CO
also will tend to be high emitters of PM (Sagebiel et al.,1997; Cadle et al., 1997). Particle
emission rates, even in newer vehicles, also are correlated with vehicle acceleration; and
emissions occur predominantly during periods of heavy acceleration (Maricq et al., 1999).
Although the data shown in Table 3D-5 indicate that sulfur (mainly in the form of sulfate)
is a minor component of PM2 5 emissions, sulfur (S) may be the major component of the ultrafme
particles emitted by either diesel or internal combustion engines (Gertler et al., 2000). It is not
clear what the source of the small amount of Pb seen in the auto exhaust profile is. It is very
difficult to find suitable tracers for automotive exhaust, because Pb has been removed from
gasoline.
Several tunnel studies have measured the distribution of OC and EC in the integrated
exhaust of motor vehicle fleets comprising several classes of motor vehicles (Pierson and
Brachaczek, 1983; Weingartner et al., 1997a; Fraser et al., 1998a). The study by Fraser et al.
(1998a) found that OC constituted 46% of the carbonaceous PM emissions from the vehicles
operating in the Van Nuys tunnel in Southern California in the Summer of 1993. Although
diesel vehicles constituted only 2.8% of the vehicles measured by Fraser et al. (1998a), the
contribution of OC to the total particulate carbon emissions obtained in the Van Nuys tunnels is
in reasonable agreement with the dynamometer measurements shown in Table 3D-4b.
Very few studies have reported comprehensive analyses of the organic composition of
motor vehicle exhaust. The measurements by Rogge et al. (1993c) are the most comprehensive
but are not expected to be the best representation of current motor vehicle emissions, because
3D-15
-------
these measurements were made in the mid-1980s. Measurements reported by Fraser et al.
(1999) were made in a tunnel study conducted in 1993 and represent integrated gasoline and
diesel powered vehicle emissions. In addition, exhaust emissions from two medium-duty diesel
vehicles operating over an FTP cycle were analyzed by Schauer et al. (1999). A unique feature
of both the measurements by Fraser et al. (1999) and Schauer et al. (1999) is that they include
the quantification of unresolved complex mixture (UCM), which comprises aliphatic and cyclic
hydrocarbons that cannot be resolved by gas chromatography (GC) (Schauer et al., 1999).
Schauer et al. (1999) have shown that all of the organic compound mass in their diesel exhaust
samples could be extracted and eluted by CG/MS techniques even though not all of the organic
compound mass can identified on a single compound basis. Table 3D-6 summarizes the
composition of motor vehicle exhaust measured by Fraser et al. (1999) and Schauer et al. (1999).
Several studies have measured the distribution of poly cyclic aromatic hydrocarbons
(PAHs) in motor vehicle exhaust from on-road vehicles (Westerholm et al., 1991; Lowenthal
et al., 1994; Venkataraman et al., 1994; Westerholm and Egeback, 1994; Reilly et al., 1998;
Cadle et al., 1999, Weingartner et al., 1997b; Marr et al., 1999). Cadle et al. (1999) found that
high molecular weight PAHs (PAHs with molecular weights greater than or equal to 202 g/mole)
constitute 0.1 to 7.0% of the PM emissions from gasoline- and diesel-powered light-duty
vehicles. It is important to note, however, that PAHs with molecular weights of
202 (fluoranthene, acephenanthrylene, and pyrene), 226 (benzo[ghi]fluoranthene and
cyclopenta[cd]pyrene), and 228 (benz[a]anthracene, chrysene, and triphenylene) exist in both the
gas-phase and particle-phase at atmospheric conditions (Fraser et al., 1998b), although those
with molecular weight of 228 are predominantly associated with particles, with only traces in the
gas-phase (Arey et al., 1987). Excluding these semivolatile PAHs, the contribution of
nonvolatile PAHs to the PM emitted from the light-duty vehicles sampled by Cadle et al. (1999)
ranges from 0.013 to 0.18%. These measurements are in good agreement with the tunnel study
conducted by Fraser et al. (1999) and the heavy-duty diesel truck and bus exhaust measurements
by Lowenthal et al. (1994), except that the nonvolatile PAH emissions from the heavy-duty
diesel vehicles tested by Lowenthal et al. (1994) were moderately higher, making up -0.30% of
the PM mass emissions.
3D-16
-------
TABLE 3D-6. SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
EMITTED FROM MOTOR VEHICLES
Source
Gasoline and diesel-
powered vehicles
driving through the
Van Nuys Tunnel
(Eraser et al., 1999) a
Medium-duty diesel
vehicles operated over
an FTP Cycle
(Schauer et al., 1999)
Contribution to Dominant Contributors to
Compound Class Particulate Mass (%) Emissions of Compound Class
n-Alkanes
Petroleum biomarkers
PAHs
Aromatic acids
Aliphatic acids
Substituted aromatic
UCMb
n-Alkanes
Petroleum biomarkers
PAHs
Aliphatic acids
Aromatic acids
Saturated cycloalkanes
UCMb
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
C-21 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.
b Unresolved complex mixture.
Biomass Burning
In contrast to the mobile and stationary sources discussed earlier, emissions from biomass
burning in wood stoves and forest fires are strongly seasonal and can be highly episodic within
their peak emissions seasons. The burning of fuelwood is confined mainly to the winter months
and is acknowledged to be a major source of ambient air PM in the northwestern United States
during the heating season. Forest fires occur primarily during the driest seasons of the year in
different areas of the country and are especially prevalent during prolonged droughts.
Particulate matter produced by biomass burning outside the United States (e.g., in Central
America during the spring of 1988) also can affect ambient air quality in the United States.
3D-17
-------
An example of the composition of fine particles (PM25) produced by wood stoves is shown
in Figure 3D-4. These data were obtained in Denver during the winter of 1987-1988 (Watson
and Chow, 1994). As was the case for motor vehicle emissions, OC and EC are the major
components of particulate emissions from wood burning. It should be remembered that the
relative amounts shown for OC and EC vary with the type of stove, the stage of combustion, and
the type and condition of the fuelwood. Fine particles are dominant in smoke studies of wood
burning emissions. For instance, the mass median diameter of wood particles was found to be
-0.17 |im in a study of the emissions from burning hardwood, softwood, and synthetic logs
(Dasch, 1982).
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).
3D-18
-------
Kleeman et al. (1999) showed that the particles emitted by the combustion of wood in
fireplaces are predominately < 1.0 jim in diameter, such that the composition of fine PM (PM25)
emitted from fireplace combustion of wood is representative of the total PM emissions from this
source. Hildemann et al. (1991) and McDonald et al. (2000) reported that fireplace and wood
stove smoke consists of 48 to 71% OC and 2.9 to 15% EC. Average OC and EC 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 smoke samples but show good agreement for total carbon (OC + EC)
measurements (Chow et al., 1993).
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
(g/kg of wood
burned)
13.0
5.14
5.28
5.66
3.96
Number of
Tests
2
5
3
5
8
Percent
Organic
Carbon3
48.4
58.5
48.4
63.2
71.2
Percent
Elemental
Carbon3
5.2
15.0
2.9
7.0
9.0
References
Hildemann et al.
(1991)
McDonald et al.
(2000)
Hildemann et al.
(1991)
McDonald et al.
(2000)
McDonald et al.
(2000)
a 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.
Hawthorne et al. (1988) and Hawthorne et al. (1989) measured gas-phase and particle-
phase derivatives of guaiacol (2-methoxyphenol), syringol (2,6-dimethoxyphenol), phenol, and
catechol (1,2-benzenediol) in the downwind plume of 28 residential wood stoves and fireplaces.
Rogge et al. (1998) reported a broad range of particle-phase organic compounds in the wood
3D-19
-------
smoke samples collected by Hildemann et al. (1991), which include n-alkanes, n-alkanoic acids,
n-alkenoic acids, dicarboxylic acids, resin acids, phytosterols, PAHs, and the compounds
reported by Hawthorne et al. (1989). Supplementing these measurements, McDonald et al.
(2000) reported the combined gas-phase and particle-phase emissions of PAHs and the
compounds quantified by Hawthorne et al. (1989). The measurements by Rogge et al. (1998),
which represent a comprehensive data set of the organic compounds present in wood smoke
aerosol, are summarized in Table 3D-8. It should be noted, however, that these nearly 200
compounds account for only -15 to 25% of the OC particle mass emitted from the residential
combustion of wood. Simoneit et al. (1999) have shown that levoglucosan constitutes a
noticeable portion of the organic compound mass not identified by Rogge et al. (1998). In
addition, Elias et al. (1999) used high-temperature gas chromatography/mass spectrometry
(HTGC-MS) to measure high-molecular-weight organic compounds in smoke from South
American leaf and stem litter biomass burning. These compounds cannot be measured by the
analytical techniques employed by Rogge et al. (1998) and, therefore, are strong candidates to
make up some of the unidentified organic mass in the wood smoke samples analyzed by Rogge
et al. (1998). These compounds, which include triterpenyl fatty acid esters, wax esters,
triglycerides, and high-molecular-weight n-alkan-2-ones, are expected to be present in North
American biomass smoke originating from agricultural burning, forest fires, grassland fires, and
wood stove/fireplace emissions.
Measurements of aerosol composition, size distributions, and aerosol emissions factors
have been made in biomass burning plumes, either on towers (Susott et al., 1991) or aloft on
fixed-wing aircraft (e.g., Radke et al., 1991) or helicopters (e.g., Cofer et al., 1988). As was
found for wood stove emissions, the composition of biomass burning emissions is strongly
dependent on the stage of combustion (i.e., flaming, smoldering, or mixed), and the type of
vegetation (e.g., forest, grassland, scrub). Over 90% of the dry mass in particulate biomass
burning emissions is composed of OC (Mazurek et al., 1991). Ratios of OC to EC are highly
variable, ranging from 10:1 to 95:1, with the highest ratio found for smoldering conditions and
the lowest for flaming conditions. Emissions factors for total particulate emissions increase by
factors of two to four in going from flaming to smoldering stages in the individual fires studied
by Susott etal. (1991).
3D-20
-------
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
Compound Class I
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Dicarboxylic acids
Resin acids
Substituted phenols
Phytosterols
PAHs
Oxygenated PAH
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Dicarboxylic acids
Resin acids
Substituted phenols
Phytosterols
PAHs
Oxygenated PAHs
Contribution to
'articulate Mass (1
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
Dominant Contributors to Emissions
Yo) of Compound Class
C21 through C31
/~i /~i /~i /~i /~i /~i
Oleic and linoleic acid
Malonic acid
Abietic, dehydroabietic, isopimaric,
pimaric, and sandaracopimaric acids
Benzenediols and guaiacols
p-Sitosterol
Fluoranthene and pyrene
IH-phenalen- 1 -one
C21 through C29
y~i y~i y~i y~i
Oleic and linoleic acid
Succinic acid
Dehydroabietic acid
Benzediols, guaiacols, and syringols
p-sitosterol
No dominant compounds
IH-phenalen- 1 -one
Note: Measurements were made using a dilution sampler and no semivolatile organic compound sorbent.
Source: Rogge et al. (1998).
Particles in biomass burning plumes from a number of different fires were found to have
three distinguishable size modes: (1) a nucleation mode, (2) an accumulation mode, and
(3) a coarse mode (Radke et al., 1991). Based on an average of 81 samples, -70% of the mass
was found in particles < 3.5 jim in aerodynamic diameter. The fine particle composition was
found to be dominated by tarlike, condensed hydrocarbons; and the particles were usually
3D-21
-------
spherical in shape. Figure 3D-2 shows additional information for the size distribution of
particles produced by vegetation burning.
An example of ambient data for the composition of PM2 5 collected at a tropical site that
was heavily affected by biomass burning is shown in Table 3D-9. The samples were collected
during November 1997 on the campus of Sriwijaya University, located in a rural setting on the
island of Sumatra in Indonesia (Pinto et al., 1998). The site was subjected routinely to levels
of PM2 5 well in excess of the U.S. NAAQS as a result of biomass fires in Indonesia from
summer 1997 through spring 1998. As can be seen from a comparison of the data shown in
Table 3D-9 with those shown in Figure 3D-4, there are a number of similarities and differences
(especially with regard to the heavy metal content) in the abundances of many species. The
abundances of some crustal elements (e.g., Si, Fe) are higher in Table 3D-9 than in Figure 3D-4,
perhaps reflecting additional contributions of entrained soil dust.
TABLE 3D-9. MEAN AEROSOL COMPOSITION AT TROPICAL SITE
(SRIWIJAYA UNIVERSITY, SUMATRA, INDONESIA) AFFECTED
HEAVILY BY BIOMASS BURNING EMISSIONSa
Component
OC
EC
S042
Al
Si
Cl
K
Ca
Ti
V
Abundance (%)
76
1.2
11
BDb
9.3 x 1(T2
4.4
0.7
4.5 x 1(T2
4.2 x 1(T3
BDb
Component
Cr
Mn
Fe
Ni
Cu
Zn
As
Se
Br
Pb
Abundance (%)
BDb
BDb
3.9 x 1(T2
<3.8 x 1(T5
4.8 x 1(T4
3.1 x 1(T3
6.4 x 1(T4
2.8 x 1(T4
3.6 x 1(T2
3.1 x 1(T3
a The mean PM2 5 concentration during the sampling period (November 5 through 11, 1997) was 264 ug/m3.
b Beneath detection limit.
Source: Pinto etal. (1998).
3D-22
-------
Limited emissions data that include organic compound speciation information have been
reported for agricultural burning (Jenkins et al., 1996), forest fires (Simoneit, 1985), and
grassland burning (Standley and Simoneit, 1987). Jenkins et al. (1996) present PAH emissions
factors for the combustion of cereals (barley, corn, rice, and wheat), along with PAH emissions
factors for wood burning. Profiles of organic compounds in emissions from meat cooking
(Rogge et al., 1991) and cigarette smoke (Rogge et al., 1994) also have been obtained.
Natural Sources
Although sea-salt aerosol production is confined to salt water bodies, it is included here
because many marine aerosols can exert a strong influence on the composition of the ambient
aerosol in coastal areas. In some respects, the production of sea-salt aerosols is like that of
windblown dust, in that both are produced by wind agitation of the surface. The difference
between the two categories arises because sea-salt particles are produced from the bursting of air
bubbles rising to the sea surface. Air bubbles are formed by the entrainment of air into the water
by breaking waves. The surface energy of a collapsing bubble is converted to kinetic energy in
the form of a jet of water that can eject drops above the sea surface. The mean diameter of the
jet drops is about 15% of the bubble diameter (Wu, 1979). Bubbles in breaking waves range in
size from a few |im to several mm in diameter. Field measurements by Johnson and Cooke
(1979) of bubble size spectra show maxima in diameters at around 100 jim, with the bubble size
distribution varying as (d/d0) with d0 = 100 jim.
Because sea-salt particles receive water from the surface layer, which is enriched in
organic compounds, aerosol drops are composed of this organic material in addition to sea salt
(about 3.5% by weight in seawater). The major ionic species by mass in seawater are: Na+
(30.7%), Cr (55.0%), SO42 (7.7%), Mg2+ (3.6%), Ca2+ (1.2%), K+ (1.1%), HCCV (0.4%),
and Br" (0.2%) (Wilson, 1975). The composition of the marine aerosol also reflects the
occurrence of displacement reactions that enrich sea-salt particles in SO42 and NO3 while
depleting them of Cr and Br.
Sea salt is concentrated in the coarse mode, with a mass median diameter of ~7 jam for
samples collected in Florida, the Canary Islands, and Barbados (Savoie and Prospero, 1982).
The size distribution of sulfate is distinctly bimodal. Sulfate in the coarse mode is derived from
sea water, but sulfate in the submicron aerosol arises from the oxidation of dimethyl sulfide
3D-23
-------
(CH3SCH3) or DMS. DMS is produced during the decomposition of marine micro-organisms.
DMS is oxidized to methane sulfonic acid (MSA), a large fraction of which is oxidized to sulfate
(e.g., Hertel et al., 1994).
Apart from sea spray, other natural sources of particles include the suspension of organic
debris and volcanism. Profiles of organic compounds in vegetative detritus have been obtained
by Rogge et al. (1993d). Particles released from plants in the form of seeds, pollen, spores, leaf
waxes, and resins, range in size from ~1 to 250 jim (Warneck, 1988). Fungal spores and animal
debris, such as insect fragments, also are to be found in ambient aerosol samples in this size
range. Although material from all the foregoing categories may exist as individual particles,
bacteria usually are found attached to other dust particles (Warneck, 1988). Smaller bioaerosol
particles include viruses, individual bacteria, protozoa, and algae (Matthias-Maser and Jaenicke,
1994). In addition to natural sources, other sources of bioaerosols include industry (e.g., textile
mills), agriculture, and municipal waste disposal (Spendlove, 1974). Size distributions for
bioaerosols have not been as well characterized as for other categories of airborne particles.
Trace metals are emitted to the atmosphere from a variety of sources such as sea spray,
wind-blown dust, volcanoes, wildfires, and biotic sources (Nriagu, 1989). Biologically mediated
volatilization processes (e.g., biomethylation) are estimated to account for 30 to 50% of the
worldwide total Hg, As, and Se emitted annually; whereas other metals are derived principally
from pollens, spores, waxes, plant fragments, fungi, and algae. It is not clear, however, how
much of the biomethylated species are remobilized from anthropogenic inputs. Median ratios of
the natural contribution to globally averaged total sources for trace metals are estimated to be
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),
0.25 (V), and 0.34 (Zn), suggesting a significant natural source for many trace elements.
It should be noted, however, that these estimates are based on emissions estimates that have
uncertainty ranges of an order of magnitude.
3D-24
-------
REFERENCES
Arey, I; Zielinska, B.; Atkinson, R.; Winer, A. M. (1987) Polycyclic aromatic hydrocarbon and nitroarene
concentrations in ambient air during a wintertime high-NO^ episode in the Los Angeles basin.
Atmos. Environ. 21: 1437-1444.
Bagley, S. T.; Gratz, L. D.; Johnson, J. H.; McDonald J. F. (1998) Effects of an oxidation catalytic converter and a
biodiesel fuel on the chemical, mutagenic, and particle size characteristics of emissions from a diesel engine.
Environ. Sci. Technol. 32: 1183-1191.
Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of paniculate diesel
exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
Cadle, S. H.; Mulawa, P. H.; Ball, J.; Donase, C; Weibel, A.; Sagebiel, J. C; Knapp, K. T.; Snow, R. (1997)
Paniculate emission rates from in-use high-emitting vehicles recruited in Orange county, California. Environ.
Sci. Technol. 31:3405-3412.
Cadle, S. H.; Mulawa, P.; Hunsanger, E. C.; Nelson, K.; Ragazzi, R. A.; Barrett, R.; Gallagher, G. L.; Lawson,
D. R.; Knapp, K. T.; Snow, R. (1999) Light-duty motor vehicle exhaust paniculate matter measurement in the
Denver, Colorado, area. J. Air Waste Manage. Assoc. 49: PM-164-174.
Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some
reflections on controlling vehicle emissions. Science (Washington, DC) 261: 37-45.
Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere;
April 1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
Chow, J. C.; Watson, J. G.; Houck, J. E.; Pritchett, L. C.; Rogers, C. F.; Frazier, C. A.; Egami, R. T.; Ball, B. M.
(1994) A laboratory resuspension chamber to measure fugitive dust size distributions and chemical
compositions. Atmos. Environ. 28: 3463-3481.
Cofer, W. R., Ill; Levine, J. S.; Sebacher, D. I.; Winstead, E. L.; Riggin, P. J.; Brass, J. A.; Ambrosia, V. G. (1988)
Paniculate emissions from a mid-latitude prescribed chaparral fire. J. Geophys. Res. [Atmos.] 93: 5207-5212.
Dasch, J. M. (1982) Paniculate and gaseous emissions from wood-burning fireplaces. Environ. Sci. Technol.
16: 639-645.
Durbin, T. D.; Smith, M. R.; Norbeck, J. M.; Truex, T. J. (1999) Population density, paniculate emission
characterization, and impact on the paniculate inventory of smoking vehicles in the South Coast Air Quality
Management District. J. Air Waste Manage. Asssoc. 49: 28-38.
Elias, V. O.; Simoneit, B. R. T.; Pereira, A. S.; Cabral, J. A.; Cardoso, J. N. (1999) Detection of high molecular
weight organic tracers in vegetation smoke samples by high-temperature gas chromatography-mass
spectrometry. Environ. Sci. Technol. 33: 2369-2376.
Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. (1998a) Gas-phase and particle-phase organic compounds emitted
from motor vehicle traffic in a Los Angeles roadway tunnel. Environ. Sci. Technol. 32: 2051-2060.
Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T.; Rasmussen, R. A. (1998b) Air quality model evaluation data for
organics. 5. C6-C22 nonpolar and semipolar aromatic compounds. Environ. Sci. Technol. 32: 1760-1770.
Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. (1999) Paniculate organic compounds emitted from motor vehicle
exhaust and in the urban atmosphere. Atmos. Environ. 33: 2715-2724.
Fujita, E.; Watson, J. G.; Chow, J. C.; Robinson, N. F.; Richards, L. W.; Kumar, N. (1998) Northern front range air
quality study, volume C: source apportionment and simulation methods and evaluation. Reno, NV: Desert
Research Institute; prepared for Colorado State University, Cooperative Institute for Research in the
Atmosphere.
Gertler, A. W.; Abu-Allaban, M.; Coulombe, W.; Gilles, J. A.; Pierson, W. R.; Rogers, C. F.; Sagebiel, J. C.;
Tarnay, L. (2000) Preliminary results of a tunnel study to characterize mobile source paniculate emissions.
Presented at: PM2000: paniculate matter and health—the scientific basis for regulatory decision-making,
specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management
Association.
Gillies, J. A.; Gertler, A. W. (2000) Comparison and evaluation of chemically speciated mobile source PM2 5
paniculate matter profiles. J. Air Waste Manage. Assoc. 50: 1459-1480.
Hawthorne, S. B.; Miller, D. J.; Barkley, R. M.; Krieger, M. S. (1988) Identification of methoxylated phenols as
candidate tracers for atmospheric wood smoke pollution. Environ. Sci. Technol. 22: 1191-1196.
3D-25
-------
Hawthorne, S. B.; Krieger, M. S.; Miller, D. J.; Mathiason, M. B. (1989) Collection and quantitation of
methoxylated phenol tracers for atmospheric pollution from residential wood stoves. Environ. Sci. Technol.
23:470-475.
Hertel, O.; Christensen, J.; Hov, 0. (1994) Modelling of the end products of the chemical decomposition of DMS in
the marine boundary layer. Atmos. Environ. 28: 2431-2449.
Hildemann, L. M.; Markowski, G. R.; Jones, M. C.; Cass, G. R. (1991) Submicrometer aerosol mass distributions of
emissions from boilers, fireplaces, automobiles, diesel trucks, and meat-cooking operations. Aerosol Sci.
Technol. 14: 138-152.
Hopke, P. K.; Lamb, R. E.; Natusch, D. F. S. (1980) Multielemental characterization of urban roadway dust.
Environ. Sci. Technol. 14: 164-172.
Houck, J. E.; Chow, J. C.; Watson, J. G.; Simons, C. A.; Pritchett, L. C.; Goulet, J. M.; Frazier, C. A. (1989)
Determination of particle size distribution and chemical composition of paniculate matter from selected
sources in California: volume I and executive summary (final report). Sacramento, CA: California Air
Resources Board.
Houck, J. E.; Goulet, J. M.; Chow, J. C.; Watson, J. G.; Pritchett, L. C. (1990) Chemical characterization of emission
sources contributing to light extinction. In: Mathai, C. V., ed. Visibility and fine particles: an A&WMA/EPA
international specialty conference; October 1989; Estes Park, CO. Pittsburgh, PA: Air & Waste Management
Association; pp. 437-446. (A&WMA transactions series no. TR-17).
Huffman, G. P.; Huggins, F. E.; Shah, N.; Huggins, R.; Linak, W. P.; Miller, C. A.; Pugmire, R. J.; Meuzelaar, H. L.
C.; Seehra, M. S.; Manivannan, A. (2000) Characterization of fine paniculate matter produced by combustion
of residual fuel oil. J. Air Waste Manage. Assoc. 50: 1106-1114.
Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning.
Atmos. Environ. 30: 3825-3835.
Johnson, B. D.; Cooke, R. C. (1979) Bubble populations and spectra in coastal waters: a photographic approach.
J. Geophys. Res. C: Oceans Atmos. 84: 3761-3766.
Kerminen, V. M.; Makela, T. E.; Ojanen, C. H.; Hillamo, R. E.; Vilhunen, J. K.; Rantanen, L.; Havers, N.;
Von Bohlen, A.; Klockow, D. (1997) Characterization of the paniculate phase in the exhaust from a diesel
car. Environ. Sci. Technol. 31: 1883-1889.
Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
Kleeman, M. J.; Schauer, J. J.; Cass, G. R. (1999) Size and composition distribution of fine paniculate matter
emitted from wood burning, meat charbroiling, and cigarettes. Environ. Sci. Technol. 33: 3516-3523.
Kleeman, M. J.; Schauer, J. J.; Cass, G. R. (2000) Size and composition distribution of fine paniculate matter
emitted from motor vehicles. Environ. Sci. Technol. 34: 1132-1142.
Lawson, D. R.; Smith, R. E. (1998) The northern front range air quality study: a report to the Governor and General
Assembly. Fort Collins, CO: Colorado State University; December.
Linak, W. P.; Miller, C. A.; Wendt, J. O. L. (2000) Comparison of particle size distributions and elemental
partitioning from the combustion of pulverized coal and residual fuel oil. J. Air Waste Manage. Assoc.
50: 1532-1544.
Lowenthal, D. H.; Zielinska, B.; Chow, J. C.; Watson, J. G.; Gautam, M.; Ferguson, D. H.; Neuroth, G. R.; Stevens,
K. D. (1994) Characterization of heavy-duty diesel vehicle emissions. Atmos. Environ. 28: 731-743.
Maricq, M. M.; Podsiadlik, D. H.; Chase, R. E. (1999) Gasoline vehicle particle size distributions: comparison of
steady state, FTP, andUS06 measurements. Environ. Sci. Technol. 33: 2007-2015.
Marr, L. C.; Kirchstetter, T. W.; Harley, R. A.; Miguel, A. H.; Hering, S. V; Hammond, S. K. (1999)
Characterization of polycyclic aromatic hydrocarbons in motor vehicle fuels and exhaust emissions.
Environ. Sci. Technol. 33: 3091-3099.
Mason, B. (1966) Principles of geochemistry. 3rd ed. New York, NY: John Wiley & Sons, Inc.
Matthias-Maser, S.; Jaenicke, R. (1994) Examination of atmospheric bioaerosol particles with radii > 0.2 um.
J. Aerosol Sci. 25: 1605-1613.
Mazurek, M. A.; Cofer, W. R., Ill; Levine, J. S. (1991) Carbonaceous aerosols from prescribed burning of a boreal
forest ecosystem. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
implications. Cambridge, MA: MIT Press; pp. 258-263.
McDonald, J. D.; Zielinska, B.; Fujita, E. M.; Sagebiel, J. C.; Chow, J. C.; Watson, J. G. (2000) Fine particle and
gaseous emission rates from residential wood combustion. Environ. Sci. Technol. 34: 2080-2091.
Miller, C. A.; Linak, W. P.; King, C.; Wendt, J. O. L. (1998) Fine particle emissions from heavy fuel oil combustion
in a firetube package boiler. Combust. Sci. Technol. 134: 477-502.
3D-26
-------
Norbeck, J. M; Durbin, T. D.; Traex, T. J. (1998) Measurement of primary paniculate matter emissions from
light-duty motor vehicles. Riverside, CA: University of California, College of Engineering, Center for
Environmental Research and Technology; prepared for Coordinating Research Council, Inc. and South Coast
Air Quality Management District.
Nriagu, J. O. (1989) A global assessment of natural sources of atmospheric trace metals. Nature (London)
338: 47-49.
Nriagu, J. O.; Pacyna, J. M. (1988) Quantitative assessment of worldwide contamination of air, water and soils by
trace metals. Nature (London) 333: 134-139.
Olmez, I.; Sheffield, A. E.; Gordon, G. E.; Houck, J. E.; Pritchett, L. C.; Cooper, J. A.; Dzubay, T. G.; Bennett, R. L.
(1988) Compositions of particles from selected sources in Philadelphia for receptor modeling applications.
JAPCA38: 1392-1402.
Oros, D. R.; Simoneit, B. R. T. (2000) Identification and emission rates of molecular tracers in coal smoke
paniculate matter. Fuel 79: 515-536.
Pierson, W. R.; Brachaczek, W. W. (1983) Paniculate matter associated with vehicles on the road. II. Aerosol Sci.
Technol. 2: 1-40.
Pinto, J. P.; Grant, L. D.; Hartlage, T. A. (1998) Report on U.S. EPA air monitoring of haze from S.E. Asiabiomass
fires. Research Triangle Park, NC: U.S. Environmental Protection Agency, National Center for
Environmental Assessment; report no. EPA/600/R-98/071.
Radke, L. F.; Hegg, D. A.; Hobbs, P. V.; Nance, J. D.; Lyons, J. H.; Laursen, K. K.; Weiss, R. E.; Riggan, P. J.;
Ward, D. E. (1991) Paniculate and trace gas emissions from large biomass fires in North America. In: Levine,
J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric implications. Cambridge, MA: MIT
Press; pp. 209-224.
Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. (1998) Real-time characterization of the organic
composition and size of individual diesel engine smoke particles. Environ. Sci. Technol. 32: 2672-2679.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine organic
aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks.
Environ. Sci. Technol. 27: 1892-1904.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
aerosol. 5. Natural gas home appliances. Environ. Sci. Technol. 27: 2736-2744.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci.
Technol. 27:636-651.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993d) Sources of fine organic
aerosol. 4. Paniculate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol.
27:2700-2711.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1994) Sources of fine organic
aerosol. 6. Cigarette smoke in the urban atmosphere. Environ. Sci. Technol. 28: 1375-1388.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997a) Sources of fine organic
aerosol. 8. Boilers burning no. 2 distillate fuel oil. Environ. Sci. Technol. 31: 2731-2737.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997b) Sources of fine organic
aerosol. 7. Hot asphalt roofing tar pot fumes. Environ. Sci. Technol. 31: 2726-2730.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1998) Sources of fine organic
aerosol. 9. Pine, oak, and synthetic log combustion in residential fireplaces. Environ. Sci. Technol. 32: 13-22.
Sagebiel, J. C.; Zielinska, B.; Walsh, P. A.; Chow, J. C.; Cadle, S. H.; Mulawa, P. A.; Knapp, K. T.; Zweidinger,
R. B.; Snow, R. (1997) PM-10 exhaust samples collected during IM-240 dynamometer tests of in-service
vehicles in Nevada. Environ. Sci. Technol. 31: 75-83.
Savoie, D. L.; Prospero, J. M. (1982) Particle size distribution of nitrate and sulfate in the marine atmosphere.
Geophys. Res. Lett. 9: 1207-1210.
Sawyer, R. F.; Johnson, J. H. (1995) Diesel emissions and control technology. Diesel exhaust: a critical analysis of
emissions, exposure, and health effects. Cambridge, MA: Health Effects Institute.
Schauer, J. J.; Cass, G. R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants
using organic compounds as tracers. Environ. Sci. Technol. 34: 1821-1832.
3D-27
-------
Schauer, J. I; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. (1999) Measurement of emissions from air pollution
sources. 2. Q through C30 organic compounds from medium duty diesel trucks. Environ. Sci. Technol.
33: 1578-1587.
Simoneit, B. R. T. (1985) Application of molecular marker analysis to vehicular exhaust for source reconciliations.
Int. J. Environ. Anal. Chem. 22: 203-233.
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.
(1999) Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles. Atmos. Environ.
33: 173-182.
Spendlove, J. C. (1974) Industrial, agricultural, and municipal microbial aerosol problems. In: Current problems in
aerobiology: proceedings of the thirtieth general meeting of the Society for Industrial Microbiology; August
1973; Evanston, IL. Dev. Ind. Microbiol. 15: 20-27.
Standley, L. J.; Simoneit, B. R. T. (1987) Characterization of extractable plant wax, resin, and thermally matured
components in smoke particles from prescribed burns. Environ. Sci. Technol. 21: 163-169.
Susott, R. A.; Ward, D. E.; Babbitt, R. E.; Latham, D. J. (1991) The measurement of trace emissions and combustion
characteristics for a mass fire. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and
biospheric implications. Cambridge, MA: MIT Press; pp, 245-257.
Turekian, K. K. (1971) Elements, geochemical distribution of. In: Encyclopedia of science and technology, v. 4.
New York, NY: McGraw-Hill Book Company; pp. 627-630.
U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
U.S. Environmental Protection Agency. (1999) SPECIATE. Washington, DC: Technology Transfer Network-
Clearinghouse for Inventories & Emissions Factors. Available:
http://www.epa.gov/ttn/chief/software/speciate/index.html (2 June 2003).
U.S. Environmental Protection Agency. (2002) Health assessment document for diesel engine exhaust. Washington,
DC: Office of Research and Development; report no. EPA/600/8-90/057.
Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci.
Technol. 28: 555-562.
Vinogradov, A. P. (1959) The geochemistry of rare and dispersed chemical elements in soils. 2nd ed. New York,
NY: Consultants Bureau, Inc.
Warneck, P. (1988) Chemistry of the natural atmosphere. New York, NY: Academic Press, Inc.
Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ.
19: 241-266.
Watson, J. G.; Fujita, E. M.; Chow, J. C.; Zielinska, B.; Richards, L. W.; Neff, W.; Dietrich, D. (1998) Northern Iron
range air quality study final report. Reno, NV: Desert Research Institute; prepared for Colorado State
University, Cooperative Institute for Research in the Atmosphere.
Weingartner, E.; Burtscher, H.; Baltensperger, U. (1997a) Hygroscopic properties of carbon and diesel soot
particles. Atmos. Environ. 31: 2311-2327.
Weingartner, E.; Keller, C.; Stahel, W. A.; Burtscher, H.; Baltensperger, U. (1997b) Aerosol emission in a road
tunnel. Atmos. Environ. 31: 451-462.
Westerholm, R.; Egeback, K.-E. (1994) Exhaust emissions from light- and heavy-duty vehicles: chemical
composition, impact of exhaust after treatment, and fuel parameters. In: Symposium of risk assessment of
urban air: emissions, risk identification, and risk quantitation; May-June 1992; Stockholm, Sweden. Environ.
HealthPerspect. 102(suppl. 4) 13-23.
Westerholm, R. N.; Almen, J.; Li, H.; Rannug, J. U.; Egeback, K.-E.; Gragg, K. (1991) Chemical and biological
characterization of particulate-, semivolatile-, and gas-phase-associated compounds in diluted heavy-duty
diesel exhausts: a comparison of three different semivolatile-phase samplers. Environ. Sci. Technol.
25: 332-338.
Wilson, T. R. S. (1975) Salinity and the major elements of sea water. In: Riley, J. P.; Skirrow, G., eds. Chemical
oceanography: v. 1. 2nd ed. London, United Kingdom: Academic Press Inc.; pp. 365-413.
Wu, J. (1979) Spray in the atmospheric surface layer: review and analysis of laboratory and oceanic results.
J. Geophys. Res. C: Oceans Atmos. 84: 1693-1704.
3D-28
-------
APPENDIX 3E
Characterization of PM25, PM10, and PM10_25
Concentrations at IMPROVE Sites
In this appendix, concentrations of PM25, PM10_25 and PM10 are presented based on daily
values in the PM data set obtained at 18 sites in the IMPROVE network. The purpose of these
analyses is to provide summaries of concentrations of PM25 and PM10_25 that could be used for
characterizing human and ecosystem exposures to PM under the least polluted conditions that
are found in the United States and to provide upper limits on policy relevant background (PRB)
PM concentrations (cf, Section 3.3.3). This is accomplished by summarizing nonsulfate PM25
and PM10_2 5 concentrations and their variability at relatively remote monitoring sites (RRMS)
that are not obviously influenced by local and/or regional anthropogenic pollution sources in
the United States. The rationale for looking at nonsulfate PM2 5 is that the sulfate component
of PM2 5 is almost all anthropogenic in origin and, therefore, nonsulfate PM25 provides a better
estimate of upper limits on PRB PM concentrations.
The potential for receiving contributions from anthropogenic sources exists at every
monitoring site. Thus, sites were chosen for this analysis where it was judged that anthropogenic
sources do not contribute extensively to the monitored concentrations. These sites are
characterized by lower interannual variability and smaller ranges of values across the percentile
distribution in concentrations and by lower probabilities of being affected by transport from
regional pollution sources than other sites. Such sites are almost all in the western United States.
This analysis includes only two sites in the eastern United States and two areas in the midwest.
The 24-h average data available at the selected 18 IMPROVE sites from 1990 through
2002 were summarized on an annual basis; and annual mean values and annual 90th percentiles
of the 24-h concentrations were calculated. The data capture requirement for these summaries
required that each site have least 11 observations in each calendar quarter, and only years in
which this requirement was satisfied for all four quarters were used. The monitoring site in
Yellowstone National Park was relocated in 1996; thus, the designations for the two locations
are Yellowstone National Park 1 and Yellowstone National Park 2. The Voyageurs National
3E-1
-------
Park site was also relocated. The designations for the two locations are analogous to those used
for Yellowstone National Park.
Both PM2 5 and PM10 mass were examined with the sulfate component removed to obtain
better insight into non-anthropogenic background. Daily fine sulfate concentrations were
calculated by 4.125 * S, where S is the measured mass of fine sulfur (Malm, et al., 1994). Since
sulfate is not a significant component of coarse fraction particles and measurements are not
available, we calculated an estimate of the coarse PM (PM10_25) by the difference between PM10
and PM2 5.
Information about the range of annual mean nonsulfate (and total) PM2 5 and PM10 mass
concentrations, and PM10_25 mass concentrations, is summarized in Table 3E-1, and the range of
90th percentile concentrations is summarized in Table 3E-2. The 90th percentile was chosen to
reduce the impact of infrequent, episodic events that can result in exceptionally high daily
values. As can be seen from Table 3E-1, the lowest annual average nonsulfate PM25
concentrations among all sites examined were recorded at Denali National Park (AK), with
concentrations there ranging from an annual mean of 0.7 to 2.4 |ig/m3. The site with the next
lowest observed annual mean (among the 18 sites analyzed) was Bridger Wilderness (WY),
where annual means ranged from 1.5 to 2.2 |ig/m3. The highest annual mean nonsulfate PM25
concentrations at the sites examined were observed in Glacier National Park (MT), where values
ranged from 3.8 to 5.5 |ig/m3. Annual means of coarse PM ranged from a low of 1.1 to
5.6 |ig/m3 at Denali National Park to a high of 5.6 to 10.8 |ig/m3 at Big Bend National Park, TX.
Over the period of the observational record, a substantial range in annual mean concentrations in
all three size ranges is observed, indicating the importance of examining a number of years in
attempting to characterize long-term averages. Inspection of Table 3E-1 shows that the range in
annual average PM10_2 5 concentrations was generally larger than for PM2 5. However, it should
be noted that PM10_2 5 concentrations are obtained by a difference technique and, so, are more
strongly affected by errors in the determination of both PM25 and PM10.
As can be seen in Table 3E-2, 90th percentile 24-h nonsulfate PM2 5 levels ranged from
1.5 to 10.5 |ig/m3. The 90th percentiles of 24-h coarse PM ranged from 1.9 to 27.6 |ig/m3.
As can be seen from Table 3E-2, concentrations of PM in all three size fractions at the
90th percentile level are often several times higher than the mean concentrations at all sites.
In addition, the range of values at the 90th percentile level is much greater than that at lower
3E-2
-------
TABLE 3E-1. RANGES OF ANNUAL MEAN PM CONCENTRATIONS AT IMPROVE
MONITORING SITES (jig/m3)
W
Site
Acadia National Park, ME
Big Bend National Park, TX
Boundary Waters Canoe Area, MN
Bryce Canyon National Park, UT
Bridger Wilderness, WY
Canyonlands National Park, UT
Denali National Park, AK
Gila Wilderness, NM
Glacier National Park, MT
Lassen Volcanic National Park, CA
Lone Peak Wilderness, UT
Lye Brook Wilderness, VT
Redwood National Park, CA
Three Sisters Wilderness, OR
Voyageurs National Park 1, MN
Voyageurs National Park 2, MN
Yellowstone National Park 1, WY
Yellowstone National Park 2, WY
PM25
Nonsulfate
2.6-4.7
2.7-4.9
2.6-3.9
1.7-2.4
1.5-2.2
1.9-3.2
0.7-2.4
2.4-3.4
3.8-5.5
1.7-4.5
3.1-5.3
2.3-4.8
2.8-4.6
2.0-5.4
3.2-3.5
2.6-5.4
2.0-3.0
1.7-4.1
(Total)
(4.9-8.2)
(5.0-7.8)
(4.4-5.8)
(2.6-3.4)
(2.1-2.9)
(2.8-4.0)
(1.1-3.2)
(3.4-4.5)
(4.8-6.5)
(2.1-5.1)
(4.1-6.9)
(4.5-8.8)
(3.6-5.4)
(2.7-6.5)
(5.1-5.9)
(4.1-7.2)
(2.6-3.6)
(2.3-4.7)
PM10
Nonsulfate
4.6-11.3
8.8-15.7
5.0-10.2
4.4-7.6
3.7-6.5
5.1-10.5
2.0-7.5
4.9-7.9
7.6-14.2
4.0-8.1
7.1-10.9
4.2-9.7
6.0-10.6
4.0-8.1
5.7-11.2
5.2-10.8
6.0-9.2
3.6-9.0
(Total)
(7.3-15.0)
(11.3-18.6)
(7.0-12.0)
(5.3-8.5)
(4.3-7.3)
(6.3-11.7)
(2.4-8.3)
(6.0-9.2)
(8.5-15.2)
(4.6-8.5)
(8.1-12.5)
(7.0-13.6)
(7.2-11.7)
(4.6-9.1)
(8.1-13.1)
(7.0-12.5)
(6.6-9.9)
(4.2-9.6)
Coarse PM
1.8-6.0
5.6-10.8
2.3-7.3
2.5-5.6
1.9-4.7
3.2-8.0
1.1-5.6
2.5-5.0
3.7-9.6
1.8-6.4
3.7-6.0
1.6-4.8
3.3-6.5
1.9-4.4
2.8-7.8
2.6-5.3
3.8-7.0
1.9-5.0
-------
TABLE 3E-2. RANGES OF ANNUAL 90th PERCENTILE VALUES OF DAILY PM CONCENTRATIONS AT
IMPROVE MONITORING SITES (jig/m3)
Site
Acadia National Park, ME
Big Bend National Park, TX
Boundary Waters Canoe Area, MN
Bryce Canyon National Park, UT
Bridger Wilderness, WY
Canyonlands National Park, UT
Denali National Park, AK
Gila Wilderness, NM
Glacier National Park, MT
Lassen Volcanic National Park, CA
Lone Peak Wilderness, UT
Lye Brook Wilderness, VT
Redwood National Park, CA
Three Sisters Wilderness, OR
Voyageurs National Park 1, MN
Voyageurs National Park 2, MN
Yellowstone National Park 1, WY
Yellowstone National Park 2, WY
PM25
Nonsulfate
4.8-10.5
4.7-9.5
5.5-8.0
2.9-5.0
2.8-4.6
3.2-7.3
1.5-5.1
4.1-6.3
7.2-10.0
3.7-7.7
5.4-9.8
5.4-9.1
5.2-9.4
4.5-8.0
5.7-6.4
5.1-5.5
3.9-5.7
3.4-4.9
(Total)
(8.8-16.5)
(8.7-12.8)
(8.1-12.0)
(4.4-6.1)
(3.9-5.8)
(4.5-7.9)
(2.2-7.4)
(5.6-8.0)
(8.0-11.0)
(4.0-9.1)
(6.7-11.1)
(11.4-18.2)
(6.3-9.9)
(5.8-10.0)
(7.7-11.3)
(7.4-8.1)
(4.7-6.6)
(4.1-6.0)
PM10
Nonsulfate
9.0-20.2
15.9-29.0
9.4-13.8
7.4-13.8
7.3-14.5
9.5-16.1
3.6-17.2
8.3-14.4
16.2-36.5
7.6-15.6
11.7-19.6
8.5-15.8
11.8-21.8
7.3-16.2
13.9-22.0
9.6-11.9
11.7-20.6
6.7-10.3
(Total)
(12.7-26.0)
(18.5-34.5)
(13.6-17.1)
(8.9-14.8)
(8.3-15.1)
(10.9-16.9)
(4.3-17.9)
(10.2-15.9)
(16.5-36.8)
(8.5-16.0)
(13.4-21.5)
(14.4-22.9)
(13.2-23.7)
(8.5-18.0)
(17.9-23.1)
(12.0-15.5)
(12.3-21.3)
(7.4-10.8)
Coarse PM
3.6-12.0
10.4-18.7
4.8-7.5
4.7-12.0
3.5-12.3
6.3-14.7
1.9-11.7
4.2-9.6
8.1-27.6
3.7-14.7
6.6-10.9
3.3-8.8
5.9-13.0
4.0-9.6
9.0-17.4
5.4-6.0
8.8-16.5
4.2-5.8
-------
levels. These findings indicate that extreme events are important for shaping the frequency
distribution of PM that is observed. Wildfires play a major role in defining the year-to-year
variability at specific sites, especially in the West. Notable examples include the massive forest
fire in 1988 in Yellowstone National Park and those that occurred in 2000 in many western
states. Dust storms also play a role in arid climates. The variability in annual average PM10_2 5
concentrations is again related to several factors. Variability in factors that govern the
production of soil dust from the surface and the production of primary biological aerosol
particles may be largely responsible. Wildfires also produce PM in this size range (cf,
Appendix 3D).
Examples of the interannual variability in the percentile distributions of PM25
concentrations for 1997 through 2001 are shown in Figure 3E-l(a,b) for two sites. As can be
seen, at concentrations less than the 90th percentile (P90) level, year-to-year differences in PM2 5
can be less than 1 |ig/m3 at sites such as Bridger, WY. Differences at the P90 level are much
larger, both on an absolute and on a relative basis at eastern sites.
Percentile distributions by season are shown in Figure 3E-2(a,b) for PM2 5 and
Figure 3E-3(a,b) for PM10_2 5. For most sites, the variability in concentrations on a quarterly
basis increases substantially beyond the P90 level, as can be seen from inspection of
Figure 3E-2(a,b) for PM2 5 and Figure 3E-3(a,b) for PM10_2 5. Perhaps the most striking features
seen in Figure 3E-2(a,b) and Figure 3E-3(a,b) are the concentration changes associated with the
95th (P95) and 99th (P99) percentile events, which represent extreme value events. Most of these
events occur in the third calendar quarter; however, high concentrations also do occur at some
sites during the second and fourth calendar quarters. If locally derived climatologic seasons
were used instead of calendar quarters, a more accurate depiction of the seasonal variability
of these events may have been obtained. In most cases, there is consistency in the behavior
of PM2 5 and PM10, which suggests that the episodes of higher concentrations could be associated
with sources that produce mainly PM2 5 such as wildfires and/or anthropogenic combustion
sources. Wildfires are limited to hotter and drier times of the year, but anthropogenic sources
can contribute to high concentrations during other seasons. Additional factors which would tend
to produce a third quarter maximum include: the enhanced production of secondary PM from
anthropogenic and biogenic precursors during summer months; wildfires that are located in the
East or elsewhere during summer; and surface dust produced locally and/or in northern Africa.
3E-5
-------
REFERENCES
Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
3E-6
-------
Bridger Wilderness WY
Min
10
25
Max
Min
Yellowstone National Park WY
PWIa.5
10
P75 P90
P95 P99
Max
Figure 3E-la,b. Interannual variability in 24-h average PM2 5 concentrations observed at
selected IMPROVE sites: (a) Bridger Wilderness, WY; (b) Yellowstone
National Park, WY.
3E-7
-------
Bridger Wilderness WY
PM2.5 by Season
Min
25
50
75
90
95
99
Max
Quarter 1
Quarter 2
A Quarter 3
Quarter 4
25
20
Yellowstone National Park2 WY
PM2.5 by Season
»
E
*3>
I <6
G
0>
o
c
o
O
u>
-------
C
o
1!
*l«l
E
0>
u
G
o
o
Min
Bridger Wilderness WY
PM 10-2.5 by Season
99
Max
Quarter 1
Quarter 2
Quarter3
Quarter 4
Min
Yellowstone National Park 2 WY
PM 10-2.5 by Season
25
50
75
90
95
99
Max
Quarter 1
Quarter 2
A Quarter3
Quarter 4
Figure 3E-3a,b. Seasonal variability in 24-h average PM10_2 5 concentrations observed at
selected IMPROVE sites: (a) Bridger Wilderness, WY; (b) Yellowstone
National Park, WY.
3E-9
-------
4. ENVIRONMENTAL EFFECTS OF AIRBORNE
PARTICULATE MATTER
4.1 INTRODUCTION
This chapter assesses information providing inputs to U.S. EPA decision making on
secondary National Ambient Air Quality Standards (NAAQS) aimed at protecting against
welfare effects of ambient airborne particulate matter (PM). Specifically, it assesses the effects
of atmospheric PM on the environment, including: (a) direct and indirect effects on vegetation
and natural ecosystem integrity, (b) effects on visibility, and (c) effects on man-made materials,
as well as (d) relationships of atmospheric PM to climate change processes.
4.2 EFFECTS OF AMBIENT AIRBORNE PARTICULATE MATTER ON
VEGETATION AND NATURAL ECOSYSTEMS
Introduction
The effects of airborne particles are manifested via physical and chemical effects exerted at
the individual plant level. However, plants are key members of ecosystems, which are
structurally complex communities comprised of populations of plants, animals (including
humans), insects, and microorganisms that interact with one another and with their non-living
(abiotic) chemical and physical environment in which they exist (Odum, 1989; U.S.
Environmental Protection Agency, 1993). All life on Earth is dependent on chemical energy in
the form of carbon compounds to sustain their life processes. Terrestrial vegetation, via the
process of photosynthesis, provides approximately half of the carbon that annually cycles
between the Earth and the atmosphere (Chapin and Ruess, 2001).
Ecosystems respond to stresses through their constituent organisms. The responses of
plant species and populations to environmental perturbations (such as those caused by
atmospheric PM) depend on their genetic constitution (genotype), their life cycles, and their
microhabitats. Stresses that produce changes in their physical and chemical environment apply
selection pressures on individual organisms (Treshow, 1980). The changes that occur within
populations and plant communities reflect these new and different pressures. A common
4-1
-------
response in a community under stress is the elimination of the more sensitive populations and an
increase in abundance of species that tolerate or are favored by stress (Woodwell, 1970,
Guderian et al., 1985).
The present section is organized to discuss: (1) factors affecting deposition of airborne
PM on plants and ecosystems and then (2) the effects of PM deposition on individual plants,
plant populations, forest trees and terrestrial ecosystems. As such, the section is organized to
follow, in rough outline, the Framework for Assessing and Reporting on Ecological Condition
recommended in a report by the Ecological Processes and Effects Committee (EPEC) of the
EPA's Science Advisory Board (Science Advisory Board, 2002), which states:
"The purpose of this report is to provide the Agency with a sample framework that
may serve as a guide for designing a system to assess, and then report on, ecological
condition at local, regional, or national scale. The sample framework is intended as
an organizing tool that may help the Agency decide what ecological attributes to
measure and how to aggregate those measurements into an understandable picture of
ecological integrity." This framework is not actually a risk assessment per se, but it
can be used to "construct a report of ecological condition" that characterizes the
ecological integrity of an ecosystem based on "the relationship between common
anthropogenic stressors and one or more of the six Essential Ecological Attributes."
The EPEC report provides a useful approach for organizing discussions of stressor effects on
ecosystem components at successive levels of complexity.
4.2.1 Ecological Attributes
The EPEC Framework provides a checklist of generic ecological attributes that should be
considered when evaluating the integrity of ecological systems (see Table 4-1). The six generic
ecological attributes, termed Essential Ecological Attributes (EEAs), represent groups of related
ecological characteristics (Science Advisory Board, 2002; Harwell et al., 1999) and include:
landscape conditions, biotic conditions, chemical and physical characteristics, ecological
processes, hydrology and geomorphology, and natural disturbance regimes. All of the EEAs
are interrelated (i.e., changes in one EEA may directly or indirectly affect other EEAs).
The first three ecological attributes listed in Table 4-1 are primarily "patterns," whereas the
last three are "processes." Ecological science has used "patterns" and "processes" as terms to
describe features of ecological systems for many years (e.g., Bormann and Likens, 1979).
4-2
-------
TABLE 4-1. ESSENTIAL ECOLOGICAL ATTRIBUTES AND
REPORTING CATEGORIES
Landscape Condition
• Extent of Ecological System/Habitat Types
• Landscape Composition
• Landscape Pattern and Structure
Biotic Condition
• Ecosystems and Communities
- Community Extent
- Community Composition
- Trophic Structure
- Community Dynamics
- Physical Structure
• Species and Populations
- Population Size
- Genetic Diversity
- Population Structure
- Population Dynamics
- Habitat Suitability
• Organism Condition
- Physiological Status
- Symptoms of Disease or Trauma
- Signs of Disease
Chemical and Physical Characteristics
(Water, Air, Soil, and Sediment)
• Nutrient Concentrations
- Nitrogen
- Phosphorus
- Other Nutrients
• Trace Inorganic and Organic Chemicals
- Metals
- Other Trace Elements
- Organic Compounds
• Other Chemical Parameters
-pH
- Dissolved Oxygen
- Salinity
- Organic Matter
- Other
• Physical Parameters
Ecological Processes
• Energy Flow
- Primary Production
- Net Ecosystem Production
- Growth Efficiency
• Material Flow
- Organic Carbon Cycling
- Nitrogen and Phosphorus Cycling
- Other Nutrient Cycling
Hydrology and Geomorphology
• Surface and Groundwater flows
- Pattern of Surface flows
- Hydrodynamics
- Pattern of Groundwater flows
- Salinity Patterns
- Water Storage
• Dynamic Structural Characteristics
- Channel/Shoreline Morphology, Complexity
- Extent/Distribution of Connected Floodplain
- Aquatic Physical Habitat Complexity
• Sediment and Material Transport
- Sediment Supply/Movement
- Particle Size Distribution Patterns
- Other Material Flux
Natural Disturbance Regimes
• Frequency
• Intensity
• Extent
• Duration
Source: Science Advisory Board (2002).
Of main concern in this chapter are relationships between a certain class of diverse airborne
stressors from anthropogenic sources (i.e., PM) and one or more of the EEAs. Changes in
patterns resulting from responses of vegetation and ecosystems to the effects of fine and coarse
-------
PM deposition, along with known or possible effects on ecological processes associated with
changes in the patterns, are discussed in the subsections that follow.
The reader is also referred to several other sources for more detailed discussions of several
topics only briefly alluded to or addressed here. For example, an extensive discussion of various
types of effects of acidic deposition is presented in the U.S. National Acid Precipitation
Assessment Program (NAPAP) Biennial Report to Congress: An Integrated Assessment
Program (National Science and Technology Council, 1998). Additionally, ecological effects of
acidic precipitation and nitrate deposition on aquatic systems are discussed in the EPA document
Air Quality Criteria Document for Nitrogen Oxides (U.S. Environmental Protection Agency,
1993); and sulfate deposition and effects, as related to wetlands and aquatic habitats, are
discussed in U.S. Environmental Protection Agency (1982). Effects of lead on crops, vegetation,
and ecosystems are assessed in the EPA document Air Quality Criteria for Lead(U.S.
Environmental Protection Agency, 1986). Lastly, effects of "certain pesticides, metal
compounds, chlorinated organic compounds, and nitrogen compounds" are discussed in
Deposition of Air Pollutants to the Great Waters, Third Report to Congress (U.S. Environmental
Protection Agency, 2000a).
4.2.2 Ecosystem Exposures - Particle Deposition
Airborne particles, their precursors, and their transformation products are removed from
the atmosphere by wet and dry deposition processes. This atmospheric cleansing process
fortunately lowers the long-term buildup of lethal concentrations of these pollutants in the air
and moderates the potential for direct human health effects caused by their inhalation.
Unfortunately, these deposition processes also mediate the transfer of PM pollutants to other
environmental media where they can and do alter the structure, function, diversity, and
sustainability of complex ecosystems.
The potential effects of PM deposition on vegetation and ecosystems encompass the full
range, scales, and properties of biological organization listed under Biotic Condition, Table 4-1.
Exposure to a given mass concentration of airborne PM, however, may lead to widely differing
responses, depending on the particular mix of deposited particles. Particulate matter is not a
single pollutant, but rather a heterogeneous mixture of particles differing in size, origin, and
chemical composition. This heterogeneity exists across individual particles within samples from
4-4
-------
individual sites and, to an even greater extent, between samples from different sites. Thus far,
atmospheric PM has been defined, for regulatory purposes, mainly by size fractions and less
clearly so in terms of chemical nature, structure, or source. While size is related to the mode and
magnitude of deposition to vegetated landscapes and may be a useful surrogate for chemical
constitution, PM size classes do not necessarily have specific differential relevance for
vegetation effects (Whitby, 1978; U.S. Environmental Protection Agency, 1996a); i.e., both fine-
and coarse-mode particles may affect plants. Much of the burden of sulfates (SO42), nitrates
(NO3 ), ammonium salts (NH4+), and hydrogen ions (H+) resides in the atmosphere either
dissolved in fog water or as liquid or solid aerosols. Therefore, assessment of atmospheric PM
deposition and effects on vegetation unavoidably include discussion of NO3 and SO42 and
associated compounds involved in acidic and acidifying deposition. Other important issues
relate to trace elements and heavy metals often found in ambient airborne PM.
4.2.2.1 Fine and Coarse Particulate Matter
Fine and coarse PM have a number of contrasting properties that affect their impact on
vegetated systems (see Chapter 2, Table 2-1 of this document). The model results of Wiman and
Agren (1985) and the measurements of Lovett and Lindberg (1993) addressing the complexity of
deposition processes in patchy forested landscapes and vertical stratification within canopies
reveal clear distinctions between the deposition behavior of fine and coarse particles. For one,
coarse particles settle nearer their site of formation than do fine particles. Also, the chemical
constitution of individual particles is strongly correlated with size (i.e., most S and much N is
present on fine particles, whereas much of the base cation and heavy metal burden is present on
coarse particles) and influences the predicted landscape loading of specific elements.
Atmospheric PM may also act as a carrier for other directly phytotoxic materials (e.g.,
herbicides). Fine PM dominates the surface area of particles suspended in the atmosphere, while
coarse PM dominates the mass of such airborne particles. Surface area may become more
central to ecological impact assessment as recognition of the oxidizing capacity of fine particles,
their interactions with other pollutants such as ozone (O3), and their adsorption of phytoactive
organic compounds such as herbicides become more fully appreciated. Fine and coarse particles
respond to changes in atmospheric humidity, precipitation, and wind through different
mechanisms, differentially altering their deposition characteristics.
4-5
-------
Fine PM may exhibit similar mass concentrations at different sites (Figure 4-1) and yet be
composed of very different constituents. In eastern North America, sulfate is typically the major
component of this fraction, in contrast to the West where nitrate is a key component (Figure 4-1;
cf, the eastern Appalachian site and the western California site). On the other hand, in the urban
center of Mexico City (Hidy et al., 2000), an environment more similar to the western than
eastern U.S., concentrations of fine PM of about 300 |ig/m3 are found, and sulfate concentrations
are three times that of nitrate. In contrast to sulfur and nitrogen, the contributions of organic and
elemental carbon to the eastern and western U.S. sites were similar (Figure 4-1), although soil-
derived geologic material was greater at the more arid western site.
PM2.5(ug/m3)
Southern California
%
% Organic
% Geologic
% Carbon
PM2.5(ug/m3)
804
%
% Organic
% Geologic
% Carbon
10 20 30 40 50 60
Figure 4-1. The diversity of fine PM from sites in the western and eastern United States.
Source: Modified from Sisler and Malm (2000).
Fine PM is typically more diverse than coarse PM and is secondary in nature, having
condensed from the vapor phase or been formed by chemical reaction from gaseous precursors
4-6
-------
in the atmosphere (see Chapter 2). Fine PM derives from atmospheric gas-to-particle conversion
reactions involving nucleation, condensation, and coagulation, and from evaporation of water
from contaminated fog and cloud droplets. Sulfur and nitrogen oxides (SOX and NOX) are often
oxidized to their respective acids and neutralized with ammonium cations as particulate salts.
Fine PM may also contain condensates of volatile organic compounds, volatilized metals, and
products of incomplete combustion (see Chapter 3).
Coarse PM, in contrast, is mainly primary in nature, having been emitted from area, point,
or line sources as fully formed particles derived from abrasion and crushing processes, soil
disturbances, desiccation of marine aerosol from bursting bubbles, hygroscopic fine PM
expanding with humidity to coarse mode, and/or gas condensation directly onto preexisting
coarse particles (see Chapters 2 and 3). Suspended primary coarse PM may contain iron, silica,
aluminum, and base cations from soil, plant and insect fragments, pollen, fungal spores, bacteria,
and viruses, as well as fly ash, brake linings, debris, and automobile tire fragments. Coarse-
mode particles can be altered by chemical reactions and/or physical interactions with gaseous
or liquid contaminants.
The coating of coarse particles with semivolatile materials can substantially affect their
deposition and potential for biological effects. For example, nitrogen exhibits a strongly
bimodal size distribution: the peak above 1 jim can be attributed to nitric acid (HNO3)
adsorption onto coarse alkaline particles; and that below 1 jim can be attributed to gas phase
condensation of ammonia with either sulfuric or nitric acid yielding either (NH4)2SO4 or
NH4NO3 aerosol. HNO3 has an extremely high deposition velocity, nearly independent of the
physiography of the surface. Therefore, formation of ammonium nitrate reduces local nitrogen
deposition, because the deposition velocity of these particles is much less than that of HNO3 gas.
Similarly, anthropogenic emissions of sulfur are mostly as sulfur dioxide (SO2), which is
hydrophilic, rapidly hydrated, and subsequently oxidized to sulfate (SO42 ), which is about
30-fold less phytotoxic than SO2. The ratio of SO42 :SO2 increases with aging of the air mass
and, therefore, with distance from the source. Sulfate is thus a widespread regional/global
pollutant and is sufficiently hygroscopic that, in humid air, it exists significantly in the coarse
PM fraction. It is unusual for injurious levels of particulate sulfate to be deposited upon
vegetation in the vicinity of emission sources, while direct injury due to SO2 is commonly
observed near uncontrolled point sources. In this case, gas-to-particle conversion is of benefit
4-7
-------
to vegetation. The chemical composition of gaseous precursors of PM and the formation of
sulfates and nitrates is discussed in Section 2.1.3 of Chapter 2.
Since enactment of the 1990 Clean Air Act Amendments, the atmospheric mix of PM
precursors in the United States has changed substantially. That is, as emissions of SO2 have
declined, emissions of oxides of nitrogen (NOX) have remained about the same, but emissions of
cations have decreased (Hedin et al., 1994; Likens et al., 1996). This is almost certainly due to
increased suspension of wind-borne geologic material from exposed soils.
For characterization of tropospheric chemistry, the deposition of O3, NOX, peroxides, and
ammonia are first-order problems, followed by deposition of organics, SO2, and particulate
sulfate and nitrate (Wesely and Hicks, 2000). For impact on vegetation, however, the order may
be different and may include different species — notably SO2, fluoride (where it still exists as a
problem), particulate heavy metals, base cations, sulfate and nitrate. In spite of the current
regulatory focus on non-speciated PM, exposure to a given mass concentration of PM may lead
to widely differing phytotoxic outcomes depending upon the particular mix of PM constituents
involved. This variability has not been characterized adequately. Though effects of specific
chemical fractions of PM have been described, there has been relatively little research aimed at
defining the effects of unspeciated PM on plants or ecosystems.
4.2.2.2 Diversity of Deposition Modes
Atmospheric deposition of particles to ecosystems takes place via both wet and dry
processes, through three major routes: (1) wet, by precipitation scavenging in which particles
are deposited in rain and snow; (2) much slower dry deposition; and (3) occult deposition
(so named because it was hidden from measurements which determined the previous two types
of deposition) by fog, cloud-water, and mist interception (Table 4-2). Unlike gaseous dry
deposition, neither the solubility of the particle nor the physiographical activity of the surface are
likely to be of first order importance in determining particulate dry deposition velocity (Vd).
Factors that contribute to surface wetness or stickiness may be critical determinants of deposition
efficiency. Available tabulations of deposition velocity are highly variable and suspect. Recent
evidence indicates that all three modes of deposition (wet, occult, and dry) must be considered in
determining inputs to watersheds or ecosystems, because each may dominate over specific
4-8
-------
TABLE 4-2. TYPES AND DETERMINANTS OF PARTICIPATE DEPOSITION
AND IMPACT TO VEGETATION
Type of Deposition Determinant of Deposition
Quantifiable Factors
Dry deposition
Wet deposition
Ambient Concentration
Atmospheric Conditions
Aerosol Properties
Surface Roughness
Occult deposition
Vegetation Condition
Ambient Concentration
Atmospheric Conditions
Aerosol Properties
Surface Roughness
As Above
Distance from Source
Emission Strength
Wind Speed
Stability
Mixing Height
Temperature
Humidity
Dew Formation
Chemical Reactivity
Particle Solubility
Aerodynamic Diameter
Biological Availability
Hygroscopicity
Terrain Discontinuity
Leaf Pubescence
Leaf Shape
Plant Density
Branch Spacing
Tissue Flexibility
Surface Wetness
Salt Exudates
Organic Exudates
Insect Excreta
Distance from Source
Emission Strength
Mixing Height
Timing of Precipitation
Intensity of Precipitation
Duration of Precipitation
Chemical Reactivity
Particle Solubility
Biological Availability
Terrain Discontinuity
Leaf Pubescence
Leaf Area Index
Nature of Exposed Bark and Stem
Combination of Above Factors
4-9
-------
intervals of time or space and ultimately, by interception and impaction, onto vegetation or other
rough elements in the landscape.
The distribution of deposition between wet, dry, and occult modes varies substantially
between locations for both nitrogen and sulfur (Figure 4-2). Clearly, rainfall and snowfall will
determine the magnitude of wet deposition. Precipitation events clean the air so that dry
deposition is eliminated or reduced during subsequent dry periods. Occult deposition depends
upon landscape interception of the cloud base (Cape, 1993). This may occur at high elevation
sites, in coastal areas subject to onshore advection, or in low-lying interior areas subject to
radiation fogs. Thus, ecosystem exposure is determined by the mode, and to some extent the
magnitude, of deposition. Total deposition among mountain sites is related to the magnitude of
the occult deposition, particularly for nitrogen (Figure 4-2). Topography and vegetation
characteristics influence the deposition modes differently (Table 4-3). In general, dry deposition
is the most sensitive to features of the vegetated surface, and wet deposition is the least sensitive.
Comparison of micrometeorological and other methods for estimating particle deposition
velocity (Erisman et al., 1997) suggests that there is little discrepancy between contrasting
methodologies for estimating particle deposition and that this conclusion holds for both anions
and cations, with the exception of nitrogenous species, which appear to interact with foliage in
more complex ways. These comprehensive studies in the Speulder forest in the Netherlands
indicated that aerosol deposition represents a considerable fraction of total deposition to the
landscape. At this location, occult deposition was relatively insignificant, but dry deposition
accounted for about one-fourth of alkaline-earth cation deposition.
Wet Deposition
Wet deposition results from the incorporation of atmospheric particles and gases into cloud
droplets and their subsequent precipitation as rain or snow, or from the scavenging of particles
and gases by raindrops or snowflakes as they fall (Lovett, 1994). Precipitation scavenging, in
which particles are incorporated in hydrometers and deposited in the resulting rain and snow,
includes rainout (within-cloud incorporation by nucleation) and washout (below-cloud
scavenging by impaction). Wet deposition generally is confounded by fewer factors than dry or
occult deposition and has been easier to quantify. Total inputs from wet deposition to vegetative
canopies can be significant (Table 4-3), although not all wet deposition involves particle
4-10
-------
80 -
60-
40-
20-
0
140
120
100
80-
60-
40-
20-
0
Keilder
Mitchell
Whitetop
Total ' Occult ' Dry
% Deposition
Wet
Figure 4-2. Relative importance of three modes of deposition of nitrate (A) and sulfate (B)
at high elevation sites (Unsworth and Wilshaw, 1989; Fowler et al., 1989;
Mueller, 1988; Aneja and Murthy, 1994).
4-11
-------
TABLE 4-3. 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/ha)
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/ha)
17.20
7.60
7.79
a 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 was derived as in Hicks et al. (1987).
b Wet nitrogen consists of NO3 and NH4+; dry nitrogen consists of vapor phase HNO3 and NO2; and paniculate
nitrogen consists of NO3 .
c Wet sulfur consists of SO42~, dry sulfur consists of vapor phase SO2, and paniculate sulfur consists of pSO4~2
scavenging because gaseous pollutants also dissolve in raindrops during precipitation events
(Lovett, 1994). This contribution is obscured during measurements, because wet deposition is
measured simply by chemical analysis of total precipitation collected in clean, nonreactive
buckets. Exclusion of dry deposited material (as opposed to dissolved gaseous species) requires
closure or covering of the vessels except during precipitation.
Wet deposition is largely a function of precipitation amount and ambient pollutant
concentrations. Surface properties (Table 4-2) have little effect on wet deposition, compared to
dry or occult deposition, although leaves (depending on their surface properties of wettability,
exposure, and roughness) retain liquid and solubilized PM. Extensive vegetative canopies
typically develop leaf area indices (LAI; ratio of projected leaf area to ground area) much greater
than 1. Thus, any material deposited via precipitation to the upper stratum of foliage is likely to
be intercepted by several foliar surfaces before reaching the soil.
Landscape characteristics may also affect wet deposition. Forested hillsides often receive
four- to six-fold greater inputs of wet deposition than short vegetation in nearby valleys. This
effect is due to a variety of orographic effects (Unsworth and Wilshaw, 1989) and the closer
aerodynamic coupling to the atmosphere of tall forest canopies than of the shorter canopies in
the valleys. This leads to more rapid foliar drying in forest camopies, which reduces the
4-12
-------
residence time but concentrates the solubilized participate materials available for foliar uptake
on the cuticular surface more quickly; and concentration increases the thermodynamic driving
force for foliar uptake (Fowler et al., 1991; Unsworth, 1984; Schonherr and Huber, 1977).
Following wet deposition, humidity and temperature conditions strongly influence the extent of
biological effects, because of the competing effects of drying versus concentrating the solutions,
which influence the rate of metabolic uptake of surface solutes (Swietlik and Faust, 1984). The
net consequence of these factors on direct physical effects of wet deposited PM on leaves is not
known.
Rainfall introduces new wet deposition and redistributes throughout the canopy previously
dry-deposited PM, particularly coarse particles that were preferentially deposited in the upper
foliage (Peters and Eiden, 1992). Both effects scale the likelihood of foliar contact and potential
direct PM effects on vegetation nearly linearly with canopy leaf area. The concentrations of
suspended and dissolved materials are typically highest at the onset and decline with duration of
individual precipitation events (Lindberg and McLaughlin, 1986; Hansen et al., 1994).
Sustained rainfall removes much of the accumulation of dry-deposited PM from foliar surfaces,
reducing direct foliar effects and combining the associated chemical burden with the wet-
deposited material (Lovett and Lindberg, 1984; Lovett, 1994) for transfer to the soil. Intense
rainfall may contribute substantial total particulate inputs to vegetated land surfaces, mostly via
the soil, but is less effective as a source of directly bioavailable or injurious pollutants to foliar
surfaces. This washing effect, combined with differential foliar uptake and foliar leaching of
different chemical constituents of PM, alters the composition of the rainwater that reaches the
soil. Low intensity precipitation events, in contrast, may deposit significantly more particulate
pollutants to foliar-surfaces than high intensity precipitation events. Because of the short
duration and limited atmospheric cleansing, the concentration of PM in the final precipitation
that remains in contact with foliar surfaces may be high. Additionally, low-intensity events may
enhance foliar uptake through the hydrating of some previously dry-deposited particles. The
combination of dry deposition to foliage and subsequent wet removal increases the potential for
PM to exert effects via soil pathways: first by enhancing dry deposition relative to adjacent
unvegetated surfaces; and second by accelerating passage along with wet deposited PM material
by throughfall and stemflow to the soil. Once, in the soil, PM may affect the important
ecosystem-level biogeochemical cycles of major, minor, and trace elements.
4-13
-------
Dry Deposition
Dry particulate deposition is a complex, poorly characterized process controlled primarily
by atmospheric stability, macro- and micro-surface roughness, particle diameter, and surface
characteristics (Table 4-2; Hosker and Lindberg, 1982). The range of particle sizes, the diversity
of canopy surfaces, and the variety of chemical constituents in airborne PM have slowed
progress in both prediction and measurement of dry particulate deposition. Deposition of
particles suspended regionally and throughout the full depth of the planetary boundary layer
(PEL) is controlled by different mechanisms within the three distinct atmospheric transport
zones above the surface. In the lower atmosphere, fine particles are transported by turbulent
eddies of mechanical and convective origin. In the relatively unstirred, laminar boundary layer
surrounding individual surface elements, Brownian diffusion dominates. Near the surface,
actual deposition and contact with the surface is mediated by impaction (El-Shobokshy, 1985).
Dry deposition of atmospheric particles to plant and soil is a much slower process than wet
or occult deposition, but it occurs nearly continuously and affects all exposed surfaces (Hicks,
1986). In dry deposition, particles at the large end of the spectrum (i.e., > 5 jim diameter) are
deposited mainly by gravitational sedimentation and inertial impaction. Smaller particles,
especially those with diameters between «0.2 and 2 jim, are not readily dry-deposited and tend
to travel long distances in the atmosphere until their eventual deposition, most often via
precipitation. This long-distance transport of fine aerosols is largely responsible for the regional
nature of acid deposition (Lovett, 1994). A major conclusion from atmospheric deposition
research is the realization that dry deposition is usually a significant and, in some cases, a
dominant portion of total atmospheric deposition to an ecosystem (Lovett, 1994). Plant parts of
all types, including those not currently physiologically active, along with exposed soil and water
surfaces, receive steady deposits of dry dusts, elemental carbon encrustations, grease films, tarry
acidic coatings, and heterogeneous secondary particles formed from gaseous precursors (U.S.
Environmental Protection Agency, 1982).
Deposition fluxes may be calculated from measurements, estimates, or modeled values of
mass concentration, C, at a specified measurement height and the total conductance or deposition
velocity, Vd, from this height to the surface (equation 4-1; Hicks et al., 1987). These modeling
techniques are closely allied with the micrometeorological techniques used to measure such
fluxes. The flux, F, may be inferred as:
4-14
-------
F=Vd*(Cz-C0), (4-1)
where F is flux to the surface, Cz is the particle concentration at measurement height z, C0 is the
particle concentration at receptor sites in the canopy (usually assumed equal to 0), and Vd is the
overall deposition velocity. The flux is controlled by Vd and Cz.
Vertical transport of particles through the lower atmosphere to the vicinity of the
vegetation elements is by turbulence and sedimentation, such that:
Vd = Vt + Vs, (4-2)
in which V, (inner, left-hand pathway of Figure 4-3) is a turbulent diffusion term, and Vs is a
sedimentation term that dominates deposition of very coarse particles (Figure 4-4) and increases
with particle size (Figure 4-5; dotted line). Sedimentation may be considered a pathway parallel
to turbulent transport (Figure 4-4), but this is an oversimplification. Vs affects the concentration
of particles near the surface where eddy transport may occur and also governs the redeposition
of some fraction of the particles that were lost to resuspension or rebound following deposition
by impaction. For this reason, Vs is included (Figure 4-3) in the composite surface resistance
term (RaRcpVs) as well as in the parallel sedimentation term.
For submicron particles for which sedimentation is negligible (Hicks et al., 1987; Monteith
and Unsworth, 1990; Wesely, 1989), the resistance catena Ohm's Law Analogy analogous to the
law used to describe transport of heat, momentum, or gases may be adequate, as:
Vd =Vt=[ra + rb+rc]-', (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 via
stomata in the leaf surface. Significant departures from the analogy arise near the surface
4-15
-------
V
1
Particles
Atmospheric
Source
Particles
CP
v
Figure 4-3. A simplified resistance catena representing the factors controlling deposition
of particles to the surface. Vegetation-specific parameters are not explicitly
considered.
Source: Modified after Hicks et al. (1987).
(Chamberlain, 1975; Sehmel, 1980), where particles that were transported efficiently by
turbulent eddies are slowed substantially in the laminar boundary layer, reducing the efficiency
of impaction. The preservation of momentum in this zone declines with decreasing diameter;
however, this is offset by an increase in Brownian diffusivity with decreasing diameter
(Figure 4-3). Aerodynamic streamlines are parallel to the surface of each roughness element,
so that deposition ultimately depends on diffusion to the surface. The transition from impaction
to diffusion is likely blurred in the presence of leaf pubescence extending beyond the boundary
layer. These conflicting trends lead to a broad range over which empirical measurements of Vd
and particle size are relatively independent (Figure 4-3), further demonstrating the importance of
the quasilaminar boundary layer (Lamaud et al., 1994; Shinn, 1978).
4-16
-------
8
7-
'J/T
£ 6^
£ 5-
o
£
04-
.1 3^
0
Q
1 -
0
Mn
Cu
Cl
1234
Coarse/Fine Ratio (M9/M9)
Figure 4-4. The relationship between deposition velocity of selected particulate materials
and the distribution of the material between the coarse- and fine-aerosol
fractions.
Source: Data from Foltescu et al. (1994). Ranges for Mn and Fe are from Davidson and Wu (1989).
The aerodynamic term (ra) decreases with increasing wind speed, turbulence, and friction
velocity and increases with measurement height and atmospheric stability. It describes the
capacity of turbulent eddies to transport material, momentum, and heat between the
measurement height and the roughness height of the surface. Coarse particles may not be carried
efficiently by the high frequency eddies near the surface and may fall more rapidly than they
diffuse by either Brownian or turbulent processes. Thus, the relevance of ra breaks down as Vs
increases. Indeed because Vs (equation 4-2) is independent of a concentration gradient, the
electrical analogy is a theoretically flawed approximate approach (Venkatram and Pleim, 1999).
4-17
-------
(0
E
u
o
$
c
o
I?
'55
o
Q.
0)
Q
1,000-
100 —
10 —
1 —
0.1 —
0.01 —
0.001 —
0.000
i Stokes Law
Brownian Diffusion
' Peters and Eider (1992)
• Little and Wiffen (1977)
0.001 0.01 0.1 1 10
Particle Diameter (pm)
100
Figure 4-5. 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 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 um) are influenced strongly by both
particle size and sedimentation velocity, and deposition is independent of size.
Deposition Velocity
Because the final stage of deposition for particles involves either impaction following
deceleration through a quasilaminar boundary layer or diffusion through this boundary layer, its
effective depth is a critical determinant of Vd (Wiman et al., 1985; Peters and Eiden, 1992). The
term corresponding to the boundary layer resistance for gases (rb; equation 4-3) incorporates the
absence of form drag for gases. This parameter decreases with increasing turbulence and
particle diffusivity; it is poorly characterized for gases (depending critically on canopy
morphology, vertical wind profiles, and gust penetration) and is of extremely limited usefulness
for particles.
4-18
-------
Once delivered by turbulent transport or sedimentation to the vicinity of vegetative surface
elements, a variety of particle size-dependent mechanisms come into play, some differing
substantially from those governing gaseous deposition. The concepts of rb (the still air or
boundary layer resistance) and rc (the canopy or surface resistance) are not generally applicable
to deposition of poly disperse particles. Because of the roles of momentum and bounce-off and
the complication by reentrainment back into the airstream following deposition of a particle to
the surface, the factors determining the effective rb and rc for particle deposition are not as
independent as for gases. They are replaced in some resistance formulations (e.g., Hicks et al.,
1987) by the term, rcp, that combines near-surface and surface effects and by a mathematically
derived composite term, RaRcpVs, that combines atmospheric, surface, and sedimentation effects
(Figure 4-4). This latter term was insignificant for the submicron sulfate component considered
originally in its derivation (Hicks et al., 1987); however, it scales with the square of particle
diameter, so that its general applicability to polydisperse particles is unclear. In general,
transport between the turbulent air column and the leaf surface through the laminar boundary
layer remains difficult to describe (Lindberg and McLaughlin, 1986).
Current estimates of regional particulate dry deposition (e.g., Edgerton et al., 1992; Brook
et al., 1999) infer fluxes from the product of variable and uncertain measured or modeled
particulate concentrations and even more variable and uncertain measured or modeled estimates
of dry deposition velocity parameterized for a variety of specific surfaces (e.g., Brook et al.,
1999). However, even for specific sites and well defined particles, uncertainties in F are largest
in the values of Vd, which are typically characterized by the large ranges and variances described
in other sources (e.g., Bytnerowicz et al., 1987a,b and Hanson and Lindberg, 1991, for nitrogen-
containing particles; McMahon and Denison, 1979 and Hicks et al., 1987, for general treatment).
The nature of the vegetative cover to which particulate deposition occurs has a moderate to
substantial effect on the components of Vd. The surface resistance (Hicks et al., 1987) is a
significant and highly site-specific component of total resistance that is difficult to predict along
with site, seasonal, and diurnal effects on the atmospheric components of total resistance.
Early models of dry particulate deposition to vegetation (e.g., U.S. Environmental
Protection Agency, 1982; Chamberlain, 1975; Davidson and Friedlander, 1978; Garland, 1978;
Little and Wiffen, 1977; McMahon and Denison, 1979; Sehmel, 1980; Sehmel and Hodgson,
1976; Slinn, 1977, 1978) used this paradigm (e.g., equation 4-3) to deal with transport to the
4-19
-------
near-surface regime explicitly including conventional micrometeorological and particle size
considerations. Alternative modeling treatments have attempted to parameterize the geometry of
vegetative receptor surfaces and within-canopy micrometeorology (Wiman and Agren, 1985;
Peters and Eiden, 1992). Chemical reactivity, particle shape and density, rates of physiological
sequestration, and reentrainment by gusts of wind remain to be addressed. Modeling the
deposition of particles to vegetation is at a relatively early stage of development, and it is not
currently possible to identify a best or most generally applicable modeling approach. These
approaches have been further elaborated with canopy-specific choices among the available
models and with specific incorporation of capture efficiencies by Brook et al. (1999).
Methods of Measuring Dry Deposition
Methods of measuring dry deposition of particles are more restricted than for gaseous
species and fall into two major categories: surface analysis methods, which include all types of
measurements that examine contaminant accumulation or surfaces of interest, and atmospheric
flux methods, which measure contaminants in the atmosphere from which one may estimate the
flux (Davidson and Wu, 1990). Surface extraction or washing methods characterize the
accumulation of particles on natural receptor surfaces of interest or on experimental surrogate
surfaces. These techniques rely on methods designed specifically to remove only
surface-deposited material (Lindberg and Lovett, 1985). Total surface rinsate may be equated to
accumulated deposition or to the difference in concentrations in rinsate between exposed and
control (sheltered) surfaces and may be used to refine estimates of deposition (John et al., 1985;
Dasch, 1987). In either case, foliar extraction techniques may underestimate deposition to leaves
because of uptake and translocation processes that remove pollutants from the leaf surface
(Taylor et al., 1988; Garten and Hanson, 1990). Foliar extraction methods also cannot
distinguish sources of chemicals (e.g., N) deposited as gases from those deposited as particles
(e.g., HNO3 or NO3 from nitrogen dioxide [NO2], or NH3 from NH4+ [Bytnerowicz et al.,
1987a,b; Dasch, 1987; Lindberg and Lovett, 1985; Van Aalst, 1982]). Despite these limitations,
these methods are often used in the development of in-canopy deposition models (McCartney
and Aylor, 1987).
Deposition of pollutants by wet deposition is relatively straightforward to determine
through analysis of precipitation samples. Dry deposition of pollutants, on the other hand, is
4-20
-------
more difficult to measure. A significant limitation on current capacity to estimate regional
impacts of PM is inadequate knowledge of the mechanism and factors governing particle dry
deposition to diverse surfaces. This has hindered efforts to develop robust measurement
techniques for particle deposition (distinct from atmospheric concentration) and has
compromised efforts to develop generally applicable deposition models for particles. The
National Dry Deposition Network (NDDN) was established in 1986 to document the magnitude,
spatial variability, and trends in dry deposition across the United States. Currently, the network
operates as a component of the Clean Air Status and Trends Network (CASTNet) (Clarke et al.,
1997).
Dry deposition is not measured directly, but is determined by a inferential approach (i.e.,
fluxes are calculated as the product of measured ambient concentration and a modeled deposition
velocity). This method is appealing and widely used because atmospheric concentrations are
relatively easy to measure when compared to dry deposition fluxes, and models have been
developed to calculate deposition velocities (Lovett, 1994). Ambient pollutant concentrations,
meteorological conditions, and land use data required for the inferential model are routinely
collected at CASTNet dry deposition sites. Monitored chemical species include ozone, sulfate,
nitrate, ammonium, sulfur dioxide, and nitric acid. The temporal resolution for the ambient
concentration measurements and dry deposition flux calculations is hourly for ozone and weekly
for the other chemical substances (Clarke et al., 1997). Isotopic labeling of dry deposited PM
(e.g., sulfate with 35S) prior to experimental surface exposures and extractions (Garten et al.,
1988) can provide more precise differentiation between the deposition rates of related chemical
species (e.g., SO42 from SO2).
At the whole-canopy level, natural surface washing by rainfall may be used to estimate dry
deposition of PM and gases during the preceding dry period (Cape et al., 1992; Davidson and
Wu, 1990; Draaijers and Erisman, 1993; Erisman, 1993; Fahey et al., 1988; Lindberg and Lovett,
1992; Lovett and Lindberg, 1993; Reiners and Olson, 1984; Sievering, 1987). Collection and
analysis of stem flow and throughfall provides useful estimates of particulate deposition when
compared to directly sampled precipitation. The method is most precise for strictly PM
deposition when gaseous deposition is a small component of the total dry deposition and when
leaching or uptake of compounds of interest out of or into the foliage (i.e., N, S, base cations) is
not a significant fraction of the total deposit! onal flux (Davidson and Wu, 1990; Draaijers and
4-21
-------
Erisman, 1993; Lindberg and Lovett, 1992; Lovett and Lindberg, 1993). Throughfall sampling
of sulfate deposition (Garten et al., 1988; Lindberg and Garten, 1988; Lindberg et al., 1990)
often suggests substantial foliar exchange. Other throughfall studies (e.g., Erisman, 1993; Fahey
et al., 1988) may lack sufficient specificity for evaluating dry particle deposition.
Careful timing of throughfall and stemflow measurements after prolonged dry periods,
with simultaneous direct measurement of unchanged precipitation, can be used to determine the
magnitude of dry deposition between precipitation events. Indeed, this foliar washing technique,
whether using subsequent precipitation or experimental lavage, is one of the best available
methods to determine dry deposition of PM to vegetated ecosystems. Major limitations include
the extreme site specificity of the measurements, the substantial labor requirement that normally
precludes regional coverage, and the restriction to elements that are conserved within the
vegetative system (thereby excluding semivolatile organics, ammonium salts, and gases such as
ozone).
Surrogate surfaces have not been found that can adequately replicate essential features of
natural surfaces; and therefore the surfaces currently used do not produce reliable estimates of
particle deposition to the landscape. Deposition to surrogate surfaces deployed in extensive
plant canopies may provide a measure of particle deposition to the surrounding foliage or soil
surfaces. For example, a uniform population of 0.8 jim gold colloid particles deposited with
similar coverage to leaves ofPhaseolus vulgaris and to upward-facing inert surfaces (Klepper
and Craig, 1975). However, dry deposition of particles to foliage and to flat inert surrogate
surfaces (polycarbonate Petri dishes) in a deciduous forest resulted in greater accumulation on
the inert surfaces, although both surfaces accumulated particles of a similar range of sizes
(Lindberg and Lovett, 1985). These differences in deposition/accumulation remain to be fully
characterized and hinder the surrogate techniques in providing quantitative deposition estimates.
Micrometeorological methods employ an eddy covariance, eddy accumulation, or flux
gradient protocol in contrast to washing or extracting of receptor surfaces for quantifying dry
deposition. These techniques require measurements of PM concentrations and of atmospheric
transport processes. They are currently well developed for ideal conditions of flat,
homogeneous, and extensive landscapes and for chemical species for which accurate and rapid
sensors are available. Additional studies have expanded the variety of such species and extended
4-22
-------
these techniques to more complex terrain (McMillen, 1988; Hicks et al., 1984; Wesely and
Hicks, 1977).
The eddy covariance technique measures vertical fluxes of gases and fine particles directly
from calculations of the mean covariance between the vertical component of wind velocity and
pollutant concentration (Wesely et al., 1982). This technique is particularly limited by its
requirement for sensors capable of acquiring concentration data at 5 to 20 Hz. For the flux-
gradient or profile techniques, vertical fluxes are calculated from a concentration difference and
an eddy exchange coefficient determined at discrete heights (Erisman et al., 1988; Huebert et al.,
1988). Businger (1986), Baldocchi (1988), and Wesely and Hicks (1977) evaluated the benefits
and pitfalls of these micrometeorological flux measurements for gases. Most measurements of
eddy transport of PM have used chemical sensors (rather than mass or particle counting) to focus
on specific PM components. These techniques have not been well developed for generalized
particles and, for the same reasons that limit mathematical description of particle deposition,
may be less suitable for coarse particles that are transported efficiently in high frequency eddies
(Gallagher et al., 1988).
Factors Affecting Dry Deposition
Ambient Concentration. The ambient concentration of particles (Cz; equation 4-1), the
parameter for which there is the most data (for example, see Chapter 3, this document), is at best
an indicator of exposure. However, it is the amount of PM actually entering the immediate plant
environment that determines the biological effect. The linkage between ambient concentration
and delivery to vegetation is the deposition velocity (Vd; equation 4-1). Cz is determined by
regional and local emission sources, regional circulation, and weather. It may be locally
sensitive to removal from the atmosphere by deposition, but this effect is generally small.
Average annual concentrations for NO3 exhibit much more variability than SO4~2and a definite
pattern of higher concentrations in the Midwest than elsewhere. The highest concentrations are
observed (i.e, > 2 jig nT3) in the agricultural areas of the Midwest, while the lowest are seen at
forested sites in New England and the southern Appalachian Mountains. Annual average
concentrations SO42 of 5.0 jig nT3 are observed over nearly the entire eastern United States from
New York and Michigan to northern Mississippi and Alabama (Edgerton et al., 1992).
4-23
-------
Deposition increases linearly with the concentration of many materials over a broad range.
This allows atmospheric cleansing to take place and accounts for the greater surface impact of
pollutant particles during pollution episodes. A serious limitation of the Vd formulation used to
infer deposition of specific chemical species that exist in a range of particle sizes is an
appropriate specification of their concentration. Most sulfur emissions are readily oxidized to
sulfite, bisulfite, and SO42 . In the presence of atmospheric ammonia, particulate ammonium
sulfate is formed. However, this material is hygroscopic and will increase in mass and diameter
in the presence of high humidity, altering its deposition behavior. Similarly, coalescence of
small particles into larger aggregates and adsorption of gaseous pollutants onto existing coarse
particles complicate the association of particle size with concentration of individual chemical
species.
Distance from the source, and the resulting residence time in the atmosphere, control the
relative concentrations of surface-reactive materials (NO, SO2) of secondary particles that take
some time to form in the atmosphere and of coarse particles that exhibit high rates of deposition
by sedimentation near the source. These interacting processes affect the time required for the
formation of secondary particles by gas-to-particle conversion reactions, resulting in a greater
ratio of dry to wet deposition near emission sources where gaseous sulfur dioxide (gSO2)
deposition predominates than at greater distances where rainout of particulate SO42 (pSO42~)
may dominate (Barrie et al., 1984) and where dry deposition of pSO42~ may be greater than that
of gSO2. The effect of gas-to-particle conversion on dry deposition of a specific chemical
species can be substantial because Vd for SO2 is approximately 0.33 ± 0.17 cm/s; whereas it is
approximately 0.16 ± 0.08 cm/s for SO42 . These phase conversions affect both Cz and the
effective Vd which together control dry depositional fluxes (equation 4-1). The neutralization of
acidic gaseous and particulate species by alkaline coarse particles has been described in arid
regions, but it may be more prevalent in urban New York where coarse particles are observed to
be neutral because alkaline cations approximately balance gaseous acidic species (Lovett et al.,
2000). The deposition of the acidic materials in the urban environment is likely enhanced by
incorporation into these previously formed coarse particles.
Similarly, the ratio of coarse to fine particle concentrations determines the effective Vd for
chemically speciated particles (Figure 4-4). This ratio reflects the size-dependent deposition
processes that govern delivery of PM to receptor surfaces (Figure 4-5). For example, SO42 was
4-24
-------
found predominantly on fine submicron particles; whereas potassium (K+), calcium (Ca2+), and
nitrate (NO3 ) ions were associated most often with coarse particles larger than 2 jim (Lindberg
and Lovett, 1985). However, concentrations of particulate S and K+ within a coniferous canopy
were strongly correlated (Wiman and Lannefors, 1985), suggesting a primary source of
coarse-mode sulfur particles. Many researchers reported biomodal mass-size distributions of
NO3 with one peak in the fine mode and the other peak in the coarse mode (Wu and Okada,
1994). The behavior of NO3 m ambient aerosols depends not only on the concentrations of
gaseous nitric acid and ammonia, but also on the chemical composition of the particles and
atmospheric conditions (Tang, 1980). When sea-salt particles were transported atmospherically
from the Pacific Ocean in the Nagoya, Japan area, the amount of nitrate in the coarse particle
size range increased. Coarse particle formation on sea salt under these conditions becomes a
major pathway for nitrate. The heterogeneous reaction of NaCl with gaseous HNO3 is
considered to be the dominating pathway for nitrate formation (Wu and Okada, 1994). As a
result, marine and continental particle size spectra for both N and S differ substantially: a peak
in the coarse mode is generally apparent near marine sources (Milford and Davidson, 1987).
The issue for NO3 is further confounded by uncertain discrimination between gaseous and
parti culate N species by current sampling methods. The substantial effect of particle size on Vd
(Figure 4-5) implies a need for size resolution as well as chemically speciated ambient PM
concentrations, even within the PM10 fraction.
Particle Effects on Vd. Particle size is a key determinant of Vd as noted above; but,
unfortunately, the size spectra may be quite complex. The particles in the study of Lindberg and
Lovett (1985) at Walker Branch Watershed had median diameters ranging from 3 to 5 jim; but
approximately 25% of the particles had diameters < 1 |im (0.2 to 0.3 jim), and 5 to 20% of the
particles were much larger aggregates. The aggregated particles are significant in that
chemically they reflect their fine particle origins, but physically they behave like large particles
and deposit by sedimentation. Direct observations with SEM demonstrate that particle
morphology can be highly variable. Many submicron particles can be observed on trichomes
(leaf hairs), although most particles are in the 5 to 50 jim diameter range. Large aggregated
particles in excess of 100 jim also are seen; larger carbonaceous aggregate particles are
4-25
-------
especially common (Smith, 1990a). Trichomes are efficient particle receptors; however, they are
reduced in size by "weathering" and occasionally break off during the growing season.
In the size range around approximately 0.1 to 1.0 jim, where Vd is relatively independent of
particle diameter (Figure 4-5), deposition is controlled by macroscopic roughness properties of
the surface and by the stability and turbulence of the atmospheric surface layer. The resistance
catena (Figure 4-3) is less useful in this size range and, in some treatments, has been abandoned
entirely (e.g., Erisman et al., 1994; equation 4-4). Impaction and interception dominate over
diffusion, and the Vd is considerably (up to two orders of magnitude; Figure 4-2) lower than for
particles either smaller or larger (Shinn, 1978). The deposition velocity may be parameterized in
this size range as a function of friction velocity,
Vd - (a/b)u*, (4-4)
where a depends on atmospheric stability and b depends on surface roughness (Wesely et al.,
1985; Erisman et al., 1994). Similar formulations have been presented in terms of turbulence
(standard deviation of wind direction) and wind speed (e.g., Wesely et al., 1983), both
determinants of u*.
Deposition of particles between 1 and 10 jim in diameter, including the coarse mode
of PM10, is strongly dependent on particle size (Shinn, 1978). Larger particles within this size
range are collected more efficiently at typical wind speeds than are smaller particles (Clough,
1975), suggesting the importance of impaction. Impaction is related to wind speed, the square of
particle diameter, and the inverse of receptor diameter as a depositing particle fails to follow the
streamlines of the air in which it is suspended around the receptor. When particle trajectory
favors a collision, increasing either wind speed or the ratio of particle size to receptor cross
section makes collision nearly certain. Similarly, as these parameters become very small, the
probability of collision becomes negligible. However, the shape parameter for the more
common range of situations between these extremes remains poorly characterized (Peters and
Eiden, 1992; Wiman and Agren, 1985).
As particle size increases above 1 |im, deposition is governed increasingly by
sedimentation (Figure 4-5) and decreasingly by turbulence and impaction. Particles between
-10 and 24 jim (Gallagher et al., 1988) are both small enough to be transported efficiently by
4-26
-------
turbulent eddies to the surface and large enough to impact with sufficient momentum to
overcome boundary layer effects. These particles deposit highly efficiently and relatively
independently of particle size. Deposition of the largest suspended particles (e.g., > 50 jim) is
governed, through sedimentation and the corresponding terminal settling velocity, Vs, almost
entirely by size. These particles are not transported efficiently by small-scale eddies near the
surface.
Theoretically based models for predicting particle deposition velocities have been
published by Bache (1979a,b), Davidson et al. (1982), Noll and Fang (1989), Slinn (1982), and
Wiman (1985). These models deal primarily with low canopies or individual elements of
canopy surfaces. Wiman and Agren (1985) developed an aerosol deposition model that
specifically treats the problem of particle deposition to forests where turbulence plays a
particularly important role, especially at roughness transitions such as forest edges. They found
that deposition of supermicron particles is controlled by complex interactions among particle
size and concentration, forest structure, and aerodynamics; whereas deposition of fine
(submicron) particles is controlled by particle concentration and forest structure.
Empirical measurements of Vd for fine particles under wind tunnel and field conditions are
often several-fold greater than predicted by available theory (Unsworth and Wilshaw, 1989).
A large number of transport phenomena, including streamlining of foliar obstacles, turbulence
structure near surfaces, and various phoretic transport mechanisms remain poorly parameterized
in current models. The discrepancy between measured and predicted values of Vd may reflect
such model limitations or experimental limitations in the specification of the effective size and
number of receptor obstacles, as suggested by Slinn (1982).
Available reviews (Davidson and Wu, 1990; McMahon and Denison, 1979; Nicholson,
1988; Sehmel, 1980; Slinn, 1982; U.S. Environmental Protection Agency, 1982, 1996a) suggest
the following generalizations: (1) particles > 10 jim exhibit a variable Vd between -0.5 and
1.1 cm/s depending on friction velocities; whereas a minimum particle Vd of 0.03 cm/s exists for
particles in the size range 0.1 to 1.0 |im; (2) the Vd of particles is approximately a linear function
of friction velocity; and (3) deposition of particles from the atmosphere to a forest canopy is
from 2 to 16 times greater than deposition in adjacent open terrain (i.e., grasslands or other
vegetation of low stature).
4-27
-------
Leaf Surface Effects on Vd. The term rc (Equation 4-3) reflects the chemical, physical, or
physiographical characteristics of the surface that govern its ability to capture, denature, or
otherwise remove particulate material from the atmospheric surface layer. For gases, relevant
surface properties involve the physiological state of the vegetation, including stomatal opening
and mesophyll antioxidant activity, as well as the chemical reactivity of the exposed surface with
the specific gas. For particles, relevant surface properties involve stickiness, microscale
roughness, and cross-sectional area. These properties determine the probability of impaction and
bounce (e.g., Shinn, 1978). The chemical composition of PM is not usually a primary
determinant of deposition velocity. At the microscopic scale where Van der Waals forces may
determine particle bounce and reentrainment, the chemical properties of both surface and particle
may be significant but remain poorly understood.
Stickiness may itself depend on previous deposition of deliquescent particles that prolong
leaf wetness, on the wettability of foliar surfaces, and on the presence of sticky residues such as
honeydew deposited by aphids. These factors increase deposition by decreasing bounce-off,
wind reentrainment, and, to some extent, wash-off by precipitation.
The distribution of particles on, and the efficiency of deposition to, vegetation also varies
with leaf shape and plant part. Particles are more prevalent on the adaxial (upper surface facing
twig) surface than on the abaxial (lower surface away from the twig). Peripheral leaf areas tend
to be the cleanest, with most particles accumulating in the midvein, central portion of leaves.
The rough area surrounding the stomatal pores was not found to be a preferential site for particle
deposition or retention (Smith and Staskawicz, 1977). Most particles were located near veins
with smaller particles localized on the trichomes. The greatest particle loading on
dicotyledonous leaves is frequently on the adaxial (upper) surface at the base of the blade, just
above the petiole junction. Precipitation washing probably plays an important role in this
distribution pattern. Lead particles accumulated to a greater extent on older than younger
needles and twigs of white pine, indicating that wind and rain were insufficient to fully wash the
foliage. Fungal mycelia (derived from windborne spores) were frequently observed in intimate
contact with other particles on other leaves, which may reflect minimal reentrainment of the
spore due to shelter by the particles, mycelia development near sources of soluble nutrients
provided by the particles, or simply codeposition (Smith and Staskawicz, 1977). This pattern is
significant and could yield further insight into deposition mechanisms.
4-28
-------
Leaves with complex shapes collect more particles than those with regular shapes, but
conifer needles are more effective than broad leaves in accumulating particles. The edge to area
ratio (Woodcock, 1953) is also a key determinant of salt deposition to individual artificial leaves.
A strong negative correlation was observed under wind tunnel conditions between the area of
individual leaves and the deposition of coarse particles (Little, 1977). Small twigs and branches
were more effective particle collectors than were large branches and trunks of trees (Smith,
1984). Lead particles accumulated 20-fold more on woody stems than on needles of white pine
(Pinus strobus), even though needles displayed a 10-fold greater total surface area (Heichel and
Hankin, 1976). Deposition has heaviest at the tips of individual leaves.
Rough, pubescent broadleaf discs collected coarse (5.0 jim) particles up to 7-fold more
efficiently than glabrous leaf discs (Little, 1977). Laminae, petioles, and stems all differed in
collection efficiency. Pubescent leaves of sunflower (Helianthus annuus) collected coarse
particles nearly an order of magnitude more efficiently than the glabrous leaves of tulip poplar
(Liriodendron tulipifera) under wind tunnel conditions (Wedding et al., 1975). The rough
pubescent leaves of nettle (Urtica dioica) were more effective in capturing coarse PM10 than
were the densely tomentose leaves of poplar (Populus alba) or smooth leaves of beech
(Fagus sylvaticd).
Conifer needles are more efficient than broad leaves in collecting particles by impaction.
This reflects the small cross section of the needles relative to the larger leaf laminae of
broadleaves and the greater penetration of wind into conifer than broadleaf canopies. Conifers
were more effective in removing coarse («20 jim) particles of ragweed pollen from the
atmosphere than were broadleaf trees (Steubing and Klee, 1970) and in intercepting even coarser
rain particles (Smith, 1984). Conifers were also more effective in retaining and accumulating
particles against reentrainment by wind and removal by rain, particularly on foliar surfaces
where they are likely to be most biologically active. Seedlings of white pine (Pinus strobus) and
oak (Quercus rubrd) initially retained between a quarter (pine) and a third (oak) of very coarse
(88 to 175 jim) 134Cs-tagged quartz particles applied under field conditions (Witherspoon and
Taylor, 1971). After 1 h, the pine retained over 20% of the 134Cs-tagged particles; whereas the
oak retained only approximately 3%. Long-term retention of the particles was concentrated at
the base of the fascicles in pine and near the surface roughness caused by the vascular system on
leaves of oak. The sheltered locations available in the conifer foliage contribute substantially to
4-29
-------
greater retention of particles. For similar reasons, grasses also are efficient particle collectors
(Smith and Staskawicz, 1977) with long-term retention mostly in the ligule and leaf sheath.
Because of the strong relationship between particle size and deposition, the sharply
increasing humidity gradient near transpiring foliar surfaces may cause hygroscopic particles to
behave, at the immediate surface, as larger particles than reflected in ambient measurements at
reference heights. This needs to be further considered (Wesely and Hicks, 2000). Recent
deposition models (Ruijgrok et al., 1997; Zhang et al., 2001) have introduced the role of ambient
humidity but lack sufficient emphasis on the role of vegetation itself in modifying humidity near
the surface. This may be significant, as the size of a dry, 0.5 jim particle of ammonium sulfate
may increase to about 3.5 jim at saturated humidity (Ruijgrok et al., 1997). Kinetic analyses
have suggested that full (95 to 100%) equilibration to the new diameter will occur during the
deposition process in ambient humidity.
Wind tunnel studies also demonstrated equivalent deposition properties for 3.4 jim
particles of dense lead chloride and the 6.8 jim particles of the less dense uranine dye. These
particles were shown to be aerodynamically equivalent, substantiating the use of aerodynamic
diameter as a classification parameter for particle deposition.
Canopy Surface Effects on Vd. In general, surface roughness increases particulate
deposition. As a result, Vd is typically greater for a forest than for a field or nonwoody wetland
and greater for a field than for a water surface. The contrasting transport properties and
deposition velocities of different size particles lead to predictable patterns of deposition.
For coarse particles, the upwind leading edges of forests, hedge rows, and individual plants, as
well as of individual leaves, are primary sites of deposition. Impaction at high wind speed and
the sedimentation that follows the reduction in wind speed and carrying capacity of the air in
these areas lead to preferential deposition of larger particles.
Air movement is slowed in proximity to vegetated surfaces. Resulting log profiles of wind
and pollutant concentrations in the near-surface turbulent boundary layer above canopies reflect
surface characteristics of roughness length, friction velocity, and displacement height. Plasticity,
streamlining, and oscillations of foliar elements also alter the aerodynamic roughness and the
level of within-canopy turbulence. Canopies of uneven age or with a diversity of species are
typically aerodynamically rougher and receive larger inputs of pollutants than do smooth, low, or
4-30
-------
monoculture vegetation (Garner et al., 1989; Wiman and Agren, 1985). Canopies on slopes
facing the prevailing winds and individual plants on the windward edges of discontinuities in
vegetative cover over which roughness increases receive larger inputs of pollutants than more
sheltered, interior canopy regions. For example, some 80% of coarse particulate sea salt was
deposited on the upwind edge of a hedgerow (Edwards and Claxton, 1964), and the ragweed
pollen concentration was reduced by 80% within 100 m of the upwind edge of a forest
(Neuberger et al., 1967).
Beier et al. (1992) and Beier (1991) discussed two methods for estimating the dry
deposition of base cations to forest edges: (1) a difference method between measured
precipitation and throughfall concentrations of base cations, and (2) a calculation method based
on known ratios of Na+ deposition in wet and dry forms (Ulrich, 1983). A combination of these
two approaches has produced the best estimates of SO42 , Ca2+, Mg2+, and K+ particle deposition.
Using these methods, particulate SO42 (Beier, 1991) and particulate Ca2+, Mg2+, and K+ (Beier
et al., 1992) were found to decrease by an order of magnitude from the forest edge to the forest
interior. A number of authors also have shown that particle deposition is elevated at forest edges
when compared to a uniform forest canopy (Draaijers et al., 1988; Grennfelt, 1987; Lindberg and
Owens, 1993), and Draaijers et al. (1992) reported that differences are likely to exist between
forest types because of variable canopy structure. Draaijers et al. (1988) further emphasized that
enhanced particle deposition at or near forest edges is strongly dependent on the velocity and
wind direction during observations.
The factors leading to horizontal gradients are confounded by time- and distance-related
sedimentation. For example, geologic dust (mostly around 7 jim aerodynamic diameter) collects
on stems of wild oats (Avena spp.; Davidson and Friedlander, 1978) and on eastern white pine
(Heichel and Hankin, 1972; Smith 1973) downwind of roadways. Rapid sedimentation of coarse
crustal particles suggests that potential direct effects may be restricted to roadway margins,
forest edges, and, because of the density of unpaved roads in agricultural areas, crop plants.
Simulated deposition to an ecologically complex, mixed canopy was considerably higher
than to a pure spruce stand in which most of the leaf area was concentrated in regions of low
wind speed. Limitations to the application of these models to predict deposition over large
regions include an incomplete understanding both of the nature of microscopic particle-surface
4-31
-------
interactions and of the effects of complex terrain and species composition on macroscopic
transport processes.
Macroscopic turbulent transport processes, related to ra at successive layers through the
canopy, can be separated from microscopic processes related to rb and rc (or rcp) at each
deposition surface (e.g., Peters and Eiden, 1992; Wiman and Lannefors, 1985). The
macroscopic approach deals with deposition as the product of a turbulent diffusion coefficient
and a given concentration (Cz) at each canopy layer, both of which vary with particle size and
with height (z) in the canopy. The microscale parameters involve those factors that determine
absorption of a particle at each surface as captured imperfectly by rc. Shelter effects caused by
the crowding of foliar elements within the canopy can be ignored if the wind speed within each
canopy layer is specified. This approach requires knowledge of the vertical distribution of
particle concentration and foliage density in the canopy airspace along with profiles of wind
speed or turbulence.
Once introduced into a forest canopy, elements associated with coarse particles tend to
decrease markedly with canopy depth; whereas elements associated with fine particles do not
(Lovett and Lindberg, 1992). Trace elements and alkaline-earth elements are enriched below the
canopies of both southern (Lindberg et al., 1986) and northern (Eaton et al., 1973) hardwood
forests. Vertical gradients in concentration of coarse particles and of elements associated with
coarse particles were observed in a mixed conifer/birch forest canopy (Wiman and Lannefors,
1985; Wiman et al., 1985) and in a mixed oak forest (e.g., Ca2+, Figure 4-6a; Lovett and
Lindberg, 1992). The highly reactive gas, HNO3, also exhibited a vertical gradient but with a
steep decline at the top of the canopy (Figure 4-6b). Lovett and Lindberg (1992) studied
concentration profiles of various gases and particles within an closed canopy forest and
concluded that coarse particle concentrations associated with elements like Ca2+ would decrease
markedly with depth in the canopy, but they found only minor reductions with depth in the
concentrations of fine aerosols containing SO42 , NH4+, and FT. These data suggest that all foliar
surfaces within a forest canopy are not equally exposed to particle deposition: upper canopy
foliage tends to receive maximum exposure to coarse and fine particles, but foliage within the
canopy tends to receive primarily fine aerosol exposures. Fine-mode particles (e.g., sulfate,
Figure 4-6c) and unreactive gases typically do not exhibit such vertical profiles, suggesting that
their uptake is smaller in magnitude and more evenly distributed throughout the canopy.
4-32
-------
25 -
^ 20-
E,
^7 15 -
•5 10-
5-
0
25 -
^ 20-
E,
^T 15 -
10 -
5-
0
A = 34%
i i i i
0.6 0.7 0.8 0.9
A = 6%
[df
9.5 10.0 0.08 0.10
Concentration (|jg/m3)
A = 43%
K+
A = 0%
0.12
0.14
Figure 4-6. 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.
Source: Based on data from Lovett and Lindberg (1992).
In multilayer canopies, simultaneous reentrainment and deposition may effectively uncouple
deposition from local concentration. Poly disperse size distributions of many chemical species
effectively prevent the use of a single estimate of Vd for any element if highly accurate results
are required.
Although gradients (Figures 4-6 and 4-7) may be related to local Vd within the canopy
(Bennett and Hill, 1975), the absence of a gradient may reflect either low or very high rates of
4-33
-------
i .u —
_0
2
"= -sr 0-8-
Ql 0)
0 =
5 re
o >
0 >,
-------
Simulation of the horizontal deposition patterns at the windward edge of a spruce forest
downwind of an open field with the canopy between 1 and 25 m above the ground indicated that
deposition was maximal at the forest edge, where wind speed and impaction were greatest.
Simulation of the vertical deposition pattern was more complex. Deposition was not greatest at
the top of the canopy where wind speed was highest, but at z = 0.75 h, where the balance
between leaf area (obstacles for impaction) and wind speed (momentum for impaction) was
optimal, although neither parameter alone was maximal. Simulated deposition in this spruce
forest increased considerably with increasing LAI at the forest edge, where wind speed was
insensitive to LAI but the number of obstacles increased. Inside the forest, where both wind
speed and impaction increasingly were attenuated by increasing LAI, deposition increased only
marginally in spite of the increase in obstacle frequency.
To scale surface-specific measurements of particle deposition to forest or crop canopies,
conversions of the following type have been suggested:
Vd.canopy = Vd surface * scaling factor, (4-5)
with empirical scaling factors proposed by Lindberg et al. (1988).
To appropriately scale surface-specific measurements of particle deposition to landscapes,
one must consider the complexity of grassland, crop, and forest canopies in order to avoid
serious over- or underestimates of particle deposition. Individual species exposed to similar
ambient concentrations may receive a range of particulate loading that is more closely related to
foliar damage than the ambient concentration (Vora and Bhatnagar, 1987).
Both uptake and release of specific constituents of PM may occur within a single canopy
(e.g., K+; Lovett and Lindberg, 1992). The leaf cuticular surface is a region of dynamic
exchange processes occurring through leaching and uptake. Exchange occurs with epiphytic
microorganisms and bark and through solubilization and erosion of previously deposited PM.
Vegetation emits a variety of particles and parti culate precursor materials. Terpenes and
isoprenoids predominate and, on oxidation, become condensation nuclei for heterogeneous
particle formation. Salts and exudates on leaves and other plant parts are continually abraded
and suspended as particles, as are plant constituents from living and dead foliage (Rogge et al.,
1993a). Soil minerals, including radioactive strontium, nutrient cations and anions, and trace
4-35
-------
metals are transferred to the active upper foliage and then to the atmosphere in this way.
Although not representing a net addition to an ecosystem, particle release from vegetation is a
mechanism for redistributing chemical pollutants derived from the soil or prior deposition within
a canopy, potentially enhancing direct effects and confounding estimates of Vd.
Range of Deposition Velocity
As noted in an earlier criteria document (U.S. Environmental Protection Agency, 1982)
and in McMahon and Denison (1979), estimates of Vd for PM10 particles to vegetation are
variable and suggest a minimum between 0.1 and 1.0 |im, as predicted using first principles
(Monteith and Unsworth, 1990; Sehmel, 1980). Determinations in wind tunnels with passive
collectors and micrometeorological methods tend to converge in this range. The range of Vd for
sulfate from passive collectors was found to be from 0.147 to 0.356 cm/s; and, from eddy
covariance techniques, a mean Vd of 0.27 cm/s was observed (Dolske and Gatz, 1984).
Micrometeorological techniques over grass (Wesely et al., 1985); indirect, inert collector
techniques within an oak forest (Lindberg and Lovett, 1985); and many other empirical
determinations (e.g., McMahon and Denison, 1979; Table 4-4) generally support this range.
Over aerodynamically smooth snow (Duan et al., 1988; Table 4-5), measurements of Vd were
an order of magnitude smaller. Very coarse particles, often non-size-specified primary geologic
material, frequently exhibit Vd greater than 1.0 cm/s (e.g., Clough, 1975). The increase in Vd
with decreasing size below 0.1 |im is probably hidden in most empirical determinations of Vd
because the total mass in this fraction is very small despite the large number of individual
particles. Table 4-6 shows published estimates of Vd with variability estimates for fine particles
of specified aerodynamic diameters dominated by a range of chemical species.
Ibrahim et al. (1983) evaluated the deposition of ammonium sulfate particles to a range of
surfaces and found that particles having a mean diameter of 0.7 jim had deposition velocities
ranging from 0.039 to 0.096 cm/s. Larger particles (having mean diameters of 7 |im) had greater
deposition velocities (between 0.096 and 0.16 cm/s). The authors further concluded that the
hygroscopic nature of the sulfate particle could increase its size and enhance deposition near
sources of water (e.g., snow). Using eddy correlation approaches, Hicks et al. (1989) found a
mean daily Vd for sulfur-containing PM to be 0.16 cm/s. However, they suggested that the Vd
value could be as high as 1 cm/s during the day and near zero at night.
4-36
-------
TABLE 4-4. REPORTED MEAN DEPOSITION VELOCITIES (Vd) FOR SULFATE,
CHLORINE, NITRATE, AND AMMONIUM-ION-CONTAINING PARTICLES
Chemical Species/ Surface
so42-
Inert plates
Inert plates
Inert plates
Inert bucket
Foliage
Chaparral
Grass canopy
Grass canopy
Pine foliage
Plant canopies
Grass canopy
cr
Inert bucket
Inert bucket
Beech canopy
Spruce canopy
Inert plates
Ceanothus
Glycine max
Ligustrum
Quercus
Quercus
summer
winter
Quercus
summer
winter
Pinus
Pasture
Ulmus
NH4+
Calluna/Molina
Ceanothus
Kalmia
Pinus
Vd(cm/s)a
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 etal. (1981)
Dasch (1987)
1 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.
4-37
-------
TABLE 4-5. REPRESENTATIVE EMPIRICAL MEASUREMENTS OF DEPOSITION
VELOCITY (Vd) FOR PARTICIPATE DEPOSITION
vd
x ± SE (cm/s)
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
(urn)
0.15-0.30
0.5-1.0
37987
0.1-2.0
(SO/l
(K+)
(Ca2+)
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
Reference
Duanetal. (1988)
Allen etal. (1991)
Wesely et al. (1985)
Lindberg and Lovett
(1985)
Chamberlain and
Little (1981)
TABLE 4-6. REPORTED MEAN DEPOSITION VELOCITIES FOR POTASSIUM
(K+), SODIUM (Na+), CALCIUM (Ca2+), AND MAGNESIUM (Mg2+) BASE CATION
CONTAINING PARTICLES
Chemical Species/Surface
K+
Inert plates
Inert bucket
Na+
Inert bucket
Inert plate
Ca2+
Inert plates
Inert plates
Inert bucket
Mg2+
Inert bucket
Vd (cm/s)
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)
4-38
-------
Lindberg et al. 1990 found a wide discrepancy between deposition velocities for NO3
between study sites in Oak Ridge, TN («2 cm/s) and Gottingen, Germany (-0.4 cm/s). They
suggest that the increased Vd at Oak Ridge could be explained by the primary occurrence of
NO3 in coarse particles that exhibit greater Vd than fine particles (Davidson et al., 1982). Large
values of Vd for base-cation-containing particles (> 1 cm/s) suggest their occurrence in coarse
particles (Lindberg and Lovett, 1985).
The several attempts to estimate Vd for SO42 , NO3 , and NH4+ with the throughfall mass
balance approach (Davidson and Wu, 1990; Gravenhorst et al., 1983; Hofken and Gravenhorst,
1982) have produced higher Vd values that are considered suspect. They have not been included
in Tables 4-4 and 4-6. Overestimates of Vd for SO42 and NO3 particles derived from
throughfall mass balance approaches may be the result of gaseous SO2 and HNO3 gaseous
deposition to foliar surfaces (Lindberg and Lovett, 1985). A similar contribution of NH3
deposition may lead to erroneously high Vd values for NH4+ when the throughfall method is
attempted in areas of high NH3 concentrations. Dolske's (1988) reported Vd values for NO3
deposition to soybean ranged from 0.4 to 31 with a mean of 0.24 cm/s. However, because
Dolske's leaf extraction measurements included a component of HNO3 vapor, the Vd values may
represent more than deposition caused by aerosol nitrate alone.
The quantitative importance of dry particulate deposition depends upon the chemical
species, topography, precipitation regime, and surface characteristics, including vegetation
properties. Across the diverse landscapes of the Integrated Forest Study (Johnson and Lindberg,
1992a), the relative contribution of dry deposition for Ca2+ ranged from about 0% to nearly 90%
(Figure 4-8). In contrast, for S the total range was from just over 0% to about 30%. An average
for these forest systems demonstrates that deposition of (usually coarse) base cations was nearly
50% by dry parti culate deposition (Figure 4-9). Both N and S were around 15%. While the
relative significance of dry particle deposition varies from site to site, it cannot be excluded from
the analysis at any site.
In a recent review, Wesely and Hicks (2000) concluded that a comprehensive
understanding of parti culate deposition remains a distant goal. In general, there is only modest
confidence in available paniculate deposition parameterizations at this time, although recent
experimental and theoretical efforts to improve this situation have been made (e.g., Erisman
et al., 1997, and companion articles; Zhang et al., 2001; Kim et al., 2000).
4-39
-------
100
M 60^
I_
1
1
Figure 4-8. Range of percent of total deposition delivered in the dry particulate fraction,
across the sites of the Integrated Forest Study.
Source: Johnson and Lindberg (1992a).
c
o
o
0.:
3
.2
(0
a.
*KJ
40
*JC am
35
OA «
aO
on _
£\1
H| 1% •
n.
N
Base Cations
Figure 4-9. Contribution of particulate deposition to total deposition of nitrogen, sulfur,
and base cations.
Source: Modified from Johnson and Lindberg (1992a); Lovett (1992); Lovett and Lindberg (1993); Lindberg et al.
(1990); Kelly and Meagher (1986).
The successful treatment of dry deposition of gaseous pollutants (e.g., SO2 and O3) linking
turbulent deposition to surface physiographical properties has allowed incorporation of these
models into landscape and regional scale models. This has allowed gaseous deposition to be
4-40
-------
scaled up for purposes of atmospheric chemistry and vegetation damage assessment. These
advances have not been matched by progress in deposition of PM. A serious remaining
impediment is the lack of suitable techniques for measuring deposition of heterogeneous and
polydisperse particles, such as the gradient and eddy covariance techniques that are used for
gaseous pollutant species. As with gaseous pollutants, parameterization of particle deposition in
hilly terrain, to patchy surfaces (small agricultural fields, forest edges), and under extremely
windy conditions, remains to be fully developed. These limitations must be addressed before a
full accounting of regional PM effects on vegetation in natural and managed ecosystems can be
achieved.
Occult Deposition
Gaseous pollutant species may dissolve in the suspended water droplets of fog and clouds.
The stability of the atmosphere and persistence of the droplets often allow a gas/liquid phase
equilibrium to develop. This permits the use of air mass history or ambient concentrations of
specific pollutants to estimate fog or cloud water concentrations. Further estimates of the
deposition velocity of the polluted droplets allows calculation of deposit!onal fluxes.
Unfortunately, interception of fog or cloud droplets by plant parts or other receptor surfaces
remains difficult to predict and to measure. Fog formation influences the total atmospheric
burden and deposition of particulate matter (Pandis and Seinfeld, 1989) by accreting and
removing particles from the air, by facilitating particle growth through aqueous oxidation
reactions, and by enhancing deposition as noted. Aqueous condensation may occur onto
preexisting fine particles, and such particles may coalesce or dissolve in fog or cloud droplets.
Material transported in fog and cloud water and intercepted by vegetation escapes detection by
measurement techniques designed to quantify either dry or wet deposition; hence it is hidden
(i.e., "occult") from the traditional measurements.
Occult deposition of fog and cloud droplets is by impaction and gravitational settling in
concert with the instantaneous particle (droplet) size. This mode of deposition may be
significant at high elevation sites, particularly near the base of orographic clouds. While coastal
fog may be generally more pristine than high-elevation continental fogs, this is not true in areas
subject to inland radiation fog (e.g., the increasingly polluted San Joaquin Valley of Central
California) and in areas where the marine layer is advected through a highly polluted urban area
4-41
-------
(e.g., the Los Angeles basin of Southern California). Fog water in these areas may be at least as
contaminated as that occurring at higher elevations.
Low elevation radiation fog has different formation and deposition characteristics than
high elevation cloud or coastal fog water droplets. A one-dimensional deposition model has
recently been described for a radiation fog episode (Von Glasow and Bott, 1999).
A substantially greater concentration of key polluting species (e.g., NO3 , SO42 , organics) may
be observed in smaller than in larger droplets in fog (Collett et al., 1999). Acidity differences
exceeding 1 pH unit were also observed in the San Joaquin Valley winter radiation fog, with
smaller particles being more acidic. This has implications for aqueous phase oxidation of sulfur
and nitrogen compounds, in particular, while sulfur oxidation by ozone (the dominant reaction in
this environment even during winter) is well known in typically acidic fog droplets. However,
the alkaline larger droplets in the San Joaquin Valley could lead to greater nitrate production
through aqueous ozonation reactions (Collett et al., 1999). The size class distinctions have
substantial implications for deposition of particulate pollutant species in the fog droplets due to
the larger Vd for impaction and occult deposition of the larger fog particles.
Acidic cloud water deposition has been associated with forest decline in industrialized
areas of the world (Anderson et al., 1999). Clouds can contain high concentrations of acids and
other ions. Cloud water typically is 5 to 20 times more acid than rain water. This can increase
pollutant deposition and increases the exposure of vegetation and soils at high-elevation sites by
more than 50% when compared with rainfall and dry deposition.
The widespread injury to mountain forests documented in West Germany since the 1970s
and other parts of Europe, and more recently in the Appalachian Mountains, has been attributed
to cloud water exposure reducing the cold tolerance of red spruce. Forest injury also has been
attributed to increased leaching of cations and amino acids and increased deposition of nitrogen
and aluminum toxicity resulting from acidic deposition and the combined effect of acidic
precipitation, acid fog, oxidants, and heavy metals (Anderson et al., 1999). The Mountain Acid
Deposition Program (MADPro) was initiated in 1993 as part of the Clean Air Status and Trends
Network (CASTnet). MADPro monitoring efforts focused on the design and implementation of
an automated cloud water collection system in combination with continuous measurement of
cloud liquid water content (LWC) and meteorological parameters relevant to the cloud
deposition process.
4-42
-------
Using the MADPro automated cloud-water collectors at three selected mountain sites
(Whiteface Mt, NY; Whitetop Mt., Va; and Clingman's Dome, TN), samples taken hourly from
nonprecipitating clouds during nonfreezing seasons of the year from 1994 to 1997 and promptly
analyzed for pH, conductivity, and the concentrations of dissolved ions gave an indication of
exposures at each of the three sites. Cloud LWC was measured at each site. The mean cloud
water frequencies and LWC were higher at Whiteface Mountain, NY than in the southern
Appalachians. The four most prevalent ions found in cloud water samples, in order of
decreasing concentrations, were usually SO42 , H+, NH4+, and NO3 . The concentrations of these
ions tended to co-vary within cloud events and typically there was an inverse relationship
between LWC of the cloud and ionic concentration of the cloudwater. Highest ionic
concentrations were seen in mid-summer. Ionic concentrations of samples from southern sites
were significantly higher than samples from Whiteface Mountain, but further analysis indicated
that this was at least partly due to north to south differences in the LWC of clouds (Anderson
etal., 1999).
Several factors make occult deposition particularly effective for the delivery of dissolved
and suspended materials to vegetation. Concentrations of particulate-derived materials are often
many-fold higher in cloud or fog water than in precipitation or ambient air in the same area due
to orographic effects and from gas-liquid partitioning coefficients of specific chemical species.
Fog and cloud water deliver PM in a hydrated (and, therefore, bioavailable) form to foliar
surfaces. Previously dry-deposited PM may also become hydrated through delinquence or by
dissolution in the film of liquid water from fog deposition. The presence of fog itself maintains
conditions of high relative humidity and low radiation, thus reducing evaporation and
contributing to the persistence of these hydrated particles on leaf surfaces. Deposition of fog
water is very efficient (Fowler et al., 1991) with a Vd (fog 10 to 24 jim; Gallagher et al., 1988)
essentially equal to the aerodynamic conductance for momentum transfer (ra)-1. This greatly
enhances deposition by sedimentation and impaction of submicron aerosol particles that exhibit
very low Vd prior to fog droplet formation (Fowler et al., 1989). The near equivalence of Vd
and (ra)-1 simplifies calculation of fog water deposition and reflects the absence of vegetative
physiological control over surface resistance. Fog particles outside this size range may exhibit
Vd below (rj"1. For smaller particles, this decline reflects the increasing influence of still air and
boundary layer effects on impaction as particle size and momentum decline. For larger particles,
4-43
-------
momentum is sufficient to overcome these near surface limitations, but Vd may decline as
turbulent eddy transport to the surface becomes inefficient with increasing inertia (Gallagher
et al., 1988). The deposition to vegetation for PM in fog droplets is directly proportional to wind
speed, droplet size, concentration, and fog density (liquid water content per volume air) although
the latter two may be inversely related. In some areas, typically along foggy coastlines or at high
elevations, occult deposition represents a substantial fraction of total deposition to foliar surfaces
(Fowler et al., 1991; Figure 4-2).
4.2.2.3 Magnitude of Deposition
Dry deposition of PM is most effective for coarse particles including primary geologic
material and for elements such as iron and manganese. Wet deposition is most effective for fine
particles of atmospheric (secondary) origin (e.g., nitrogen and sulfur, Table 4-6) and elements
such as cadmium, chromium, lead, nickel, and vanadium (Reisinger, 1990; Smith, 1990a,b,c;
Wiman and Lannefors, 1985). The occurrence of occult deposition is more restricted. The
relative magnitudes of the different deposition modes varies with ecosystem type, location,
elevation, and chemical burden of the atmosphere. For the Walker Branch Watershed, a
deciduous forest in rural eastern Tennessee, dry deposition constituted a major fraction of the
total annual atmospheric input of cadmium and zinc (-20%), lead («55%), and manganese
(-90%). Whereas wet deposition fluxes during precipitation events exceeded dry deposition
fluxes by one to four orders of magnitude (Lindberg and Harriss, 1981), dry deposition was
nearly continuous. Immersion of high-elevation forests in cloudwater may occur for 10% or
more of the year, significantly enhancing transfer of PM and dissolved gases to the canopy.
Occult deposition in the Hawaiian Islands dominated total inputs of inorganic N (Heath and
Huebert, 1999). Much of this N was volcanically derived during the generation of volcanic fog
in part through reactions with seawater. In this humid climate, the dominance of occult rather
than wet deposition is notable.
High-elevation forests receive larger particulate deposition loadings than equivalent low
elevation sites. Higher wind speeds enhance the rate of aerosol impaction. Orographic effects
enhance rainfall intensity and composition and increase the duration of occult deposition.
Coniferous species in these areas with needle-shaped leaves also enhance impaction and
retention of PM delivered by all three deposition modes (Lovett, 1984).
4-44
-------
In more arid regions, such as the western United States, the importance of dry deposition
may be larger. In the San Gabriel Mountains of southern California, for example, while annual
deposition of SO42 (partly of marine origin) was dominated by wet deposition (Fenn and Kiefer,
1999), deposition of NO3 was dominated by dry deposition, as was that of NH4+at two of three
sites. Similarly, at a series of low elevation sites in southern California (Padgett et al., 1999),
dry deposition of NO3 was dominated by dry deposition. In both cases, however, the
contribution of gaseous HNO3 was probably substantial.
Nitrates, Sulfates, and Cations
Much particulate sulfate and nitrate is found on particles in the 0.1- to 1.0-|im size range
(U.S. Environmental Protection Agency, 1982). However, most sulfate and nitrate, base cation
and heavy metal inputs to forested ecosystems result from the deposition of larger particles
(Chapter 2) (Lindberg and Lovett, 1985; Lindberg et al., 1982). The influence of aerodynamic
diameter is particularly critical for nitrogen species, because they exist as a wide range of
particle sizes in the atmosphere (Milford and Davidson, 1987). For example, at many sites in
North America, NO3 is characterized by a bimodal size distribution with modes above and
below 1 |im. The supermicron particles are often the result of reactions between HNO3 and
coarse alkaline aerosols (Wolff, 1984) as, for example, in the San Joaquin Valley of California
(Lindberg et al., 1990). Although the annual deposition of NH4+ is distributed similarly among
the fine and coarse particles, particulate NO3 is found predominantly in the coarse-particle
fraction (Table 4-7). Similar to the pattern for NH4+, the estimated annual deposition of SO42
particles occurs in both the fine- and coarse-particulate fractions (Table 4-8), while base cation
deposition is virtually restricted to contributions from coarse particles (Table 4-9).
Although the annual chemical inputs to ecosystems from particle deposition is significant
by itself, it is important to compare it with the total chemical inputs from all sources of
atmospheric deposition (i.e., precipitation, particles, and gaseous dry deposition). Figure 4-10
shows the mean percentage contribution of NO3 andNH4+, SO42 , and base cation-containing
particles to the total nitrogen, sulfur, and base cation deposition load to forest ecosystems
(derived from Tables 4-7 through 4-9). Although the mean contribution of particulate deposition
to cumulative nitrogen and sulfur deposition is typically less than 20% of annual inputs from all
4-45
-------
o =
^H O
ll
o
a.
—
u
'•£
sS
0.
sS
O
U
o
z
M*
z
z
Type
Region/Fo
o
— 10
O
2
.a S
(D
(D
O
s
g
-s
OO
OS
60
•a
OS
OS
OS
OS
Lovett and Lindb
TO CN
D< OS
•rt OS
* S
bO h-1
O ^H 1>
^H ^ ' ^_^ U1
a bo ^O bo
S IH oo a
P M^»
^ 2
13 M
S J3
£?P,
S^
position \
and Lin
udes
Johnson
nd
ude
4-46
-------
b-
-< H
O W
Q ^
Z >
w o
a ^H
to t/2
g W
O p^
Cw * — i
? J
"e3 • Q H
Q ^ t/3
ftH g W
0^0
&H ^J ^^
H cfl
•^ J
^ ^
hJ ^j
t1 O
IEAN ANNUAL S
3MPARED TO T(
« U
00 CC
TABLE 4-
PARTICLE
aj
c
i
R
=
_0
its
o
P
*H
"3
—
s
%
1
g
-*^
o
8.
P ^
O r:
U ®
"C
cS
a.
p
o
tM
Location
ion/Forest Type
6X
(2
CN CN CN CN c^i CN CN CN CN CN CN CN "^" CO CN
O^OO^H"-^-rsi»n^i-^o»nooo "m
fNi>|r)'^tp:^Hinoo^HasO */") iO CN O CO "^ ^O ' — ; ^O
T— r fsf -—^ i— r
^SS^^S^s^ss^l S^
(N^|-OSOS^|-0^in'£ir^_,.(N(NS t^OS
^o c — in ^o ^~ | in " ^ ^ ^o r^ w' os r —
tst-t-t-ts«nm'!j-U)Oi,-\f^o ooos
^H^t-infNfNfNmmSlj^H^"1 t- in^n
o guu 'Q "S^IQ^IU
t. soiled aofsi | ^
l^^ii^^l^^l^i II
|3gll|g|gg|g| 0^
1 ^ a a a a -s II
S^2 *• *• *• *• GJ QJ QJ QJ fll j t o , o ,
t^aaaawgooo u to &o^
^^^^•^^••I'aSaa 53 22 w ^> ^
__ ^4 i-H i-H i-H i-H WH O. O. Q. 7^ Q ^, . ^ C^
•g^^^^^^H^ S?S?S?g^^ g-g g
OOoOOoJS>:S2l^3'ra3OO
ZQ^K-lK-lK-lCO^C^C^C^^SOW^^
OO
Q>,
13
6P
1
i
ii
^
o"
OS
OS
^H
13
1
hJ
II
CN
OS
OS
t
J
_^
rH
S3
£?
42
3^'
T3 OO
to C Os
"y fs1 ^
t'^ n\ o
OS O
CS tj
precipitation, gases, an
g (1992a), 2 = Lindberg
y season from April to (
e s .s
O 42 fe
c£^ ^ o
ludes deposition
Johnson and Lir
ludes only the gr
c3 ll ' fj
S^r" !T
4-47
-------
H
^
Ro
^J r.
Q c/j
Z W
^^ 25
M^ PH
z %
HH O
S ^
O ^
TO
SS p^
>; P*
~~G& ^ ^fl
jfi Q S
^ hH r~i
"^ HH H
1 o &
•= PH =«
w o
s|° ^
^§H
O S sc
K ^ W
G ^ w
^ u o
^^ &i
0- Si FT-,
S ^(J O
Q « S
*~^r 1^^
n d H
O -^ td
^H r , LM
w g$
K H ^f-
< Q
P3 Ed
ANNUAL
COMPAR
Z 03
<< W
ON ^
4 *
TABLE
P
u
=
1
a
-f^
o
0
^
s
a
a
"o
H
13
°a "o
•2 H
'K
o
0
—
_y
-w
sS 85
cu is
a o
O Qe)
"S
U
OJ
sS
CO
S 1
S
_0
u
a
&
H*
-f^
Region/Fores
CN CN CN CN
O O O O
t- O OO ^H
^O c*~l t^ ^t"
OS 1^ O \O
OO -st t- CN
^H ^H rn CN
0000
oo m ^t- CN
^H >-H m CN
os t- o •""!
g
I 1 «
f .a S 8
1 1 « I
1 S g £
CS
u
•5 ^ s a a
g « ft ft ft
^ J iT sT sT
S M rt ,2 r2
o o o o o
Z Q j j j
m
OS
OS
oo
^
m
^H
CN
"Si
OS
Tennessee
1
Mixed Hardw
CS CS CS CS CS CS CS
o o o o o o o
O 00 Tt O m V£> O
CN CN CN W> CN \O CN
"
C-- -"xl" >/".>/". OO ^- >/".
I-H i/~) i/~) ^O f") OO OO
^0 ^H T-H Tt ^H
§O O O ^O O ^f
•/". •/". "^ f") OO OO
c-- ^- »/". »/". rs -^t I-H
«— i o o u
•i 'ft S g g m
£ £ & & & - 1
1 | -2 -2 -2 5 f
&o c? pi pi pi %
*3 so
S C
-g 8
o
O rt
a * ^
^ 1 ffi 1
4-48
-------
50-
c
;§ 40-
'35
o
Q.
a)
Q 30-
20H
0)
10-
Nitrate and
Ammonium
Sulphate Base Cations
Dry Particle Chemicals
Figure 4-10. 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.
atmospheric sources, particulate inputs of base cations average half the total base cations
entering forest ecosystems from the atmosphere.
An extensive comparison of particle to total chemical deposition is provided by the
Integrated Forest Study (IPS; Johnson and Lindberg, 1992a; Lovett, 1994; Lovett and Lindberg,
1993; Lindberg and Lovett, 1992; Ragsdale et al., 1992). Other similar data sets are available
(Kelly and Meagher, 1986; Miller et al., 1993; Lindberg et al., 1986, 1990). These data in
(Tables 4-7 through 4-9) clearly indicate that the contribution of coarse and fine aerosols to
deposition to forest ecosystems is strongly dependent on the chemical species.
Dry deposition is an important flux of sulfur and nitrogen compounds at all of the IPS sites
and ranges from 9 to 59% of total (wet + dry + cloud) deposition for sulfur, 25 to 70% for NO3
and 2 to 33% for NH4+. Only for NH4+ is wet deposition consistently greater than dry deposition
(Lovett, 1994).
After emission from their sources, air pollutants are transformed and transported by
atmospheric processes until deposited from the atmosphere to an aquatic or terrestrial ecosystem.
4-49
-------
As a result, ground-level concentrations of an air pollutant depend on the proximity to the
sources, prevailing meteorology, and nature and extent of atmospheric reactions between the
source and the receptor (Holland et al., 1999). A more direct relationship exists between source
strength and downwind ambient concentrations for primary air pollutants (e.g., SO2) than for
secondary pollutants (e.g., SO42 ). Interaction of the chemical and physical atmospheric
processes and source locations for all of the pollutants have a tendency to produce data patterns
that show large spatial and temporal variability.
Holland et al. (1999) analyzed CASTNet monitoring data and using generative additive
models (GAM) estimated the form and magnitude of trends of airborne concentrations of SO2
SO4 2, and nitrogen from 1989 to 1995 at 34 rural long-term CASTNet monitoring sites in the
eastern United States. These models provide a highly flexible method for describing potential
nonlinear relationships between concentrations, meteorology, seasonality, and time (e.g., how
weekly SO2 varies as a function of temperature). For most of the 34 sites in the eastern United
State, estimates of change in SO2 concentrations showed a decreasing functional form in 1989 to
1990, followed by a relatively stable period during 1991 to 1992), then a sharper decline
beginning in 1994 (Holland et al., 1999).
Regional trends of seasonal and annual wet deposition and precipitation-weighted
concentrations (PWCs) of sulfate in the United States over the period 1980 to 1995 were
developed by Shannon (1999) from monitoring data and scaled to a mean of unity. In order to
reduce some effects of year to year climatological variability, the unitless regional deposition
and PWC trends were averaged (hereafter referred to a CONCDEP). During the 16-year period
examined in the study, estimated aggregate emissions of SO2 in the United States and Canada
fell -12% from about 1980 to 1982, it remained roughly level for a decade and then fell another
-15% from 1992 to 1995 — for an overall decrease of about 18%. Eastern regional trends of
sulfate concentrations and deposition and their average CONCDEP, also exhibited patterns of
initial decrease, near steady state, and final decrease with year to-to-year variability. The overall
relative changed in CONCDEPs are greater than the changes in SO2 emissions.
Concentrations and calculated deposition (concentration times amount of water) of SO42
at the Hubbard Brook Experimental Forest (HBEF) in the White Mountains of central New
Hampshire have been measured since June of 1964 (Likens et al., 2001). These measurements
represent the longest continuous record of precipitation chemistry in North America. The
4-50
-------
long-term measurements generally concur with those of Shannon (1999) discussed above. Major
declines in emissions of SO2 have been observed during recent decades in the eastern United
States and have been correlated with significant decreases in SO42 concentrations in
precipitation (Shannon, 1999).
Deposition of sulfates and nitrates are clearly linked to emissions. Reduction in emissions
must occur before concentrations can be reduced below current levels (Likens et al., 2001).
Deposition is the key variable as sensitive ecosystems in the eastern North America have not yet
shown improvement in response to decreased emissions of SO2 (Driscoll et al., 1989; Likens
et al., 1996). Clearly, additions of other chemicals, such as nitric acid and base cations, must be
considered in addition to sulfur when attempting to resolve the acid rain problem (Likens et al.,
1996, 1998). The effects of sulfur and nitrogen deposition on ecosystems are discussed in
Section 4.2.2.2.
The long-term record indicates that a reduction in the deposition of basic cations (Ca2+,
Mg+2, K+, Na+) in bulk precipitation was associated with significant declines in SO42deposit!on
cited above for the HBEF region (Driscoll et al., 1989). Decreases in streamwater
concentrations of basic cations have decreased simultaneously, suggesting that streamwater
concentrations of basic cations are relatively responsive to changes in atmospheric inputs.
Regardless of the cause, the decline in atmospheric influx of basic cations could have important
effects on nutrient availability as well as on the acid/base status of soil and drainage water
(Driscoll et al., 1989).
Trace Elements
Deposition velocities for fine particles to forest surfaces have been reported in the range of
1 to 15 cm/s (Smith, 1990a). For example, total, annual heavy metal deposition amounts are
highly variable depending on specific forest location and upwind source strength (Table 4-10).
Lindberg et al. (1982) quantified the dry deposition of heavy metals to inert surfaces and to
leaves of an upland oak forest. As noted for other chemical species, Vd was highly dependent on
particle size and chemical species (Table 4-11) with the larger particles depositing more
efficiently.
The preferential association of heavy metals with fine particles results in reduced control in
emission control systems. Metal removal efficiencies for baghouse filters are typically 95 to
4-51
-------
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 (jim)
3.4 ±0.7
1.5±0.7
0.9 ±0.2
0.5
Deposition Rate
(pg/cm/h)
91 ±23
0.3 ±0.1
6±1
23 ±8
vd
(cm/s)
6.4 ±3. 6
0.37±0.18
0.38 ±0.1
0.06 ±0.01
Source: Lindbergetal. (1982).
TABLE 4-11. TOTAL HEAVY METAL DEPOSITION TO TEMPERATE
LATITUDE FORESTS
Forest Deposition kg ha/yr
Heavy Metal (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).
99% for all but mercury, but fine particle capture is much less efficient. Wet scrubber efficiency
varies with design and pressure drop, typically 50 to 90% (McGowan et al., 1993). Fine
particles also have the longest atmospheric residence times and, therefore, can be carried long
distances. Depending on climate conditions and topography, fine particles may remain airborne
for days to months and may be transported 1,000 to 10,000 km or more from their source. This
long-distance transport and subsequent deposition qualify heavy metals as regional- and global-
4-52
-------
scale air pollutants. Ecosystems immediately downwind of major emissions sources (such as
power generating, industrial, or urban complexes) may receive locally heavy inputs. Mass
balance budgets (inputs and outputs) of seven heavy metals (cadmium, copper, iron, lead,
manganese, nickel, and zinc) have been determined at the Hubbard Brook Experimental Forest
(White Mountain National Forest) in New Hampshire. This forest is about 120 km northwest of
Boston and relatively distant from major sources of heavy metal emissions. However,
continental air masses that have passed over centers of industrial and urban activity also
frequently follow storm tracks over northern New England. Resulting annual inputs for seven
heavy metals at Hubbard Brook for 1975 to 1991 are presented in Figure 4-11. Note that the
44-fold decrease in Pb deposition is correlated with removal of Pb from motor vehicle fuels.
1_
re
0)
&
BJ
^
x 1
x 10 1
X10-2
Cd
Cll
Ni
Pb
Mn
Zn
Fe
—^
••••
m-1
^^^m
^^—
— ^
J
p
m
i
i
i
D
n
•
i
I975 -1£
1980-1J
1985-1S
1990-1S
I79
)84
)89
)91
I I I I I I
0 0.005 0.01 0.015 0.02 0.025
Figure 4-11. Annual total deposition of heavy metals to Hubbard Brook Experimental
Forest, NH.
Source: Smith (1990a).
4-53
-------
Trace element investigations conducted in roadside, industrial, and urban environments
have shown that impressive burdens of particulate heavy metals accumulate on vegetative
surfaces. Lead deposition to roadside vegetation (prior to its removal from fuel) was 5 to 20,
50 to 200, and 100 to 200 times the Pb deposition to agricultural crops, grasses, and trees,
respectively, in non-roadside environments. In an urban setting, it has been estimated that the
leaves and twigs of a 30 cm (12 in.) diameter sugar maple remove 60, 140, 5800, and 820 mg of
Cd, Cr, Pb, andNi, respectively, during a single growing season (Smith, 1973).
Semivolatile Organics
Organic compounds partition between gas and particle phases, and particulate deposition
depends largely on the particle sizes available for adsorption (Pankow, 1987; Smith and Jones,
2000). Dry deposition of organic materials (e.g., dioxins, dibenzofurans, polycyclic aromatics)
is often dominated by the coarse fraction, even though mass loading in this size fraction may be
small (Lin et al., 1993) relative to the fine PM fraction. For example, measurements in Bavaria
in both summer and winter revealed that > 80% of organics were in the fine (< 1.35 jim) fraction
(Kaupp and McLachlan, 1999). Nevertheless, in most cases, calculated values of dry deposition
were dominated by the material adsorbed to coarse particles. Wet deposition, in contrast, was
dominated by the much larger amount of material associated with fine particles. In this Bavarian
environment (where monthly precipitation is about 50 mm in winter and summer), wet
deposition dominated, with dry deposition accounting for only 14 to 25% of total deposition
(Kaupp and McLachlan, 1999). Lower relative contents of more volatile species in summer than
winter (Kaupp and McLachlan, 1999) indicate the critical importance of gas-particle phase
interconversions in determining deposition.
Effective deposition of PM is required before biological effects on plants or ecosystems
can occur. It is clear that substantially improved techniques for monitoring and predicting
deposition will be required to characterize with certainty quantitative relationships between rates
and quantities of PM deposition associated with such effects.
4-54
-------
4.2.3 Assessment of Atmospheric PM Deposition Effects
Introduction
The discussion in the pages that follow assesses and characterizes the overall ecological
condition or integrity of the ecosystems within the United States affected by the deposition of
anthropogenic stressors associated with PM and indicates their status. The six Essential
Ecological Attributes (EEAs) - the landscape condition, biotic condition, and chemical/ physical
characteristics, ecological processes, hydrology/geomorphology, and natural disturbance regimes
(Table 4-1) - provide a hierarchical framework for assessing ecosystem status. Measurable
characteristics related to structure, composition, or functioning of ecological systems may be
determined by the use of endpoints or ecological indicators of condition that are significant
either ecologically or to society (Harwell et al., 1999).
The relationships among the EEAs are complex because all are interrelated (i.e., changes in
one EEA may affect, directly or indirectly, every other EEA). The ecological processes create
and maintain patterns composed of the elements in the system and their arrangement; in turn, the
patterns affect how the processes are expressed (Science Advisory Board, 2002). Changes in
patterns or processes result in changes in the status and functioning of an ecosystem. The
information in the sections that follow discusses changes in landscape and biotic patterns and
in ecological and chemical/physical processes resulting from the stressors in PM deposition
(Figure 4-12).
The elements of biotic condition are organized in a nested hierarchy with several levels.
These include the structural and composition aspects (patterns) of the biota within landscape,
ecosystem or ecological community, species/population, organism, and genetic/molecular levels
(Science Advisory Board, 2002). Within these biological levels of organization, changes in the
biodiversity, composition, and structural elements relate directly to functional integrity (such as
trophic status or structural integrity within habitats). Changes in biodiversity are of particular
significance in altering the functioning of ecosystems.
As previously stated, ecosystems are dynamic, self-adjusting, self-maintaining, complex,
adaptive systems, in which the patterns of the higher levels of biotic organization emerge from
the interactions and selection processes at localized levels (Levin, 1998). Ecosystem
components must have an adequate supply of energy, mineral nutrients, and water to maintain
themselves and function properly. During the ecological processes of energy and material flow,
4-55
-------
Hydrolic Alteration
Habitat Conversion
Habitat Fragmentation
Climate Change
invasive Non-native Species
Turbidity/Sedimentation
Pesticides
Disease/Pest Outbreaks
Nytrient Pulses
Metals
Dissolved Oxygen Depletion
Ozone (Tropospherie)
Nitrogen Oxides
Nitrates
Hydrolic Alteration
Habitat Conversion
Habitat Fragmentation
Climate Change
Over-Harvesting Vegitation
Large-Scale Invasive
Species Introduction
Large-Scale Disease/Pest
Outbreaks
Hydrolic Alteration
Habitat Conversion
Climate Change
Over-Harvesting Vegitation
Disease/Pest Outbreaks
Altered Fire Regime
Altered Flood Regime
Hydrolic Alteration
Habitat Conversion
Climate Change
Turbidity/Sedimentation
Pesticides
Nutrient Pulses
Metals
Dissolved Oxygen Depletion
Ozone (Tropospherie)
Nitrogen Oxides
Nitrates
Sulfates
Salinity
Acidic Deoosition
Hydrolic Alteration
Habitat Conversion
Climate Change
Pesticides
Disease/Pest Outbreaks
Nutrient Pulses
Dissolved Oxygen Depletion
Nitrogen Oxides
Nitrates
Sulfates
Hydrolic Alteration
Habitat Conversion
Habitat Fragmentation
Climate Change
Turbidity/Sedimentation
Figure 4-12. Sample stressors and the essential ecological attributes they affect
(after Science Advisory Board, 2002).
the energy obtained by plants (the producers) from sunlight during photosynthesis (primary
production) and chemical nutrients (e.g., nitrogen, phosphorus, sulfur) taken up from the soil are
transferred to other species (the consumers) within the ecosystem through food webs. The
movement of chemical nutrients (materials) through an ecosystem is cyclic, as the nutrients are
used or stored and eventually returned to the soil by decomposer organisms. Energy, on the
4-56
-------
other hand, is transferred from organism to organism through an ecosystem in food webs and,
finally, is dissipated into the environment as heat (Odum, 1993).
Ecosystem and community patterns are characterized by the interaction of their component
species, the ecosystems processes of energy flow, nutrient flux, water and material flow, and by
the effects their activities have on the physical and chemical environment. The flows of energy
and nutrient cycling provide the interconnectedness among the elements of the biotic hierarchy
and transform the community from a random collection of numerous species into an integrated
whole, an ecosystem. Elucidating these interactions across scales is fundamental to
understanding the relationships between biodiversity and ecosystem functioning (Levin, 1998).
Human existence on this planet depends on the life-support services provided by the
interaction of the different EEAs. Both ecosystem structure (biotic conditions) and function
(ecological processes) play essential roles in providing goods (products) and services
(Table 4-12; Daily, 1997). Ecosystem processes maintain clean water, clean air, a vegetated
Earth, and a balance of organisms — the functions that enable humans to survive. The benefits
they impart include the absorption and breakdown of pollutants, cycling of nutrients, binding of
soil, degradation of organic waste, maintenance of a balance of gases in the air, regulation of
radiation balance and climate, and the fixation of solar energy (World Resources Institute, 2000;
Westman, 1977; Daily, 1997). The economic benefits and values associated with ecosystem
functions, goods, and services as well as the need to preserve them because of their value to
human life are discussed by Costanza et al. (1997) and Pimentel et al. (1997).
Goods such as food crops, timber, livestock, fish, and drinking water usually have market
value, while ecosystem services such as flood-control benefits, wildlife habitat, cycling of
nutrients, and removal of air pollutants, usually are very difficult to measure (Goulder and
Kennedy, 1997).
Attempts have been made to calculate the value of biodiversity and the world's ecosystem
services and natural capital (Goulder and Kennedy 1997; Pimentel et al., 1997; Constanza et al.,
1997). These have been controversial because of a lack of agreement on the measurement and
philosophical basis for placing value on ecosystem services (Goulder and Kennedy, 1997; Heal,
2000; Moomaw, 2002). Constanza et al. (1997) state that it may never be possible to make a
precise estimate of the services provided by ecosystems; however, their estimates indicate the
relative importance of services, not their true values, considering that the loss of ecosystems can
4-57
-------
TABLE 4-12. PRIMARY GOODS AND SERVICES PROVIDED BY ECOSYSTEMS
Ecosystem
Goods
Services
Agmecosystems
Coastal
Ecosystems
Forest Ecosystems
Freshwater
Grassland
Ecosystems
Food crops
Fiber crops
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 aesthetic 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 Institute (2000).
4-58
-------
affect human existence. Constanza et al. (1997) refer to the loss of "total values" for all portions
of all ecosystem services, whereas the relevant economic analysis for the PM standard setting is
for marginal changes in ecosystem goods and services due to varying levels of air pollution
effects, generally not total loss of services.
The prevailing economic approach to ascertaining value discussed above is benefit-cost
analysis. Benefit-cost analysis implicitly adopts the utilitarian basis for value. That is, the value
of a given living thing is the amount of human benefit or satisfaction the thing provides (Goulder
and Kennedy, 1997). Benefit-cost analysis determines the environmental and public health
damage and then calculates an appropriate level of emissions reduction required to lower
concentrations to a standard-setting level. As the costs and benefits extend into the future at
differing rates, the question arises of what discount rate to use in determining the net present
value of the costs and benefits. A degraded ecosystem continues to provide fewer services for
the indefinite future (Moomaw, 2002).
The approach of Harwell et al. (1999) and the report of the Ecological Processes and
Effects Committee of the SAB (Science Advisory Board, 2002) also point out the need to
understand human effects and their costs on ecosystems so that management can define what
ecological conditions are to be desired. Further, they state that the establishment of ecological
goals involves a close linkage between scientists and decision makers in which science informs
decision makers and the public by characterizing the ecological conditions that are achievable
under particular management regimes. Decision makers then can make choices that reflect
societal values including issues of economics, politics, and culture. For management to achieve
their goals, the general public, scientific community, resource managers, and decision makers
need to be routinely apprised of the condition or integrity of ecosystems so that ecological goals
may be established (Harwell et al., 1999).
Biodiversity, the variety of life, encompasses all levels of biological organization,
including species, individuals, populations, and ecosystems (Wilson, 1997). Human-induced
changes in biotic diversity and alterations in the structure and functioning of ecosystems are the
two most dramatic ecological trends of the past century (Vitousek et al., 1997). There are few
ecosystems on Earth today that are not influenced by humans (Freudenburg and Alario, 1999;
Vitousek et al., 1997; Matson et al., 1997; Noble and Dirzo, 1997). The scientific literature is
filled with discussions of the importance of ecosystem structure (patterns) and function
4-59
-------
(processes). The deposition of PM from the atmosphere has the potential to alter ecosystem
structure and function by altering nutrient cycling and changing biodiversity. We must,
therefore, determine how ecosystems respond to both natural and anthropogenic stresses so as
to know whether anthropogenic stresses are affecting ecosystem services and products
(Table 4-12).
Concern has risen in recent years regarding the consequences of changing the biological
diversity of ecosystems (Tilman, 2000; Ayensu et al., 1999; Wall, 1999; Hooper and Vitousek,
1997; Chapin et al., 1998). The concerns arise because human activities are creating
disturbances that decrease biodiversity, alter the complexity and stability of ecosystems, and
produce changes in ecological processes and the structure, composition and function of
ecosystems (Figure 4-13; Pimm, 1984; Levin, 1998; Chapin et al., 1998; Peterson et al., 1998;
Tilman, 1996; Tilman and Downing, 1994; Wall, 1999; Daily and Ehrlich, 1999). The above
changes can affect ecosystem services vital to human life.
Hydrologic
CO2 and
Temperature
Changes
Food Supply
and Demand
Water Use and
Nutrient Loss
Land Precipitation / Erosion,
Transfer- and Temp- / Changes
mation erature / in Water Flow
and Temperature\
Loss of Crop
Genetic Diversity
Change in
Transpiration
and Albedo
Freshwater
Supply and Demand
Habitat
Change
Forest Product
Supply and Demand
Loss and
Fragmentation
of Habitat
Reduced Resilience
to Change
Figure 4-13. Linkages among various ecosystem goods and services (food, water,
biodiversity, forest products) and other driving forces (climate change).
Source: Modified from Ayensu et al. (1999).
4-60
-------
4.2.3.1 Effects on Vegetation and Ecosystems
Exposure to a given mass concentration of airborne PM may lead to widely differing
phytotoxic responses, depending on the particular mix of deposited particles. Effects of
particulate deposition on individual plants or ecosystems are difficult to characterize because of
the complex interactions among biological, physicochemical, and climatic factors. Most direct
effects, other than regional effects associated with global changes, occur in the severely polluted
areas surrounding industrial point sources, such as limestone quarries, cement kilns, and metal
smelting facilities. Fine particles are more widely distributed from their sources than are coarse
particles. Experimental applications of PM constituents to foliage typically elicit little response
at the more common ambient concentrations. The diverse chemistry and size characteristics of
ambient PM and the lack of clear distinction between effects attributed to phytotoxic particles
and to other air pollutants further confound understanding of the direct effects on foliar surfaces.
The majority of the documented toxic effects of particles on vegetation reflect their chemical
content (e.g., acid/base, trace metal, nutrient), surface properties, or salinity. Studies of the
direct effects of particles on vegetation have not yet advanced to the stage of reproducible
exposure experiments. The difficulties of experimental application of ambient particles to
vegetation have been discussed by Olszyk et al. (1989). Studies indicate many phytotoxic gases
are deposited more readily, assimilated more rapidly, and lead to greater direct injury of
vegetation than do most common particulate materials (Guderian, 1986). The dose-specific
responses (dose-response curves) obtained in early experiments following the exposure of plants
to phytotoxic gases generally have not been observed following the application of particles.
Unlike gaseous dry deposition, neither the solubility of the particles nor the
physiographical activity of the surface is likely to be of first order importance in determining
deposition velocity (Vd). Factors that contribute to surface wetness and stickiness may be
critical determinants of deposition efficiency. Available tabulation of deposition velocities are
highly variable and suspect. However, high-elevation forests receive larger particle deposition
loadings than equivalent lower elevations sites because of higher wind speeds and enhanced
rates of aerosol impaction, orographic effects on rainfall intensity and composition, increased
duration of occult deposition, and, in many areas, the dominance of coniferous species with
needle-shaped leaves (Lovett, 1984), (see Section 4.2.2.3 for a discussion of deposition
efficacy). Recent evidence indicates that all three modes of deposition (wet, occult, and dry)
4-61
-------
must be considered in determining inputs to ecosystems or watersheds, because each may
dominate over specific intervals of space.
Atmospheric PM may affect vegetation directly following deposition on foliar surfaces or
indirectly by changing the soil chemistry or by changing the amount of radiation reaching the
Earth's surface through PM-induced climate change processes. Indirect effects, however, are
usually the most significant because they can alter nutrient cycling and inhibit plant nutrient
uptake.
Direct Effects of Particulate Matter Deposition
Coarse and fine PM deposition affects both patterns in the elements of EEA categories
subsumed under biotic condition and processes subsumed under chemical/physical and
ecological processes in the EPEC framework described earlier. Measurable responses have been
observed as reductions in photosynthesis, changes in soil salinity, and foliar effects resulting
from nitrate, sulfate, and acidic and heavy metal deposition.
Particles transferred from the atmosphere to foliar surfaces may reside on the leaf, twig, or
bark surface for extended periods; be taken up through the leaf surface; or be removed from the
plant via resuspension to the atmosphere, washing by rainfall, or litter-fall with subsequent
transfer to the soil. Any PM deposited on above-ground plant parts may exert physical or
chemical effects. The effects of "inert" PM are mainly physical; whereas those of toxic particles
are both chemical and physical. The effects of dust deposited on plant surfaces or soil are more
likely to be associated with their chemistry, rather than their mass and chemical effects may be
more important than any physical effects (Farmer, 1993). Nevertheless, vegetative surfaces
represent filtration and reaction/exchange sites (Tong, 1991; Youngs et al., 1993).
Direct Effects of Coarse Particles
Coarse particles, ranging in size from 2.5 to 100 jim, are chemically diverse, arise
predominantly from local sources, and are typically deposited near their sources because of their
sedimentation velocities. Airborne coarse particles are derived from the following sources:
road, cement kiln, and foundry dust; fly ash; tire particles and brake linings; soot and cooking oil
droplets; biogenic materials (e.g, plant pollen, fungal spores, bacteria and viruses); abraded plant
fragments; sea salt; and hydrated deliquescent particles of otherwise fine aerosol. In many rural
4-62
-------
areas and some urban areas, the majority of the mass in the coarse particle mode derives from
the elements silicon, aluminum, calcium, and iron, suggesting a crustal origin as fugitive dust
from disturbed land, roadways, agriculture tillage, or construction activities. Rapid
sedimentation of coarse particles tends to restrict their direct effects on vegetation largely to
roadsides and forest edges.
Physical Effects — Radiation. Dust can cause physical and chemical effects. Deposition
of inert PM on above-ground plant organs sufficient to coat them with a layer of dust may result
in changes in radiation received, a rise in leaf temperature, and the blockage of stomata.
Increased leaf temperature and heat stress, reduced net photosynthesis, and leaf chlorosis,
necrosis, and abscission were reported by Guderian (1986). Road dust decreased the leaf
temperature on Rhododendron catawbiense by ~4 °C (Eller, 1977); whereas foundry dust caused
an 8.7 °C increase in leaf temperature of black poplar (Populus nigrd) under the conditions of
the experiment (Guderian, 1986). Deciduous (broad) leaves exhibited larger temperature
increases because of particle loading than did conifer (needle) leaves, a function of poorer
coupling to the atmosphere. Inert road dust caused a three- to four-fold increase in the
absorption coefficient of leaves of English ivy (Eller, 1977; Guderian, 1986) for near infrared
radiation (NIR; 750 to 1350 nm). Little change in absorption occurred for photosynthetically
active radiation (PAR; 400 to 700 nm). The increase in NIR absorption may be accounted for by
a decease in reflectance and transmission in these wavelengths. The amount of energy entering
the leaf increased by -30% in the dust-affected leaves. Deposition of coarse particles increased
leaf temperature and contributed to heat stress, reduced net photosynthesis, and caused leaf
chlorosis, necrosis, and abscission (Dassler et al., 1972; Parish, 1910; Guderian, 1986; Spinka,
1971).
Starch storage in dust-affected leaves increased with dust loading when under high
(possibly excessive) radiation, but decreased following dust loading when radiation was limiting.
These modifications of the radiation environment had a large effect on single-leaf utilization of
light. The boundary layer properties, determined by leaf morphology and environmental
conditions, strongly influenced the direct effects of particle deposition on radiation heating
(Eller, 1977; Guderian, 1986) and on gas exchange. Brandt and Rhoades (1973) attributed the
reduction in the growth of trees to crust formation from limestone dust on the leaves. Crust
4-63
-------
formation reduced photosynthesis and the formation of carbohydrates needed for normal growth,
induced premature leaf-fall, damaged leaf tissues, inhibited growth of new tissue, and reduced
starch storage. Dust may decrease photosynthesis, respiration, and transpiration; and it may
allow penetration of phytotoxic gaseous pollutants, thereby causing visible injury symptoms and
decreased productivity. Permeability of leaves to ammonia increased with increasing dust
concentrations and decreasing particle size (Farmer, 1993).
Dust also has been reported to physically block stomata (Krajickova and Mejstfik, 1984).
Stomatal clogging by PM from automobiles, stone quarries, and cement plants was also studied
by Abdullah and Iqbal (1991). The percentage of clogging was low in young leaves when
compared with old, mature leaves and the amount of clogging varied with species and locality.
The maximum clogging of stomata observed was about 25%. The authors cited no evidence that
stomatal clogging inhibited plant function. The heaviest deposit of dust usually occurs on the
upper surface of broad-leaved plants; whereas the majority of the stomata are on the lower
surface where stomatal clogging would be less likely.
Chemical Effects. The chemical composition of PM is usually the key phytotoxic factor
leading to plant injury. On hydration, cement-kiln dust liberates calcium hydroxide, which can
penetrate the epidermis and enter the mesophyll; in some cases, this has caused the leaf surface
alkalinity to reach a pH of 12. Lipid hydrolysis, coagulation of the protein compounds, and
ultimately plasmolysis of the leaf tissue reduce the growth and quality of plants (Guderian,
1986). In experimental studies, applications of cement-kiln dust of known composition for 2 to
3 days yielded dose-response curves between net photosynthetic inhibition or foliar injury and
dust application rate (Darley, 1966). Lerman and Darley (1975) determined that leaves must be
misted regularly to produce large effects. Alkalinity was probably the essential phytotoxic
property of the applied dusts.
Salinity. Particulate matter enters the atmosphere from oceans following the mixing of air
into the water column and the subsequent bursting of bubbles at the surface. The effervescence
of bubbles on the surface of the ocean forcefully ejects droplets of sea water into the air. These
droplets, concentrated by evaporation, are carried inland by wind and deposited on the seaward
side of coastal plants (Boyce, 1954). This occurs largely at the surf line (i.e., near land and
4-64
-------
potentially sensitive terrestrial receptors). This process can be a significant source of sulfate,
sodium, chloride, and trace elements (as well as living material) in the atmospheric aerosol that
impacts coastal vegetation. Sea-spray particles (Taback et al., 1979) are approximately 24%
greater in size than 10 jim, and 54% are between 3 and 10 jim. Thus, only about 20% are fine
(0 to 2.5 |im) particles; and deposition by sedimentation and impaction is concentrated near the
coast, whereas the particle size distribution shifts toward the fine fraction over longer inland
transport distances. Airborne concentrations of this marine PM decrease quickly with distance
inland from the surf line both by deposition and dilution within the atmospheric mixed layer
(McKay et al., 1994; Nelis et al., 1994). Near-shore sediments, with associated pollutants
present in coastal runoff, may be suspended in the surf and reentrained into the air. This can be a
substantial source of microorganisms and of radionuclides to coastal vegetation (Nelis et al,
1994; McKay et al., 1994).
Sea-salt particles can serve as nuclei for the absorption and subsequent reaction of other
gaseous and particulate air pollutants. Both nitrate and sulfate from the atmosphere have been
found to associate with coarse and fine sea-salt particles (Wu and Okada, 1994). Direct effects
on vegetation reflect these inputs, as well as the classical salt injury caused by the sodium and
chloride that constitute the bulk of these particles. Foliar accumulation of airborne salt particles
may lead to foliar injury, thusly affecting the species composition in coastal environments
(Smith, 1984).
The effects of winds and sea spray on coastal vegetation has been reported in the literature
since the early 1800s (Boyce, 1954). However, there has been some controversy about whether
the injury to coastal vegetation resulted from windblown aerial salts or from mechanical injury
(i.e., sand blasting) due to wind alone. Though the significance of sea water dashed on fore
dunes and rocky coasts had been recognized by several authors, Wells and Shunk (1937, 1938)
and Wells (1939) were the first to recognize the importance of salt spray in coastal ecology.
Wells and Shunk (1937) reported that salt spray carried over dunes was the most important
factor influencing growth form, zonation, and succession in coastal dunes. Salt spay injury was
recorded 1.25 miles inland on the North Carolina coast. On the basis of observations in the
Cape Fear area, they determined that the shape of coastal "wind form" shrubs were the result of
sea spray carried by high winds. They found injury on shrubs only near the coast, while those at
4-65
-------
greater distances inland showed no injury whatsoever after a strong southeast wind that persisted
for a period of 19 hours during cloudy weather and abundant soil moisture.
To determine the cause of injury, injured and uninjured shoots were titrated for chlorides.
A marked difference was observed between the injured and uninjured shoots (Wells and Shunk,
1937, 1938). Experimental spraying of shoots of woody plants with seawater resulted in a
pattern of injury similar to the injury observed on seaside shrubs. The absence of the more
inland species, such as persimmon (Diosporos virginiana), turkey oak (Quercus laevis), longleaf
pine (Pinuspalustris Mill., P. australis), and wire grass (Aristida stricta), was explained on the
basis of intolerance of these species to salt spray. The dominance of live oak (Quercus
virginiana), as a practically pure stand on Smith Island (also known as Bald Head Island), NC
and along the eastern and southern NC coast, was determined by Wells (1939) to be due to the
tree's tolerance to salt spray. He termed the long term stabilization of the live oak stand as a
new type of climax, the "salt spray climax." The later work of Costing and Billings (1942) near
Beaufort, NC corroborated the findings of Wells and Shunk (1937, 1938).
The report by Boyce (1954) is probably the most extensive on salt-spray communities.
Dune sands in many coastal areas have been shown to have extremely low concentrations of
dissolved salts. Studies have indicated that the salt content of the coastal dunes of Virginia,
Massachusetts, and California did not exceed the maximum occurring in ordinary cultivated
soils. Costing and Billings (1942) found no correlation between soil salinity and plant
distribution on the North Carolina coast. Although surface crusts of sand dunes have been
shown to have high concentrations of chlorides which could be attributed to sea spray,
concentrations of chlorides in underlying layers was low. The salt content of the surface layer
varied with exposure of the dunes to oceanic winds (Boyce, 1954).
Boyce (1954), Wells (1939), and Wells and Shunk (1938) concluded on the basis of their
studies that necrosis and death of plant tissues results from the high deposition of salt spray and
high accumulation of the chloride ion (Cl~) in the plant tissues. Very little salt is taken up by
plant roots; most enters through the aerial organs. Leaves of plants exposed to salt spray show
a distinct pattern of injury (Wells and Shunk, 1938). Necrotic areas first appear at the leaf tips
and upper margins and then progress slowly in an inverted "V" toward the petiole. This leaf
injury pattern was verified experimentally. Mechanical injury resulting from leaves and twigs
beating against each another in the wind causes the formation of small lesions through which salt
4-66
-------
can enter. After entry into the plant, the chloride ion is rapidly translocated to the apices of the
leaves and twigs where it accumulates to injurious concentrations and results in the death of only
a portion of the plant. The differential deposition and translocation of the chloride ion results in
the death of the seaward leaves and twigs. The result is the continued growth of the uninjured
branches in an inland direction. As a result, the canopy angle varies with the intensity of the
spray (Boyce, 1954).
Little or no mineral ions are available in the silicate sands of the coastal dunes.
Consequently, plants obtain mineral ions needed for growth from the salt spray. Seawater
contains all of the mineral ions required for plant growth except nitrogen and phosphorus.
The amount of nitrogen and phosphorus in seawater varies over a wide range (Boyce, 1954).
Experiments indicated that available nitrogen in sea spray was a conditioning factor.
Low nitrogen availability increased the tolerance of dune species to salt spray. Increasing
the availability of nitrogen resulted in a different pattern of plant zonation and distribution.
Dicotyledonous species were restricted to areas of lower spray intensity. The severity of
chloride injury was associated more with the amount of available potassium than with the
concentration of chlorides within the limits of 280 to 360 mg Cl/liter (Boyce, 1954).
Other sources of phytotoxic saline PM include aerosols from cooling towers and roadway
deicing salt. Cooling towers used to dissipate waste heat from steam-electric power generating
facilities may emit salt if brackish salt water is used as a coolant (McCune et al., 1977; Talbot,
1979). Foliar injury is related to salt droplets deposited by sedimentation or impaction from
cooling tower drift. The distance of the salt drift determines the amount of deposition and
location of injury. The environmental conditions most conducive to injury were absence of
precipitation, which can wash salt off leaves, and high relative humidity (Talbot, 1979).
Increased injury is associated with wind speed and salt concentrations. Typical toxicity
symptoms from acute exposures include marginal foliar necrosis and lesions, shoot-tip dieback,
leaf curl, and interveinal necrosis (McCune et al., 1977). Based on experimental data, Grattan
et al. (1981) observed that, to cause injury, salt deposited on leaf surfaces must dissolve and be
absorbed into leaf tissue. Their work also indicated the importance of RH in foliar uptake.
If RH remained below 70%, even heavy deposition of salt did not induce injury in peppers,
soybeans, or tomatoes.
4-67
-------
Injury to vegetation from the application of deicing salt was related to salt spray blown or
drifting from the highways (Hofstra and Hall, 1971; Viskari and Karenlampi, 2000). The most
severe injury was observed nearest the highways. The results presented in these studies agree
with that of Wyttenbach et al. (1989), who observed that conifers planted near roadway margins
in the eastern United States often exhibit foliar injury due to toxic levels of saline aerosols
deposited from deicing solutions. Piatt and Krause (1974) demonstrated that road and site
factors influence the spread of deicing salt into forested areas. The slope away from the road
influenced the distance from the road where injury was observed and the percent slope was
correlated with distance.
Effects of Fine Particles
Fine PM in rural areas is generally secondary in nature, having condensed from the vapor
phase or been formed by chemical reaction from gaseous precursors in the atmosphere, and is
generally smaller than 1 to 2.5 jim. Nitrogen and sulfur oxides, volatile organic gases,
condensates of volatilized metals, and products of incomplete combustion are common
precursors for fine PM. Many of these materials react with an oxidizing atmosphere to
contribute to high secondary PM concentrations during summer months in many U.S. areas or
during late fall and winter in areas with high nitrate concentrations. The conclusion that
sufficient data were not available for adequate quantification of dose-response functions for
direct effects of fine aerosols on vegetation, reached in the 1982 PM/SOX AQCD (U.S.
Environmental Protection Agency, 1982), continues to be true today. Only a few studies on the
direct effects of acid aerosols have been completed (U.S. Environmental Protection Agency,
1982). The major effects are indirect and occur through the soil (see Section 4.2.3.2).
Nitrogen. Nitrate is observed in both fine and coarse particles. Nitrates from atmospheric
deposition represent a substantial fraction of total nitrogen inputs to southeastern forests (Lovett
and Lindberg, 1986). However, much of this is contributed by gaseous nitric acid vapor; and a
considerable amount of the particulate nitrate is taken up indirectly through the soil. Garner
et al. (1989) estimated deposition of nitrogen to forested landscapes in eastern North America at
10 to 55 kg/ha/year for nitrate and 2 to 10 kg/ha/year for ammonium. About half of these values
were ascribed to dry deposition.
4-68
-------
Driscoll et al. (2003) estimated anthropogenic nitrogen inputs to eight large watersheds in
the northeastern United States for the year 1997. Inputs of total nitrogen deposition ranged from
14 kg/ha/year in the Casco Bay watershed, ME to 68 kg/ha/year in the Massachusetts Bay
watershed. Atmospheric deposition of nitrogen was the second largest source ranging in
amounts from 5 to 10 kg/ha/year (11 to 36% of the total) (Driscoll et al., 2003). Nitrogen
deposition in the western United States ranges from 1 to 4 kg/ha/year over much of the region to
as high as 30 kg/year downwind of major urban and agricultural areas. An unknown amount of
nitrogen deposited to the West Coast originates in Asia (Fenn et al., 2003a).
Atmospheric additions of particulate nitrogen in excess of vegetation needs are lost from
the system, mostly in leachate from the soil as nitrate. Managed agricultural ecosystems may be
able to utilize deposited particulate nitrogen more efficiently than native ecosystems, although
many cultivated systems also lose considerable nitrogen as nitrate in runoff, deep drainage, or
soil water. It has proven difficult to quantify direct foliar fertilization by uptake of nitrogen from
ambient particles.
There is no doubt that foliar uptake of nitrate can occur, as clearly shown by the efficacy of
foliar fertilization in horticultural systems. Potassium nitrate was taken up by the leaves of
deciduous fruit trees (Weinbaum and Neumann, 1977) and resulted in increased foliar nitrogen
concentrations. Not all forms of nitrogen are absorbed equally, nor are all equally benign.
Following foliar application of 2600 ppm of nitrogen as calcium nitrate Ca(NO3)2, (NH4)2SO4,
or urea [(NH2)2CO] to apple canopies (Rodney, 1952; Norton and Childers, 1954), leaf nitrogen
levels were observed to increase to similar levels; however, calcium nitrate and ammonium
sulfate caused visible foliar injury, whereas urea did not. Urea is generally the recommended
horticultural foliar fertilizer.
The mechanism of uptake of foliarly deposited nitrate is not well established. Nitrate
reductase is generally a root-localized enzyme: it is generally not present in leaves, but is
inducible there. Induction typically occurs when the soil is heavily enriched in NO3 . As the
root complement of nitrate reductase becomes overloaded, unreduced nitrate reaches the leaves
through the transpiration stream. Nitrate metabolism has been demonstrated in leaf tissue
(Weinbaum and Neumann, 1977) following foliar fertilization. Residual nitrate reductase
activity in leaves may be adequate to assimilate typical rates of particulate nitrate deposition.
4-69
-------
Uptake of nitrate may be facilitated by codeposited sulfur (Karmoker et al., 1991; Turner and
Lambert, 1980).
Nitrate reductase is feedback-inhibited by its reaction product NH4+. The common
atmospheric aerosol, ammonium nitrate [NH4NO3], therefore may be metabolized in two distinct
biochemical steps: first the ammonium (probably leaving nitric acid) and then the nitrate.
Losses of nitric acid by volatization during this process, if they occur, have not been quantified.
Direct foliar effects of particulate nitrogen have not been documented. Application of a
variety of fine nitrogenous aerosol particles (0.25 jim) ranging from 109 to 244 |ig/m3 nitrogen
with or without 637 |ig/m3 sulfur caused no consistent short-term (2- to 5-h) effect on gas
exchange in oak, maize, or soybean leaves (Martin et al., 1992).
Although no evidence exists for the direct transfer of nutrient parti culate aerosols into
foliage, a few studies give insights into the potential for ammonium and nitrate transfer into
leaves. Fluxes of both NO3 and NH4+, measured in wet deposition and in throughfall plus
stemflow in forests, commonly indicate higher fluxes of nitrogen above the canopy (Parker,
1983; Lindberg et al., 1987; Sievering et al., 1996), and imply net foliar uptake. Lovett and
Lindberg (1993) reported a linear relationship between inorganic nitrogen fluxes in deposition
and throughfall, suggesting that uptake may be considered passive to some extent.
Garten and Hanson (1990) studied the movement of 15N-labeled nitrate and ammonium
across the cuticles of red maple (Acer rubruni) and white oak (Quercus alba) leaves when
applied as an artificial rain mixture. Brumme et al. (1992), Bowden et al. (1989), and Vose and
Swank (1990) have published similar data for conifers. These studies show the potential for
nitrate and ammonium to move into leaves, where it may contribute to normal physiological
processes (e.g., amino acid production) (Wellburn, 1990). Garten (1988) showed that internally
translocated 35S was not leached readily from tree leaves of yellow poplar (Liriodendron
tulipiferd) and red maple (Acer rubrum), suggesting that SO42 would not be as mobile as the
nitrogen-containing ions discussed by Garten and Hanson (1990). Further, when the foliar
extraction method is used, it is not possible to distinguish sources of chemicals deposited as
gases or particles (e.g., HNO3, NO2, NO3 ), or sources of ammonium (deposited as NH3 or NH4+)
(Garten and Hanson, 1990).
Particle deposition contributes only a portion of the total atmospheric nitrogen deposition
reaching vegetation; but, when combined with gaseous and precipitation-derived sources, total
4-70
-------
nitrogen deposition to ecosystems has been identified as a possible causal factor leading to
changes in natural ecosystems (see Section 4.2.2).
Sulfur. Anthropogenic sulfur emissions are > 90% SO2. Most of the remaining emissions
of sulfur is directly as sulfate (U.S. Environmental Protection Agency, 1996a). Sulfur dioxide is
hydrophilic and is rapidly hydrated and oxidized to sulfite and bisulfite and then to sulfate,
which is -30-fold less phytotoxic. The ratio of SO42 :SO2 increases with the aging of the air
mass and, therefore, with distance from the source. Sulfate is sufficiently hygroscopic in humid
air that it may exist significantly in the coarse particulate fraction. Because dilution of both SO2
and parti culate SO42 occurs with distance from the source, damaging levels of parti culate SO4
are rarely deposited. In this case, gas to particle conversion is of benefit to local vegetation.
Sulfur is an essential plant nutrient. Low dosages of sulfur serve as a fertilizer, particularly
for plants growing in sulfur-deficient soil (Hogan et al., 1998). However, current levels of
sulfate deposition reportedly exceed the capacity of most vegetative canopies to immobilize the
sulfur (Lindberg, 1992; Johnson, 1984). Nitrogen uptake in forests may be regulated loosely by
sulfur availability, but sulfate additions in excess of needs do not typically lead to injury unless
deposited in acidic precipitation (Turner and Lambert, 1980).
There are few field demonstrations of foliar sulfate uptake (Krupa and Legge, 1986, 1998).
Sulfate in throughfall is often enriched above levels in precipitation. The relative importance of
foliar leachate and prior dry-deposited sulfate particles remains difficult to quantify (Cape et al.,
1992). Leaching rates are not constant and may respond to levels of other pollutants, including
acids. Uptake and foliar retention of gaseous and parti culate sulfur are confounded by variable
rates of translocation and accessibility of deposited materials to removal and quantification by
leaf washing. Following soil enrichment with 35SO42 in a Scots pine forest, the apparent
contribution of leachate to throughfall was only a few percent following an initial burst of over
90%, because of extreme disequilibrium in the labeling of tissue sulfate pools (Cape et al.,
1992).
Olszyk et al. (1989) provide information on the effects of multiple pollutant exposures
including particles (NO3 , 142 |ig/m3; NH4+, 101 |ig/m3; SO42 , 107 |ig/m3). They found that
only gaseous pollutants produced direct harmful effects on vegetation for the concentrations
documented, but the authors hypothesized that the long-term accumulation of the nitrogen and
4-71
-------
sulfur compounds contributed from particle deposition might have adverse effects on plant
nutrition over long periods of time. Martin et al. (1992) exposed oak (Quercus macrocarpa),
soybean (Glycine max), and maize (Zea mays) plants to acute exposures (2 to 5 h) of aerosols
(0.25 |im) containing only nitrate (109 |ig/m3), ammonium and nitrate (244 and 199 |ig/m3), or
ammonium and sulfate (179 and 637 |ig/m3). They found that these exposures, which exceeded
the range of naturally occurring aerosol concentrations, had little effect on foliar photosynthesis
and conductance. Martin et al. (1992) concluded that future investigations should focus on the
effects of particles on physiological characteristics of plants following chronic exposures.
Acidic Deposition. The effects of acidic deposition have been accorded wide attention in
the media and elsewhere (Altshuller and Linthurst, 1984; Hogan et al., 1998). Probably the most
extensive assessment of acidic deposition processes and effects is the NAPAP Biennial Report to
Congress: An Integrated Assessment (National Science and Technology Council, 1998).
Concern regarding the effects of acidic deposition on crops and forest trees has resulted in
extensive monitoring and research. Exposures to acidic rain or clouds can be divided into
"acute" exposures to higher ionic concentrations (several |imol/L) and "chronic" long-term
repeated exposures to lower concentrations (Cape, 1993). Pollutant concentrations in rainfall
have been shown to have little capacity for producing direct effects on vegetation (Altshuller and
Linthurst, 1984; Hogan et al., 1998); however, fog and clouds, which may contain solute
concentrations up to 10 times those found in rain, have the potential to cause direct effects.
More than 80% of the ionic composition of most cloudwater is made up of four major pollutant
ions: H+, NH4+, NO3 , and SO42 . Ratios of hydrogen to ammonium and sulfate to nitrate vary
from site to site with all four ions usually present in approximately equal concentrations.
Available data from plant effect studies suggest that hydrogen and sulfate ions are more likely to
cause injury than ions containing nitrogen (Cape, 1993).
The possible direct effects of acidic precipitation on forest trees have been evaluated in
experiments on seedlings and young trees. The size of mature trees makes experimental
exposure difficult, therefore necessitating extrapolations from experiments on seedlings and
saplings; however, such extrapolations must be used with caution (Cape, 1993). Both conifers
and deciduous species have shown significant effects on leaf surface structures after exposure to
simulated acid rain or acid mist at pH 3.5. Some species have shown subtle effects at pH 4 and
4-72
-------
above. Visible lesions have been observed on many species at pH 3 and on sensitive species at
pH 3.5 (Cape, 1993). The relative sensitivities of forest vegetation to acidic precipitation based
on macroscopic injury have been ranked as follows: herbaceous dicots > woody dicots >
monocots > conifers (Percy, 1991).
Huttunen (1994) described the direct effects of acid rain or acidic mist on epicuticular
waxes whose ultrastructure is affected by plant genotype and phenotype. The effects of air
pollutants on epicuticular waxes of conifers have received greater study than the waxes of other
species. Leafage and the shorter life span of broad-leaved trees make them less indicative of the
effects of acid precipitation. Many experimental studies indicate that epicuticular waxes that
function to prevent water loss from plant leaves can be destroyed by acid rain over a few weeks
(Huttunen, 1994). Epicuticular waxes are crucial to conifers because of their longevity and
evergreen foliage. Microscopic observations of epicuticular wax structures have, for a long
time, suggested links between acidic deposition and aging. In Norway spruce (Picea abies), acid
rain causes not only the aging of needles (which in northern conditions normally last from
11 to 14 years) to be shortened, but also accelerates the erosion rate of the waxes as the needles
age.
The effects of acidic precipitation and fog on red spruce (Picea rubens) have been studied
extensively (Schier and Jensen, 1992). Visible foliar injury of the needles in the form of a
reddish-brown discoloration has been observed on red spruce seedlings experimentally exposed
to acidic mist, but this visible symptom has not been observed in the field. Ultrastructural
changes in the epicuticular wax were observed both experimentally and on spruce growing at
high elevations. Laboratory studies indicate that visible injury usually does not occur unless the
pH is 3 or less (Schier and Jensen, 1992). Cape (1993) reported that, when compared with other
species, red spruce seedlings appeared to be more sensitive to acid mist. From studies of
conifers and a review of the literature, Huttunen (1994) concluded that acidic precipitation
causes direct injury to tree foliage and indirect effects through the soil. The indirect effects of
acidic precipitation are discussed in Section 4.2.3.2.
Based on a review of the many studies in the literature involving field and controlled
laboratory experiments on crops, Cape (1993) drew a number of conclusions concerning the
direct effects of acidic precipitation on crops:
4-73
-------
• foliar injury and growth reduction occur below pH 3;
• allocation of photosynthate is altered, with increased shoot to root ratios;
• expanded and recently expanded leaves are most susceptible, and injury occurs first to
epidermal cells;
• leaf surface characteristics such as wettability, buffering capacity, and transport of material
across the leaf surface contribute to susceptibility and differ among species;
• data obtained from experiments in greenhouses or controlled environmental chambers
cannot be used to predict effects on plants grown in the field;
• quantitative data from experimental exposures cannot be extrapolated to field exposures
because of differences and fluctuations in concentrations, durations, and frequency of
exposure;
• there are large differences in response within species;
• timing of exposure in relation to phenology is of utmost importance;
• plants may be able to recover from or adapt to injurious exposures; and
• sequential exposure to acidic precipitation and gaseous pollutants is unlikely to be more
injurious than exposure to individual pollutants.
Studies by Chevone et al. (1986), Krupa and Legge (1986), and Blaschke (1990) differ
with the last conclusion of Cape listed above. Their studies indicate that interactions between
acidic deposition and gaseous pollutants do occur. Acidity affects plant responses to both O3 and
SO2. Chevone et al. (1986) observed increased visible injury on soybean and pinto bean when
acid aerosol exposure preceded O3 exposure; whereas linear decreases in dry root weight of
yellow poplar occurred as acidity increased with simultaneous exposures to O3 and simulated
acid rain. Krupa and Legge (1986) also noted increased visible injury to pinto bean plants when
aerosol exposure preceded O3 exposure. In none of the studies cited above did acid rain per se
produce significant growth changes. In contrast, Blaschke (1990) observed a decrease in
ectomycorrhizal frequency and short root distribution caused by acid rain exposure in
combination with either SO2 or O3.
Trace Elements. All but 10 of the 90 elements that comprise the inorganic fraction of the
soil occur at concentrations of < 0.1% (1000 |ig/g) and are termed "trace" elements or trace
metals. Trace metals with a density greater than 6 g/cm3, referred to as "heavy metals," are of
4-74
-------
particular interest because of their potential toxicity for plant and animals. Although some trace
metals are essential for vegetative growth or animal health, they are all toxic in large quantities.
Combustion processes produce metal chlorides that tend to be volatile and metal oxides that tend
to be nonvolatile in the vapor phase (McGowan et al., 1993). Most trace elements exist in the
atmosphere in particulate form as metal oxides (Ormrod, 1984). Aerosols containing trace
elements derive predominantly from industrial activities (Ormrod, 1984). Generally, only the
heavy metals cadmium, chromium, nickel, and mercury are released from stacks in the vapor
phase (McGowan et al., 1993). Concentrations of heavy metals in incinerator fly ash increase
with decreasing particle size.
Vegetational surfaces, especially the foliage, present a major reaction and filtration surface
to the atmosphere and act to accumulate particles deposited via wet and dry processes as
described in Section 4.2.2 (long, 1991; Youngs et al., 1993). The chemical constituents of
particles deposited on foliar surfaces may be taken up through the leaf surface. The greatest
particle loading is usually on the upper leaf surface where particles accumulate in the mid-vein,
center portion of the leaves. Additionally, fungal mycelia become particularly abundant on leaf
surfaces as the growing season progresses and can be found in intimate association with
deposited particles (Smith, 1990c).
Investigations of trace elements present along roadsides and in industrial and urban
environments indicate that impressive burdens of parti culate heavy metals can accumulate on
vegetative surfaces. Foliar uptake of available metals could result in metabolic effects in above-
ground tissues. Only a few metals, however, have been documented to cause direct
phytotoxicity in field conditions. Copper, Zn, and Ni toxicities have been observed most
frequently. Low solubility, however, limits foliar uptake and direct heavy metal toxicity
because trace metals must be brought into solution before they can enter into the leaves or bark
of vascular plants. In those instances when trace metals are absorbed, they are frequently bound
in leaf tissue and are lost when the leaf drops off (Hughes, 1981). Trace metals in mixtures may
interact to cause a different plant response when compared with a single element; however, there
has been little research on this aspect (Ormrod, 1984). In experiments using chambers,
Marchwiiiska and Kucharski (1987) studied the effects of SO2 alone and in combination with the
PM components Pb, Cd, Zn, Fe, Cu, and Mn obtained from a zinc smelter bag filter. The
combined effects of SO2 and PM further increased the reduction in yield of beans caused by SO2;
4-75
-------
whereas the combination, though severely injuring the foliage, produced little effect on carrots
and parsley roots except after long-term exposures (when there was a decrease in root weight).
Trace metal toxicity of lichens has been demonstrated in relatively few cases. Nash (1975)
documented zinc toxicity in the vicinity of a zinc smelter near Palmerton, PA. The richness and
abundance of lichen species were reduced by -90% in lichen communities at Lehigh Water Gap
near the zinc smelter when compared with those at Delaware Water Gap. Zinc, Cd, and SO2
were present in concentrations toxic to some species near the smelter; however, toxic zinc
concentrations were detected farther away than the detectable limits of SO2 (Nash, 1975).
Experimental data suggest that lichen tolerance to Zn and Cd falls between 200 and 600 ppm
(Nash, 1975).
Though there has been no direct evidence of a physiological association between tree
injury and exposure to metals, heavy metals have been implicated because their deposition
pattern is correlated with forest decline. The role of heavy metals has been assessed using
phytochelatin measurements. Phytochelatins are intracellular metal-binding peptides that act as
specific indicators of metal stress. Because they are produced by plants as a response to
sublethal concentrations of heavy metals, they can indicate that heavy metals play a role in forest
decline (Gawel et al., 1996). Concentrations of heavy metals increased with altitude, as did
forest decline; and increased concentrations of heavy metals were found across the study region
that showed increased levels of forest injury.
Phytochelatin concentrations were measured in red spruce and balsam fir {Abies balsamed)
needles throughout the 1993 growing season at 1000 m on Whiteface Mountain in New York.
Mean foliar concentrations in red spruce were consistently higher than in balsam fir from June
until August, with the greatest and most significant difference occurring at the peak of the
growing season in mid-July. In July, the phytochelatin concentrations were significantly higher
than at any other time measured. Balsam fir also exhibited this peak, but maintained a
consistently low level throughout the season. Both the number of dead red spruce trees and
phytochelatin concentrations increased sharply with elevation (Gawel et al., 1996). The
relationship between heavy metals and the decline of forests in northeastern United States was
further tested by sampling red spruce stands showing varying degrees of decline at 1000 m on
nine mountains spanning New Hampshire, Vermont, and New York. The collected samples
indicated a systematic and significant increase in phytochelatin concentrations associated with
4-76
-------
the extent of tree injury. The highest phytochelatin concentrations were measured during 1994
from sites most severely affected by forest decline in the Green Mountains in Vermont and the
Adirondack Mountains in New York. These data strongly imply that metal stress causes tree
injury and contributes to forest decline in the northeastern United States (Gawel et al., 1996).
One potential direct effect of heavy metals is on the activity of microorganisms and
arthropods resident on and in the leaf-surface ecosystem. The fungi and bacteria living on and in
the surfaces of leaves play an important role in the microbial succession that prepares leaves for
decay and litter decomposition after their fall (U.S. Environmental Protection Agency, 1996b).
Numerous fungi were consistently isolated from foliar surfaces at various crown positions from
London plane trees (Platanas hispanica) growing in roadside environments in New Haven, CT.
Those existing primarily as saprophytes included Aureobasidium pullulans, Chaetomium sp.,
Cladosporium sp., Epicoccum sp., and Philaphora verrucosa. Those existing primarily as
parasites included Gnomonia platani, Pestalotioposis sp., and Pleurophomella sp. The
following cations were tested in vitro for their ability to influence the growth of these fungi:
Cd, Cu, Mg, Al, Cr, Ni, Fe, Fb, Na, and Zn. Results indicated variable fungal response with no
correlation between saprophytic or parasitic activity and sensitivity to heavy metals. Both linear
extension and dry weight data indicated that the saprophytic Chaetomum sp. was very sensitive
to numerous metals. Aureobasidium pullulans, Epicoccum sp., and especially/1, verrucosa, on
the other hand, appeared to be much more tolerant. Of the parasites, G. platani appeared to be
more tolerant than Pestalotiopsis sp. and Pleurophomella sp. Metals producing the broadest
spectrum growth suppression were Fe, Al, Ni, Zn, Mg, and Pb (Smith and Staskawicz, 1977;
Smith, 1990c). These in vitro studies employed soluble compounds containing heavy metals.
Trace metals probably occur naturally on leaf surfaces as low-solubility oxides, halides, sulfates,
sulfides, or phosphates (Clevenger et al., 1991; Koslow et al., 1977). In the event of sufficient
solubility and dose, however, changes in microbial community structure on leaf surfaces because
of heavy metal accumulation are possible.
Organic Compounds. Volatile organic compounds in the atmosphere are partitioned
between the gas and particle phases, depending on the liquid-phase vapor pressure at the ambient
atmospheric temperature, the surface area of the particles per unit volume of air, the nature of the
particles and of the chemical being adsorbed; and they can be removed from the atmosphere by
4-77
-------
wet and dry deposition (McLachlan, 1996a). Materials as diverse as DDT, poly chlorinated
biphenyls (PCBs), and polynuclear aromatic hydrocarbons (PAHs) are being deposited from the
atmosphere on rural, as well as urban, landscapes (Kylin et al., 1994). Motor vehicles emit
particles to the atmosphere from several sources in addition to the tailpipe. Rogge et al. (1993b)
inventoried the organic contaminants associated with fine particles (diameter < 2.0 jim) in road
dust, brake-lining-wear particles, and tire-tread debris. More than 100 organic compounds were
identified in these samples, including N-alkanols, benzoic acids, benzaldehydes, polyalkylene
glycol ethers, PAHs, oxy-PAH, steranes, hopanes, natural resins, and other compound classes.
A large number of PAHs, ranging from naphthalene (C10H8) to 5- and 6-ring and higher PAHs,
their alkyl-substituted analogues, and their oxygen- and nitrogen-containing derivatives are
emitted from motor vehicle sources (Seinfeld, 1989).
Plants may be used as environmental monitors to compare the deposition of PAH,
persistent organic pollutants (POPs), or semivolatile organic components (SOCs) between sites
(e.g., urban versus rural) (Wagrowski and Kites, 1997; Ockenden et al., 1998; McLachlan,
1999). Vegetation can be used qualitatively to indicate organic pollutant levels as long as the
mechanism of accumulation is considered. The substance may enter the plant via the roots or, as
noted above, be deposited as a particle onto the waxy cuticle of leaves or be taken up through the
stomata. The pathways are a function of the chemical and physical properties of the pollutant,
e.g., its lipophilicity, water solubility, vapor pressure (which controls the vapor-particle
partitioning) and Henry's law constant; environmental conditions such as ambient temperature
and the organic content of the soil; and the plant species, which controls the surface area and
lipids available for accumulation (Simonich and Kites, 1995). Ockenden et al. (1998) have
observed that, for lipophilic POPs, atmospheric transfer to plant has been the main avenue of
accumulation. Plants can differentially accumulate POPs. Results have shown differences
between species with higher concentrations in the lichen (Hypogymniaphysiodes) than in Scots
pine (Pinus sylvestris) needles. Even plants of the same species, because they have different
growth rates and different lipid contents (depending on the habitat in which they are growing),
have different rates of sequestering pollutants. These facts confound data interpretations and
must be taken into account when considering their use as passive samplers.
Vegetation itself is an important source of hydrocarbon aerosols. Terpenes, particularly
a-pinene, p-pinene, and limonene, released from tree foliage may react in the atmosphere to
4-78
-------
form submicron particles. These naturally generated organic particles contribute significantly to
the blue haze aerosols formed naturally over forested areas (Geron et al., 2000).
The low water solubility with high lipoaffinity of many of these organic xenobiotics
strongly control their interaction with the vegetative components of natural ecosystems. The
cuticles of foliar surfaces are covered with a wax layer that helps protect plants from moisture
and short-wave radiation stress. This epicuticular wax, consisting mainly of long-chain esters,
polyesters, and paraffins, has been demonstrated to accumulate lipophilic compounds. Organic
air contaminants in the particulate or vapor phase are absorbed to, and accumulate in, the
epicuticular wax of vegetative surfaces (Gaggi et al., 1985; Kylin et al., 1994). Direct uptake of
organic contaminants through the cuticle or the vapor-phase uptake through the stomata are
characterized poorly for most trace organics. The phytotoxicity and toxicity of organic
contaminants to soil microorganisms is not well studied (Foster, 1991).
4.2.3.2 Ecosystem Response to Stresses
The ecosystem response to stress begins within the biotic condition attribute at the
population level with changes in patterns resulting from the response of sensitive individual
plants or animals. The ecosystem response to pollutant deposition is a direct function of the
ecosystem's ability to ameliorate resulting change (Strickland et al., 1993). Plant responses,
changes in both structural and compositional patterns, and functional ecological processes must
be scaled in both time and space and be propagated from the individual to the more complex
levels of community interaction to produce observable changes in an ecosystem (see
Figure 4-14). Among ecosystem biota, at least three levels of biological interaction are
involved: (1) the individual plant and its environment, (2) the population and its environment,
and (3) the biological community composed of many species and its environment (Billings,
1978). Individual organisms within a population vary in their ability to withstand the stress of
environmental change. The response of individual organisms within a population is based on
their genetic constitution (genotype), stage of growth at time of exposure to stress, and the
microhabitat in which they are growing (Levin, 1998). The range within which these organisms
can exist and function determines the ability of the population to survive. Those able to cope
with the stresses survive and reproduce. Competition among the different species results in
succession (community change over time) and, ultimately, produces ecosystems composed of
4-79
-------
Level of
Organization
Leaf
(cm2)
Branch
(cm*)
Tree
(m2)
Stand
(ha)
Minute
Reaction Tir
Day Year
te
• fr
\m -
\s
W\ 1
• J
• 1
-^
ne
Decade
1
2
3
»-5
•- 8
f 9
• 1 xN
Century
10
11
12
13
^"14
^6
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 alteration 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 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
Leaf Level
Branch Leve
Impacts Within a Level of Organization
Carbon exchange-1 Tree Level
Carbon pools-2
Needle number and size-3
Needle retention/abscission-4
Carbon allocation-5 Stand Level
Branch growth-6
Branch morphology-7
Branch vigor-8
Branch retention-9
Height and diameter growth-1 0
Crown shape and size-11
Tree vigor-12
Mortality-13
Productivity-14
Mortality-15
Species composition-1 6
Evaluating Interactions Between Different Levels of Organization
The diagonal arrow indicates the interaction between any two levels of organization.
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 4-14. 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).
4-80
-------
populations of plant species that have the capability to tolerate the stresses (Rapport and
Whitford, 1999; Guderian, 1985).
The number of species in a community usually increases during succession in unpolluted
atmospheres. Productivity, biomass, community height, and structural complexity increase.
Severe stresses, on the other hand, divert energy from growth and reproduction to maintenance
and return succession to an earlier stage (Waring and Schlesinger, 1985). Ecosystems are
subject to natural periodic stresses, such as drought, flooding, fire, and attacks by biotic
pathogens (e.g., fungi, insects). Extremely severe natural perturbations return succession to an
earlier stage; reduce ecosystem structure and functions (i.e., produce a scarcity of life forms and
extinguish symbiotic interactions); disrupt the plant processes of photosynthesis and nutrient
uptake, carbon allocation, and transformation that are directly related to energy flow and nutrient
cycling; shorten food chains; and reduce the total nutrient inventory (Odum, 1993). This
transformation, however, sets the stage for recovery that permits the perturbed ecosystem to
adapt to changing environments (Holling, 1986). Therefore, these perturbations are seldom
more than a temporary setback, and recovery can be rapid (Odum, 1969).
In contrast, anthropogenic stresses usually produce severe, debilitating effects. Severely
stressed ecosystems do not recover readily, but may be further degraded (Odum, 1969; Rapport
and Whitford, 1999). Anthropogenic stresses can be classified into four main groups:
(1) physical restructuring (e.g., changes resulting from land use), (2) introduction of exotic
species, (3) over harvesting, and (4) discharge of toxic substances into the atmosphere, onto
land, and into water. Ecosystems usually lack the capacity to adapt to the above stresses and
maintain their normal structure and functions unless the stressor is removed (Rapport and
Whitford, 1999). These stresses result in a process of ecosystem degradation marked by a
decrease in biodiversity, reduced primary and secondary production, and a lower capacity to
recover and return to its original state. In addition, there is an increased prevalence of disease,
reduced nutrient cycling, increased dominance of exotic species, and increased dominance by
smaller, short-lived opportunistic species (Odum, 1985; Rapport and Whitford, 1999).
Discharge of toxic substances into the atmosphere, onto land, and into water can cause acute and
chronic stresses; and, once the stress is removed, a process of succession begins that can
ultimately return the ecosystem to a semblance of its former structure. Air pollution stress, if
acute, is usually short-term and the effects soon visible. Chronic stresses, on the other hand, are
4-81
-------
long-term stresses whose effects occur at different levels of ecosystem organization and appear
only after long-term exposures, as in the case of acidic deposition in the northeast or ozone in
California (Shortle and Bondietti, 1992; U.S. Environmental Protection Agency, 1996b).
The possible effects of air pollutants on ecosystems have been categorized by Guderian
(1977) as follows:
(1) accumulation of pollutants in the plant and other ecosystem components
(such as soil and surface- and groundwater),
(2) damage to consumers as a result of pollutant accumulation,
(3) changes in species diversity because of shifts in competition,
(4) disruption of biogeochemical cycles,
(5) disruption of stability and reduction in the ability of self-regulation,
(6) breakdown of stands and associations, and
(7) expansion of denuded zones.
How changes in these functions can result from PM deposition and influence ecosystems is
discussed in the following text. It should be remembered that, although the effects of PM are
being emphasized, the vegetational components of ecosystems also are responding to multiple
stressors from multiple sources.
Response to Direct Effects of Particulate Matter
In the previous section, it was noted that PM affects patterns in the EEA biotic condition
and processes in the chemical/physical and ecological processes categories. The presence of
PM in the atmosphere may affect vegetation directly, following physical contact with the foliar
surface (Section 4.2), but in most cases, the more significant effects are indirect. These effects
may be mediated by suspended PM (i.e., through effects on radiation and climate) and by
particles that pass through the vegetative canopies to the soil.
The majority of studies dealing with the direct effects of particulate dust and trace metals
on vegetation were focused on responses of individual plant species and were conducted in the
laboratory or in controlled environments (Saunders and Godzik, 1986). A few have considered
the effects of particles on populations, communities, and ecosystems. Most of these focused on
ecosystems in industrialized areas that are heavily polluted by deposits of both chemically inert
4-82
-------
and active dusts. Effects can result from direct deposition or indirectly from deposition onto the
soil. Reductions in growth, yield, flowering, and reproductive processes of plants from PM
deposition have been reported (Saunders and Godzik, 1986). Sensitivities of individual species
have been associated with changes in composition and structural patterns of natural ecosystems.
Evidence from studies of effects of PM deposition, specifically chemically inert and active
dusts indicates that, within a population, plants exhibit a wide range of sensitivity which is the
basis for the natural selection of tolerant individuals. Rapid evolution of certain populations of
tolerant species at sites with heavy trace element and nitrate deposition has been observed
(Saunders and Godzik, 1986). Tolerant individuals present in low frequencies in populations
when growing in unpolluted areas have been selected for tolerance at both the seedling and adult
stages when exposed to trace metal or nitrate deposition (Ormrod, 1984; U.S. Environmental
Protection Agency, 1993). Chronic pollutant injury to a forest community may result in the loss
of sensitive species, loss of tree canopy, and maintenance of a residual cover of pollutant-
tolerant herbs or shrubs that are recognized as success!onal species (Smith, 1974) (see
Table 4-13). These changes in forest patterns result from altered ecological processes.
Responses of ecosystems to stresses (unless severe or catastrophic) are difficult to
determine because the changes are subtle (Garner, 1991). This is particularly true of responses
to particles. Changes in the soil may not be observed until accumulation of the pollutant has
occurred for 10 or more years, except in the severely polluted areas around heavily industrialized
point sources (Saunders and Godzik, 1986). In addition, the presence of other cooccurring
pollutants makes it difficult to attribute the effects to PM alone. In other words, the potential for
alteration of ecosystem function and structure exists but is difficult to quantify, especially when
there are other pollutants present in the ambient air that may produce additive or synergistic
responses even though PM concentrations may not be elevated.
Physical Effects
The direct effects of limestone dust on plants and ecosystems has been known for many
years. Changes have been observed in both ecosystem patterns and processes. Long-term
changes in the structure and composition of the seedling-shrub and sapling strata of an
experimental site near limestone quarries and processing plants in Giles County in southwestern
Virginia were reported by Brandt and Rhoades (1972, 1973). Dominant trees in the control area,
4-83
-------
TABLE 4-13. 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).
a part of the oak-chestnut association of the eastern deciduous forests of eastern North America,
were chestnut oak (Quercusprinus), red oak (Q. rubra), and red maple (Acer rubrum).
An abundance of uniformly distributed saplings and seedlings were visible under the tree
canopy, and herbs appeared in localized areas in canopy openings. Chestnut oak dominated the
area, and the larger trees were 60 to 80 years old. The dusty site was dominated by white oak
(Q. alba); whereas red oak and tulip poplars (Liriodendron tulipifera) were subcodominants.
The largest trees were 100 years old and had necrotic leaves, peeling bark, and appeared to be in
generally poor condition except for tulip poplars (which thrived in localized areas). The site
contained a tangled growth of seedlings and shrubs, a few saplings, and a prevalence of green
briar (Smilax spp.) and grape (Vitis spp.). The sapling strata in the area was represented by red
maple, hickory (Carya spp.), dogwood (Cornus floridd), and hop-hornbeam (Ostrya virginiand).
Saplings of none of the leading dominant trees were of importance in this stratum. The most
4-84
-------
obvious form of vegetation in the seedling-shrub stratum, because of their tangled appearance,
were dogwood, hop-hornbeam, redbud (Cercis canadensis), and sugar maple (Acer saccarum).
Crust formation reduced photosynthesis, induced premature leaf fall and destruction of leaf
tissues, inhibited growth of new tissue, and reduced the formation of carbohydrate needed for
normal growth and storage (Brandt and Rhoades, 1973). The authors (Brandt and Rhoades,
1972), citing Odum (1969), also stated that one result of the accumulation of toxic pollutants in
the biosphere as the result of human activities is the simplification of both plant and animal
communities. In plant communities, structure is determined by sampling various strata within
the community. Each stratum comprises a particular life form (e.g., herbs, seedlings, saplings,
trees). Dust accumulation favored growth of some species and limited others. For example,
sugar maple was more abundant in all strata of the dusty site when compared with the control
site where it was present only as a seedling. The growth of tulip poplar, dogwood, hop-
hornbeam, black haw (Viburnum primifolium), and redbud (C. canadensis) appeared to be
favored by the dust. Growth of conifers and acidophiles such as rhododendron (Rhododendron
maximum)., however, was limited. Although dust accumulation began in 1945, the heaviest
accumulation occurred between 1967 and 1972 during the time of the study.
Changes in community composition were associated closely with changes in the growth of
the dominant trees. A decrease in density of seedlings and saplings and in mean basal area, as
well as lateral growth of red maple, chestnut oak, and red oak, occurred in all strata. On the
other hand, all of these characteristics increased in the tulip poplar, which was a subordinate
species before dust accumulation began but had assumed dominance by the time of the study.
Reduction in growth of the dominant trees had apparently given the tulip poplar a competitive
advantage because of its ability to tolerate dust. Changes in soil alkalinity occurred because of
the heavy deposition of limestone dust; however, the facilities necessary for critical analysis of
the soils were not available. From the foregoing, it is obvious that PM physical effects in the
vicinity of limestone quarries and processing plants can affect ecosystems.
Changes in ecosystem structure resulting from exposures to sea salt were cited previously
(Section 4.3.1.1). The dominance of live oak (Quercus virginiand) as a practically pure stand on
Smith Island (Bald Head), NC and along the eastern and southern coast of North Carolina has
been explained as due to its tolerance to salt spray. The absence of more inland species is
4-85
-------
attributed to their intolerance to salt spray. Wells (1939) termed the long-term stabilization of
live oak as "salt spray climax," a new type of climax.
Acidic Deposition
The major effects of acidic deposition occur through the soil (National Science Technology
Council, 1998) and are discussed below under indirect effects. However, included among the
direct responses of forest trees to acidic deposition are increased leaching of nutrients from
foliage; accelerated weathering of leaf cuticular surfaces; increased permeability of leaf surfaces
to toxic materials, water, and disease agents; and altered reproductive processes (Cape, 1993;
Altshuller and Linthurst, 1984).
Trace Elements
Possible direct responses of trace elements on vegetation result from their deposition and
residence on the phyllosphere (i.e., foliar surfaces). Fungi and other microorganisms living on
the leaves of trees and other vegetation play an important role in leaf decomposition after
litterfall (Miller and McBride, 1999; Jensen, 1974; Millar, 1974). Possible effects of heavy
metals on nutrient cycling and their effects on leaf microflora appear not to have been studied.
A trace metal must be brought into solution before it can enter into the leaves or bark of
vascular plants. Low solubility limits entry. In those instances when trace metals are absorbed,
they frequently are bound in the leaf tissue and then are lost from the plant when the leaf drops
off (Hughes, 1981) and transferred to the litter layer where they can affect litter decomposition,
an important source of soil nutrients. Changes in litter decomposition processes influence
nutrient cycling in the soil and limit the supply of essential nutrients. Both Cotrufo et al. (1995)
and Niklinska et al. (1998) point out that heavy metals affect forest litter decomposition.
Cotrufo et al. (1995) observed that decomposition of oak leaves containing Fe, Zn, Cu, Cr, Ni,
and Pb was influenced strongly during the early stages by metal contamination. Fungal mycelia
were significantly less abundant in litter and soil in contaminated sites when compared with
control sites. Niklinska et al. (1998) stated that toxic effects of heavy metals on soil respiration
rate have been reported by many scientists, and that, in polluted environments, this results in
accumulation of undecomposed organic matter. However, they state that results of experiments
should identify the most important "natural" factors affecting soil/litter sensitivity because the
4-86
-------
effects of heavy metals on respiration rates are dose-dependent on the dose of heavy metals, the
type of litter, types of metals deposited, and the storage time before respiration tests are made.
Trace metals, particularly heavy metals (e.g., Cd, Cu, Pb, Cr, Hg, Ni, Zn) have the greatest
potential for influencing forest growth (Smith, 1991). Experimental data indicate that the
broadest spectrum of growth suppression of foliar microflora resulted from Fe, Al, and Zn.
These three metals also inhibited fungal spore formation, as did Cd, Cr, Mg, and Ni (see Smith,
1990c). In the field, the greatest injury occurs from pollution near mining, smelting, and other
industrial sources (Ormrod, 1984). Direct metal phytotoxicity can occur only if the metal can
move from the surface into the leaf or directly from the soil into the root.
Organic Compounds
Primary and secondary organic compounds formed in the atmosphere, the effects of some
of which are discussed below, have been variably referred to under the following terms: toxic
substances, hazardous air pollutants (HAPs), air toxics, semivolatile organic compounds
(SVOCs), pesticides, and persistent organic pollutants (POPs). Again, it should be remembered
that chemical substances denoted by such headings are not criteria air pollutants controlled by
the National Ambient Air Quality Standards under Section 109 of the Clean Air Act (CAA),
but rather are controlled under Sect. 112, Hazardous Air Pollutants (U.S. Code, 1991). Their
possible effects on humans and ecosystems are discussed in a number of government documents
and in many other publications. They are mentioned here because many of the chemical
compounds are partitioned between gas and particle phases in the atmosphere. As particles, they
can become airborne, be distributed over a wide area, and affect remote ecosystems. Some of
the chemical compounds are of concern because they may reach toxic levels in food chains of
both animals and humans; whereas others tend to decrease or maintain the same toxicity as they
move through the food chain. Some examples of movement through food chains are provided
below.
Many chemical compounds from a variety of anthropogenic sources are released into the
ambient air (see Section 4.2.1). In the atmosphere, the emitted compounds initially go through a
mixing process, and the airborne particles then are distributed over a wide area and ultimately
deposited on ecosystem components. Atmospheric deposition of poly chlorinated dibenzo-p-
dioxins and dibenzofurans (PCDD/DFs), as an example, can be divided into three different
4-87
-------
forms: (1) dry gaseous, (2) dry particle-bound, and (3) wet deposition. Dry particle-bound
deposition occurs when the PM containing the pollutant is deposited on the plant surface;
whereas wet deposition ranges from hail to rain to fog and dewfall (McLachlan, 1996b).
Human exposure to PCDD/DFs has been demonstrated to be caused almost exclusively
by the ingestion of animal fat from fish, meat, and dairy products. Almost half of all human
exposure to PCDD/DFs is caused by consumption of beef and dairy products (McLachlan,
1996b). Cattle obtain most of their PCDD/DFs though grass. Therefore, the grass - cattle
- milk/beef pathway is critical for human exposure. It has been shown that root uptake/
translocation is an insignificant pathway of PCDD/DFs compared to absorption via aerial plant
parts. Wet and dry particle deposition are the most important for the accumulation of the higher
chlorinated congeners in vegetation. The persistence of PCDD/DFs in plants has not been
investigated extensively; however, biodegradation probably does not occur in that these
compounds are found primarily in the lipophilic cuticle and are very resistant to microbial
degradation (McLachlan, 1996b). Feed contaminated with soil containing the pollutant can be
another source of exposure of beef and dairy cattle, as well as chickens. The PCDD/DFs levels
are near a steady state in milk cows and laying hens; however, animals raised for meat
production (such as beef cattle and pigs) may accumulate large amounts. Beef cattle and pigs
cannot excrete the contaminants in a lipid-rich matrix such as milk or eggs. Thus, all of the
PCDD/DFs ingested are stored in the body. In agricultural food chains, there is a biodilution of
PCDD/DFs, with the fugacity decreasing by up to three orders of magnitude between the air and
cows milk (McLachlan, 1996b). Fiirst et al. (1993), based on surveys to determine the factors
that influence the presence of PCDD/DF in cows milk, earlier concluded that regardless of which
pathway, soil - grass - cow or air - grass - cow, it was the congener of the chemical that was
most important.
Chlorinated persistent organic pollutants (POPs), such as PCBs, PCDFs, and PCDDs, can
be transported as particles through the atmosphere from industrial and agricultural sources; be
brought down via wet and dry deposition in remote regions, such as the Arctic; and have been
detected in all levels of the Arctic food chain (Oehme et al., 1995). High concentrations of PCB
(1 to 10 ppm) were found in seals, but the concentrations increased to 10 to 100 ppm in polar
bears. The polar bear is the top predator in the Arctic and feeds preferentially on ringed seals
and, to a lesser extent, on other seal species. Bioconcentration factors of organochlorines in the
4-S
-------
Arctic food web, reaching 107 for fish and seals, are biomagnified in polar bears (Oehme et al.,
1995). Poly chlorinated dibenzo-p-dioxins and poly chlorinated dibenzofurans have also been
found in seals (Oehme et al., 1995). Milk taken from anesthetized polar bears was also found to
contain PCDD/DF. Very little is known regarding the intake of milk by polar bear cubs.
However, estimates of the intake of milk containing detectable levels of PCDD/DF and PCB and
the additional consumption of seal blubber confirm that these pollutants are passed on to the next
generation (Oehme et al., 1995).
Section 112 of the CAA provides the legislative basis for U.S. hazardous air pollutant
(HAP) programs. In response to mounting evidence that air pollution contributes to water
pollution, Congress included Section 112m (Atmospheric Deposition to Great Lakes and Coastal
Waters) in the 1990 CAA Amendments that direct the EPA to establish a research program on
atmospheric deposition of HAPS to the "Great Waters."
Actions taken by EPA and others to evaluate and control sources of Great Waters
pollutants of concern appear to have positively affected trends in pollutant concentrations
measured in air, sediment, and biota. Details concerning these effects may be found in
Deposition of Air Pollutants to the Great Waters, Third Report to Congress (U.S. Environmental
Protection Agency, 2000a). The Third Report (EPA-453/R-00-005, June 2000), like the First
and Second Reports to Congress, focuses on 15 pollutants of concern, including pesticides, metal
compounds, chlorinated organic compounds, and nitrogen compounds. The new scientific
information in the Third Report supports and builds on three broad conclusions presented in the
previous two EPA Reports to Congress:
(1) Atmospheric deposition from human activities can be a significant contributor of toxic
chemicals and nitrogen compounds to the Great Waters. The relative importance of
atmospheric loading for a particular chemical in a water body depends on many factors
(e.g., characteristics of the water body, properties of the chemical, and the kind and
amount of atmospheric deposition versus water discharges).
(2) A plausible link exists between emissions of toxic pollutants of concern into the air
above the Great Waters; the deposition of these pollutants (and their transformation
products); and the concentrations of these pollutants found in the water, sediments,
and biota, especially fish and shellfish. For mercury, fate and transport modeling and
exposure assessments predict that the anthropogenic contribution to the total amount
of methylmercury in fish is, in part, the result of anthropogenic mercury releases
from industrial and combustion sources increasing mercury body burdens (i.e.,
concentrations) in fish. Also, the consumption offish is the dominant pathway of
4-89
-------
exposure to methylmercury for fish-consuming humans and wildlife. However, what is
known about each stage of this process varies with each pollutant (for instance, the
chemical species of the emissions and its transformation in the atmosphere).
(3) Airborne emissions from local as well as distant sources, from both within and outside
the United States, contribute pollutant loadings to waters through atmospheric
deposition. Determining the relative roles of particular sources — local, regional,
national, and possibly global, as well as anthropogenic, natural, and reemission of
pollutants — contributing to specific water bodies is complex, requiring careful
monitoring, atmospheric modeling, and other analytical techniques.
Response to Indirect Effects of Particulate Matter
The presence of PM in the atmosphere directly affects vegetation following physical
contact with foliar surfaces (as discussed above), but in many cases the more significant effects
are indirect. These effects may be mediated by suspended PM (i.e., through effects on radiation
and climate) and by particles that pass through vegetative canopies to reach the soil. Effects
mediated through the atmosphere are considered briefly below and in greater detail later, under
Section 4.5.
The major indirect plant responses are chiefly soil-mediated and depend primarily on the
chemical composition of the individual stressors deposited in PM. The chemical stressors must
be bioavailable in order to produce an effect. The effects of exposures may result in changes in
biota patterns and in chemical/physical soil conditions that affect ecological processes, such as
nutrient cycling and uptake by plants.
The soil environment (composed of mineral and organic matter, water, air, and a vast array
of bacteria, fungi, algae, actinomycetes, protozoa, nematodes, and arthropods) is one of the most
dynamic sites of biological interactions in nature (Wall and Moore, 1999; Alexander, 1977).
The quantity of organisms in soils varies by locality. Bacteria and fungi are usually most
abundant in the rhizosphere, the soil around plant roots that all mineral nutrients must pass
through. Bacteria and fungi benefit from the nutrients in the root exudates (chiefly sugars) in the
soil and, in turn, they play an essential role by making mineral nutrients available for plant
uptake (Wall and Moore, 1999; Rovira and Davey, 1974). Their activities create chemical and
biological changes in the rhizosphere by decomposing organic matter and making inorganic
minerals available for plant uptake. Bacteria are essential in the nitrogen and sulfur cycles and
make these elements available for plant uptake and growth (see Section 4.3.3). Fungi are
4-90
-------
directly essential to plant growth. Attracted to their roots by the exudates, they develop
mycorrhizae in a symbiotic fungus-plant relationship, that is integral in the uptake of the mineral
nutrients (Allen, 1991). The impact in ecosystems of PM, particularly nitrates, sulfates, and
metals, is determined by their effects on the growth of the bacteria involved in nutrient cycling
and the mycorrhizal fungi involved in plant nutrient uptake.
Atmospheric Turbidity: Effects on Direct Beam and Photosynthetically Active
Radiation Photosynthetic processes underlie the contribution of vegetative surfaces to nutrient
and energy cycling. The characteristics and net receipts of environmental radiation determine
the rates of both photosynthesis and the heat-driven process of water cycling. Atmospheric
turbidity due to particulate loading can substantially alter the characteristics and net receipts of
solar radiation. One measure of atmospheric turbidity, Linke's turbidity factor, T, can be
derived as a direct function of light extinction by solid particles. It is defined as the ratio of the
total extinction coefficient and the extinction due exclusively to gases:
T = a/ag = 1 + waw /ag + sa5/ag (4-6)
where s and w are the relative concentrations of dust and water vapor in the atmosphere, and os
and ow are the wavelength-dependent scattering coefficients for solid, dry particles and water
vapor, respectively. The scattering coefficients are in units of inverse distance, such as km"1
(Rosenberg et al., 1983). According to this expression, a clean atmosphere would have a
turbidity value of 1. Given that turbidity and visibility are both functions of light scattering, the
trends in, and physical processes underlying, reduced visibility discussed in Section 4.3 are
directly relevant to the discussion of radiative effects on vegetation due to PM.
Turbidity, as defined above, describes the degree of scattering occurring in the atmosphere
due to particles and gases. Total, particle-based extinction, however, is the sum of both
scattering and absorption. Absorption of short-wavelength solar radiation reduces the amount of
radiation reaching the Earth's surface and leads to atmospheric heating. If the absorbing
particles re-radiate in the infrared range, some of this energy is lost as long-wave re-radiation to
space. The balance of this energy is captured at the surface as down-welling infrared radiation.
Canopy temperature and transpirational water use by vegetation are particularly sensitive to
4-91
-------
long-wave, infrared radiation. Atmospheric heating by particles reduces vertical temperature
gradients, potentially reducing the intensity of atmospheric turbulent mixing. The magnitude of
such potential effects on turbulent transport within canopies remains unknown although the
damping of eddy transport may inhibit canopy gas exchange. Suppressed tropospheric mixing
could also intensify local temperature inversions and increase the severity of pollution episodes
(Pueschel, 1993) with direct inhibitory effects on photosynthetic processes.
Atmospheric turbidity increases the intensity of diffuse (sky) radiation (Hoyt, 1978).
In a clear atmosphere, diffuse radiation may be on the order of 10% of total solar radiation
(Choudhury, 1987). However, in highly turbid, humid conditions, this fraction may increase,
to as much as 100% of the incident solar intensity in extreme cases. The direct-radiation diffuse
ratio is highest at solar noon and lowest near dawn or dusk when the path length through the
atmosphere is longest. The wavelength dependence of particle scattering induces an enrichment
of PAR with respect to total or direct beam radiation. The influence of this enrichment on plant
life is discussed in the next section.
Aerosols produced by incomplete combustion, from forest fires to specifically
anthropogenic processes such as diesel fuel combustion, contain significant fractions of black
carbon which absorbs across the solar and terrestrial radiation spectra. The presence of
absorbing aerosols reduces the ratio of photosynthetically active radiation to total radiation
received at the surface, potentially reducing photosynthetic water uptake efficiency. The net
effect of aerosol absorption on the surface depends on the relative magnitudes of the particulate
absorption coefficients in the visible and infrared area and on the albedo of the Earth's surface.
The greater effect of particulate loading on visibility and turbidity and, therefore, radiation
receipts by the biosphere, is due to scattering. Non-absorbing, scattering aerosols raise the
overall albedo of the atmosphere and reduce the amount of radiation reaching the surface by the
amount reflected or scattered back into space. Analysis of data collected by a global network of
thermopile pyranometers operated by the World Meteorological Organization (WHO) show a
50-year global trend of a 2.7% per decade reduction in the amount of solar radiation reaching the
Earth's surface. This has been associated with an increasing global albedo caused by an
increasing abundance of atmospheric particles. By evaluating the WMO data set with four
different approaches to the statistical analyses, Stanhill and Cohen (2001) have estimated that
average global solar radiation receipts have declined by 20 W nT2 since 1958. Examples of
4-92
-------
individual measurement sites include Barrow, Alaska (71 °N), where the average solar irradiance
from 1963 to 1993 was 100.9 W nT2 and the estimated linear trend was -0.23% per year; and
Jerusalem, Israel (32 °N), where the average solar irradiance from 1954 to 1994 was
244.2 W nT2 and the estimated linear trend was -0.37% per year. This phenomenon has been
labeled "global dimming."
Increased Turbidity and Altered Radiative Flux: Effects on Vegetative Processes.
In a detailed canopy-response model (e.g., Choudhury, 1987), radiation is considered in its direct
and diffuse components. Foliar interception by canopy elements is considered for both up- and
down-welling radiation (a two-stream approximation). In this case, the effect of atmospheric
PM on turbidity influences canopy processes both by radiation attenuation and by influencing
the efficiency of radiation interception throughout the canopy through conversion of direct to
diffuse radiation (Hoyt, 1978). Diffuse radiation is more uniformly distributed throughout the
canopy and increases canopy photosynthetic productivity by distributing radiation to lower
leaves. The treatment of down-welling direct-beam radiation in the two-stream approach
remains an elaboration of the simplified Beer's Law analogy with solar angle, leaf area
distribution, and orientation individually parameterized (Choudhury, 1987). Diffuse down-
welling radiation is a function of diffuse and direct radiation at the top of the canopy and
penetration within the canopy according to cumulative leaf area density and foliage orientation.
Diffuse up-welling radiation results from scattering and reflectance of both direct and diffuse
down-welling radiation within the canopy and by the soil.
Rochette et al. (1996) conducted simultaneous measurements of radiation and water use
efficiencies by maize and found that, in the absence of water stress and with adequate
fertilization, 90% of all variation in crop net photosynthesis (P-n) could be explained by
variations in PAR. Alternatively, an evaluation of the available experimental literature and
statistics on crop yields by Stanhill and Cohen (2001) indicate that plant productivity is more
affected by changes in evapotranspiration induced by changes in the amount of solar radiation
plants receive than by changes in the amount of PAR plants receive.
The enrichment in PAR present in diffuse radiation appears, however, to offset a portion of
the effect of an increased atmospheric albedo due to atmospheric particles. An observational and
theoretical study by Bange et al. (1997) of the level of radiation use efficiency (RUE) of
4-93
-------
sunflowers indicated a degree of compensation for reduced incident radiation by a proportional
increase in diffuse radiation. Variables measured by Bange et al. (1997) included biomass,
phenology, leaf area, canopy light extinction, grain size, and harvest index. Crops subject to
reduced direct beam/increased diffuse radiation produced biomass, phenology, leaf area and
canopy light extinction at leaves similar to unshaded crops but yielded smaller grains and a
lower harvest index. RUE was also seen to improve for soybeans and maize with a proportional
increase in diffuse radiation with respect to direct beam (Sinclair et al., 1992; Healey et al.,
1998) although the effect on harvest index was not indicated. Gu et al. (2002) compared the
relative efficiencies of canopy photosynthesis to diffuse and direct PAR for a Scots pine forest,
an aspen forest, a mixed deciduous forest, a tall grass prairie, and a winter wheat crop. They
concluded (1) diffuse radiation over direct radiation results in higher light use efficiencies by
plant canopies, (2) diffuse radiation has much less tendency to cause canopy photosynthetic
saturation, (3) the advantages of diffuse radiation over direct radiation increase with radiation
level, (4) temperature as well as vapor pressure deficit can cause different responses in diffuse
and direct canopy photosynthesis, indicating that their effects on terrestrial ecosystem carbon
assimilation may depend upon radiation regimes, i.e., sky conditions.
The potentially significant effect of regional haze on the yield of crops because of
reduction in solar radiation has been examined by Chameides et al. (1999). Using a case study
approach, Chameides et al. (1999) studied the effects of regional haze on crop production in
China where regional haze is especially severe. A simplified assessment of the direct effect of
atmospheric aerosols on agriculture suggests that yields of approximately 70% of crops are being
depressed by at least 3 to 5% by regional scale air pollution and its associated haze (Chameides
etal., 1999).
Effects of Nitrogen Deposition. Nitrogen is required by all organisms. It is a major
constituent of the nucleic acids that determine the genetic character of all living things and the
enzyme proteins that drive the metabolic machinery of every living cell (Galloway, 1998;
Galloway and Cowling, 2002; U.S. Environmental Protection Agency, 1993). Though nitrogen
composes 80% of the total mass of the Earth's atmosphere, it is not biologically available.
Nitrogen fixation is accomplished in nature by certain unique organisms that have developed the
capability of converting N2 to biologically active reduced forms of nitrogen such as ammonia,
4-94
-------
amines, and amino acids, which are the structural constituents of proteins and nucleic acids
(Galloway and Cowling, 2002; Hornung and Langan, 1999; U.S. Environmental Protection
Agency, 1993).
Nitrogen has long been recognized as the nutrient most important for plant growth. It is of
overriding importance in plant metabolism and, to a large extent, governs the utilization of
phosphorus, potassium, and other nutrients. Most of the nitrogen in soils is associated with
organic matter. Typically, the availability of nitrogen via the nitrogen cycle controls net primary
productivity, and possibly, the decomposition rate of plant litter. Photosynthesis is influenced by
nitrogen uptake in that -75% of the nitrogen in a plant leaf is used during the process of
photosynthesis. The nitrogen-photosynthesis relationship is, therefore, critical to the growth of
trees and other plants (Hornung and Langan, 1999; Chapin et al., 1987). Plants usually obtain
nitrogen directly from the soil through their roots by absorbing NH4+ or NO3 , or it is formed by
symbiotic organisms (e.g., bacteria, blue-green algae) in the roots.
Because nitrogen is not readily available and is usually in short supply, it is the chief
element in agricultural fertilizers (Hornung and Langan, 1999). The realization of the
importance of nitrogen in crop production resulted in a search for natural nitrogen sources such
as guano and nitrate deposits. The invention of the Haber-Bosch process in 1913 made reactive
nitrogen (Nr) available for use in food production, and more than half of the food eaten by the
peoples of the world today is produced using fertilizer produced by this process (Galloway,
1998; Galloway and Cowling, 2002).
In nature, nitrogen may be divided into two groups: nonreactive (N2) and reactive (Nr)
nitrogen. Reactive N includes all biologically, photochemically, and radioactively active
nitrogen compounds in the Earth's atmosphere and biosphere (Galloway et al., 2003). Among
those included are the inorganic reduced forms of nitrogen (e.g., NH3 and NH4+), inorganic
oxidized forms (e.g., NOX, HNO3, N2O, and NO3 ), and organic compounds (e.g., urea, amine,
proteins, and nucleic acids) (Galloway et al., 2003).
Food production continues to account for most of the newly created Nr . However, since
around 1965, the magnitude of Nr created by humans began to exceed natural terrestrial creation
of Nr and its conversion back to N2by denitrification. The overall increase in global Nr is the
result of three main causes: (1) widespread cultivation of legumes, rice, and other crops that
promote conversion of N2 to organic nitrogen through biological nitrogen fixation;
4-95
-------
(2) combustion of fossil fuels, which converts both atmospheric N2 and fossil nitrogen to
reactive NOX; and (3) the Haber-Bosch process, which converts nonreactive NH3 to sustain food
production and some industrial activities (Galloway and Cowling, 2002; Galloway et al., 2003).
Reactive nitrogen is now accumulating in the environment on all spatial scales ^pfj" local,
regional and global (Galloway, 1998; Galloway and Cowling, 2002; Galloway et al. 2003).
As a result, Nr is accumulating in various environmental reservoirs, e.g., the atmosphere, soils
and waters (Galloway and Cowling, 2002). The accumulation of Nr in the environment has
effects on humans and ecosystems (Rabalais, 2002; van Egmond et al., 2002; Galloway, 1998).
Large uncertainties regarding the rates of Nr accumulation in the various reservoirs limits
our ability to determine the temporal and spatial distribution of environmental effects. These
uncertainties are of great significance because of the sequential nature of Nr on environmental
processes (Galloway and Cowling, 2002). The sequence of transfers, transformations, and
environmental effects is referred to as the nitrogen cascade (Figure 4-15) (Galloway and
Cowling, 2002; Galloway et al, 2003). A single atom of new NHX or NOX can alter a wide array
of biogeochemical processes and exchanges among environmental reservoirs.
The results of the Nr cascade in the global system and the wide variety of changes in the
nitrogen cycle are both beneficial and detrimental to humans and to ecosystems (Galloway and
Cowling, 2002; Galloway et al., 2003). Though the synthetic fertilizers used in cultivation and
cultivation-induced bacterial nitrogen fertilization (BNF) sustain a large portion of the world's
population, there are consequences: (1) the wide dispersal of Nr by hydrological and
atmospheric transport; (2) the accumulation of Nr in the environment because the rates of its
creation are greater than the rates of its removal through denitrification to nonreactive N2;
(3) Nr creation and accumulation is projected to continue to increase in the future as per capita
use of resources by human populations increases; and (4) Nr accumulation contributes to many
contemporary environmental problems (Galloway et al., 2003).
Among the contemporary environmental problems as listed in Galloway et al. (2003) are
the following:
• increases in Nr lead to production of tropospheric ozone and aerosols and the associated
human health problems (Wolfe and Patz, 2002).
• productivity increases in forests and grasslands and then decreases wherever atmospheric
Nr deposition increases significantly and critical thresholds are exceeded; Nr additions
probably also decrease biodiversity in many natural habitats (Aber et al., 1995);
4-96
-------
Food
Production
People
(Food; Fiber)
Human Activities
NH
Agroecosystem Effects
Crop
->•
Animal
Soil
«J
The Nigrogen
Cascade
Indicates denitrification potential
Aquatic Ecosystems
••••••••••••••••••••••••••••••••••••••••S
Figure 4-15. Illustration of the nitrogen cascade showing the movement of the human-
produced reactive nitrogen (Nr) as it cycles through the various
environmental reservoirs in the atmosphere, terrestrial ecosystems, and
aquatic ecosystems.
Source: Galloway et al. (2003).
reactive nitrogen in association with sulfur is responsible for acidification and loss of
biodiversity in lakes and streams in many regions of the world (Vitousek et al., 1997);
reactive nitrogen is responsible for eutrophication, hypoxia, loss of biodiversity, and
habitat degradation in coastal ecosystems. It is now considered the biggest pollution
problem in coastal waters (Rabalais, 2002);
reactive nitrogen contributes to global climate change and stratospheric ozone
depletion, both of which have an effect on the health of humans and ecosystems
(Cowling et al., 1998).
4-97
-------
The effect of increasing nitrogen inputs (e.g., NOX, NO3 , HNO3) on the nitrogen cycle in
forests, wetlands, and aquatic ecosystems is discussed in detail elsewhere (U.S. Environmental
Protection Agency, 1993, 1997a; Garner, 1994; World Health Organization, 1997).
The deposition of nitrogen in the United States from human activity has doubled between
1961 and 1997 due mainly to the use of inorganic nitrogen fertilizers and NOX emissions from
fossil fuel combustion, with the largest increase occurring in the 1960s and 1970s (Howarth
et al., 2002). Among the most important effects of chronic nitrogen deposition are changes in
the composition of plant communities, disruptions in nutrient cycling, increased soil emissions
of nitrogenous greenhouse gases, accumulation of nitrogen compounds in the soil with enhanced
availability of nitrate or ammonium, soil-mediated effects of acidification, and increased
susceptibility to stress factors in plants (Fenn et al., 1998; Bobbink et al., 1998). A major
concern is "nitrogen saturation," the result of the atmospheric deposition of large amounts of
particulate nitrates. Nitrogen saturation results when additions to soil background nitrogen
(nitrogen loading) exceed the capacity of plants and soil microorganisms to utilize and retain
nitrogen (Aber et al., 1989, 1998; Garner, 1994; U.S. Environmental Protection Agency, 1993).
Under these circumstances, disruptions of ecosystem functions may result (Hornung and Langan,
1999).
Possible ecosystem responses to nitrate saturation, as postulated by Aber and coworkers
(Aber et al., 1989), include (1) a permanent increase in foliar nitrogen and reduced foliar
phosphorus and lignin caused by the lower availability of carbon, phosphorus, and water;
(2) reduced productivity in conifer stands because of disruptions of physiological function;
(3) decreased root biomass and increased nitrification and nitrate leaching; and (4) reduced soil
fertility, resulting from increased cation leaching, increased nitrate and aluminum concentrations
in streams, and decreased water quality. Nitrate saturation implies that some resource other than
nitrogen is limiting biotic function. Water and phosphorus and carbon are the resources most
likely to be the secondary limiting factors for plants and microorganisms, respectively (Aber
et al., 1989). The appearance of nitrogen in soil solutions is an early symptom of excess
nitrogen. In the final stage, disruption of forest structure becomes visible (Garner, 1994).
Changes in nitrogen supply can have a considerable effect on an ecosystem's nutrient
balance (Waring, 1987). Large chronic additions of nitrogen disrupt normal nutrient cycling and
alter many plant and soil processes involved in nitrogen cycling (Aber et al., 1989). Among the
4-98
-------
processes affected are (1) nutrient uptake and allocation in plants, (2) litter production,
(3) immobilization (includes ammonification [the release of ammonia] and nitrification [the
conversion of ammonia to nitrate during decay of litter and soil organic matter]), and (4) nitrate
leaching and trace gas emissions (Figure 4-16) (Aber et al., 1989; Garner 1994).
Deposition
f
Plant
Utilization
s»
Photosynthesis
N
Animal
Proteins
Process altered by
nitrogen saturation
Figure 4-16. Nitrogen cycle (dotted lines indicate processes altered by nitrogen saturation).
Source: Garner (1994).
4-99
-------
Subsequent studies have shown that, although an increase in nitrogen mineralization occurs
initially (i.e., the conversion of soil organic matter to nitrogen in available form [see item 3
above]), not nitrogen mineralization rates are reduced under nitrogen-enriched conditions.
Aber et al. (1998) hypothesized that mycorrhizal assimilation and exudation, using
photosynthate from the host plant as the carbon source, is the dominant process involved in
immobilization of nitrogen. In addition, studies suggested that soil microbial communities
change from predominantly fungal (mycorrhizal) communities predominately bacterial during
nitrate saturation (Aber et al., 1998).
The growth of most forests in North America is limited by the nitrogen supply. Severe
symptoms of nitrogen saturation, however, have been observed in high-elevation, nonaggrading
spruce-fir ecosystems in the Appalachian Mountains, as well as in the eastern hardwood
watersheds at Fernow Experimental Forest near Parsons, WV. Mixed conifer forests and
chaparral watersheds with high smog exposure in the Los Angeles Air Basin also are nitrogen
saturated and exhibit the highest stream water NO3 concentrations for wildlands in North
America (Bytnerowicz and Fenn, 1996; Fenn et al., 1998). Forests in southern California, the
southwestern Sierra Nevada in central California, and the Front Range in northern Colorado have
all been exposed to highly elevated nitrogen deposition, and nitrogen-saturated watersheds have
been reported in the above mentioned areas. Annual nitrogen additions through deposition (6 to
11 kg ha"1 year^as through fall) in the southwestern Sierra Nevada are similar to the nitrogen
storage (4 kg ha"1 year"1) in vegetation growth increment of western forests suggesting that
current nitrogen deposition rates may be near the assimilation capacity of the overstory
vegetation. Ongoing urban expansion will increase the potential for nitrogen saturation of
forests from urban sources (e.g., Salt Lake City, Seattle, Tucson, Denver, central and southern
California) unless there are improved emission controls (Fenn et al., 1998).
Atmospherically deposited nitrogen also can act as a fertilizer in soil low in nitrogen.
Not all plants, however, are capable of utilizing extra nitrogen, as plants vary in their ability to
absorb NH4+ and NO3 (Chapin, et al., 1987). Inputs of nitrogen to natural ecosystems that
alleviate deficiencies and increase growth of some plants can alter competitive relationships and
alter species composition and diversity (Ellenberg, 1987; Kenk and Fischer, 1988; U.S.
Environmental Protection Agency, 1993).
4-100
-------
Not all forest ecosystems react in the same manner to nitrogen deposition. High-elevation
alpine watersheds in the Colorado Front Range (Bowman, 2000) and a deciduous forest in
Ontario, Canada also are naturally saturated even though nitrogen deposition has been moderate
(«8 kg ha"1 year"1). The nitrogen-saturated forests in North America, including estimated inputs
and outputs, are shown in Table 4-14 (Fenn et al., 1998). The Harvard Forest hardwood stand in
Massachusetts, however, has absorbed > 900 kg N/ha without significant NO3 leaching during
an 8-year nitrogen amendment study (Table 4-14) (Fenn et al.,1998). Nitrate leaching losses
were high, on the other hand, in the Harvard Forest pine sites suggesting that deciduous forests
may have a greater capacity for nitrogen retention. In the 8-year experimental study (1988 to
1996), nitrate leaching was observed in the pine stand after the first year (1989) in the high-
nitrogen plots. Further increases were observed in 1995 and 1996, while the hardwood stand did
not show significant increases in NO3 leaching until 1996. The sharp contrast in the response of
the pine and hardwood stands indicates that the mosaic of community types across the landscape
must be considered when determining the regional scale response to nitrogen deposition (Magill
et al., 2000). Johnson et al. (1991a) reported that measurements showing the leaching of NO3
and Al+3 from high elevation forests in the Great Smoky Mountains indicate that these forests
have reached saturation.
Because the competitive equilibrium of plants in any community is finely balanced, the
alteration of one of any number of environmental parameters, (e.g., continued nitrogen
additions) can change the vegetation structure of an ecosystem (Bobbink, 1998; Skeffington and
Wilson, 1988). Increases in soil nitrogen play a selective role. When nitrogen becomes more
readily available, plants adapted to living in an environment of low nitrogen availability will be
replaced by plants capable of using increased nitrogen because they have a competitive
advantage.
Plant succession patterns and biodiversity are affected significantly by chronic nitrogen
additions in some North American ecosystems (Figure 4-17). The location of nitrogen saturated
ecosystems in North America, and the steps leading to nitrogen saturation, are indicated on the
map in Figure 4-16. Conceptual models of regional nitrogen saturation indicate saturation in
New England, in the Colorado alpine ecosystems and in California forests. Fenn et al. (1998)
reported that long-term nitrogen fertilization studies in both New England and Europe, as well,
suggest that some forests receiving chronic inputs of nitrogen may decline in productivity and
4-101
-------
TABLE 4-14. NITROGEN-SATURATED FORESTS IN NORTH AMERICA,
INCLUDING ESTIMATED N INPUTS AND OUTPUTS
Location
Adirondack Mts. northeastern New York
Catskill Mts., southeastern New York
Turkey Lakes Watershed, Ontario, Canada
Whitetop Mt, southwestern Virginia
Fernow, 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
>ij Front Range, Colorado
O
10 San Dimas, San Gabriel Mts.,
southern California
Camp Paivika, San Bernadino Mts.,
southern California
Klamath Mts, northern California
Thompson Forest, Cascade Mts.,
Washington
Forest Type
Northern hardwoods or
hardwood/ conifer mix
Mainly hardwood; some eastern
hemlock
Sugar maple and yellow birch
Red spruce
Mixed hardwood
American beech
Red spruce
Red spruce
Alpine tundra, subalpine conifer
Chapparral and grasslands
Mixed conifer
Western coniferous
Red alder
Elevation
(in)
396-661
335-675
350-400
1650
735-870
1600
1800
1740
3000-4000
580-1080
1600
NA
220
N Input
(kg/ha/year)
9.3 a
10.2 a
7.0-7.7
(as throughfall)
32 c
15-20
3.1d
10.3d
26.6
7.5-8.0
23.3
30
Mainly geologic g
4.7 plus > 100 as
N2 fixation
N Output
(kg/ha/year)
Stage 1 N loss b
Stage 1 and
2 N loss b
17.9-23.6
47 c
6.1
2.9
19.2
20.3
7.5
0.04-19.4
7-26 f
NAg
38.9
Reference
Driscoll and Van Dreason (1993)
Stoddard (1994)
Foster etal. (1989);
Johnson and Lindberg (1992a)
Joslin and Wolfe (1992); Joslin et al. (1992)
Gilliam et al. (1996); Peterjohn et al. (1996)
Johnson and Lindberg (1992b)
Johnson et al. (199 la)
Johnson et al. (199 la)
Williams et al. (1996)
Riggan et al. (1985)
Fenn et al. (1996)
Dahlgren (1994)
Johnson and Lindberg (1992b)
a Estimated total N deposition from wet deposition data is from Driscoll etal. (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/year; Johnson and Lindberg, 1992b).
Stage 1 and 2 of N loss according to the watershed conceptual model of Stoddard (1994). Nitrogen discharge (kg/ha/year) data are not available; only stream water NO3~ concentration trend
data were collected.
0 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.
Estimated total N deposition from throughfall data. Total deposition was estimated based on the throughfall/total N deposition ratio from the nearby Smokies Tower site (Johnson and
Lindberg, 1992b).
e Annual throughfall deposition to the chaparral ecosystem.
'Nitrogen output is from unpublished streamwater data (Fenn and 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 stream water and to groundwater.
g Annual 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/year (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).
-------
• 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:
> Plant biomass and soil organic matter
>-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. Ecosystems Responses to Excess Nitrogen:
»-Nitrate leaching and export
>Eutrophication of estuaries
>-Toxicity of surface waters
*-Foliar nutrient responses
>-Nitrogen mineralization and nitrification
>• Effects on soil organic matter
>Soil acidification, cation depletion, Al toxicity
> Greenhouse gas fluxes
4. Regional N Saturation Conceptual Models:
>• New England forests
> California forests
> Colorado alpine ecosystems
Figure 4-17. Diagrammatic overview of excess nitrogen (N) in North America.
Adapted from: Fernet al. (1998).
experience greater mortality. Long-term fertilization experiments at Mount Ascutney, VT
suggest that declining coniferous forest stands with slow nitrogen cycling may be replaced by
fast-growing deciduous forests that cycle nitrogen rapidly (Fenn et al., 1998).
4-103
-------
Atmospheric nitrogen deposition in the northeastern United States is largely a regional
problem (Driscoll et al., 2001). In contrast, in the western United States, vast acres of land
receive low levels of atmospheric nitrogen deposition but are interspersed with hot spots of
elevated nitrogen deposition downwind of large expanding metropolitan centers or large
agricultural operations (Fenn et al., 2003b).
Fenn et al. (1998) have documented the major effects of Nr deposition in terrestrial and
aquatic ecosystems in the western United States. Primarily these effects are in response to
nitrogen enrichment of systems that are naturally nitrogen limited. Included in these effects are
increased greenhouse gas emissions, higher nitrogen concentrations in plant tissues, and
increased nitrification rates and nitrate (NO3 ) levels in soils, streams, and lakes (Fenn et al.,
2003b). A result of chronic Nr enrichment has resulted in important community-level changes
in vegetation, lichens, mycorrhizae, and phytoplankton, occasionally occurring at relatively low
levels of nitrogen deposition (3 to 8 kgN/ha/year; Baron et al., 2000).
Developments in recent decades in the Colorado Front range have resulted in increased
nitrogen deposition at high-elevation sites since the 1980s . Total deposition values currently
range from 4 to 8 kgN/ha/year (Baron et al., 2000). Competition among species resulting in
changes in community composition is one of the most notable responses to environmental
change (Bowman, 2000). Nitrogen saturation, the result of increased deposition in the alpine
tundra of Niwot Ridge in the Front Range of the southern Rockies in Colorado, has changed
nitrogen cycling and provided the potential for replacement in plant species by more
competitive, faster growing species (Bowman and Steltzer, 1998; Bowman, 2000; Baron et al.,
2000). Plants growing in an alpine tundra, as is true of other plants growing in low resource
environments (e.g., infertile soil, shaded understory, deserts), have been observed to have certain
similar characteristics: a slow grow rate, low photosynthetic rate, low capacity for nutrient
uptake, and low soil microbial activity (Bowman and Steltzer, 1998; Bowman, 2000).
An important feature of such plants is that they continue to grow slowly and tend to respond
even less when provided with an optimal supply and balance of resources (Pearcy et al., 1987;
Chapin, 1991). Plants adapted to cold, moist environments grow more leaves than roots as the
relative availability of nitrogen increases; however, other nutrients may soon become limiting.
These patterns of vegetative development affect the plants capacities to respond to variation in
available resources and to environmental stresses such as frost, high winds, and drought.
4-104
-------
Preformation of buds 3 to 4 years in advance of emergence, reduced cell numbers, and high
biomass allocation to belowground organs also limits the ability of many alpine plants to
respond to variations in their environment (Bowman, 2000). However, significant interspecific
genetic variation influences the capacity of the alpine species to respond to changes in resource
availability. The capacity of subalpine and boreal species in particular, and gymnosperms in
general, to reduce nitrates in either roots or leaves appears to be limited. In addition, the ability
of trees to use nitrogen varies with the age of the tree and the density of the stand (Waring,
1987).
In experimental studies of nitrogen deposition conducted by Wedin and Tilman (1996)
over a 12-year period on Minnesota grasslands, plots dominated by native warm-season grasses
shifted to low-diversity mixtures dominated by cool-season grasses at all but the lowest rates of
nitrogen addition. Grasslands with high nitrogen retention and carbon storage rates were the
most vulnerable to loss of species and major shifts in nitrogen cycling. The shift to low-diversity
mixtures was associated with the decrease in biomass carbon to nitrogen (C:N) ratios, increased
nitrogen mineralization, increased soil nitrate, high nitrogen losses, and low carbon storage
(Wedin and Tilman, 1996). Naeem et al. (1994) experimentally demonstrated under controlled
environmental conditions that the loss of biodiversity, genetic resources, productivity, ecosystem
buffering against ecological perturbation, and loss of aesthetic and commercially valuable
resources also may alter or impair ecosystems services.
The long-term effects of increased nitrogen deposition have been studied in several
western and central European plant communities including lowland heaths, species-rich
grasslands, mesotrophic fens, ombrotrophic bogs, upland moors, forest-floor vegetation, and
freshwater lakes (Bobbink, 1998). Large changes in species composition have been observed in
regions with high nitrogen loadings or in field experiments after years of nitrogen addition
(Bobbink et al., 1998). The increased input of nitrogen gradually increased the availability of
nitrogen in the soil, and its retention because of low rates of leaching and denitrification,
resulting in faster litter decomposition and mineralization rates. Faster growth and greater height
of nitrophilic species enables these plants to shade out the slower growing species, particularly
those in oligotrophic or mesotrophic conditions (Bobbink, 1998; Bobbink et al., 1998). Excess
nitrogen inputs to unmanaged heathlands in the Netherlands has resulted in nitrophilous grass
species replacing the slower growing heath species (Roelofs et al., 1987; Garner, 1994).
4-105
-------
Van Breemen and Van Dijk (1988) noted that over the past several decades the composition of
plants in the forest herb layers has been shifting toward species commonly found in nitrogen-rich
areas. They also observed a decrease in number of the fruiting bodies of mycorrhizal fungi.
Other studies in Europe point out the effects of excessive nitrogen deposition on mixed-oak
forest vegetation along a deposition gradient largely controlled by soil acidity, nitrogen supply,
canopy composition, and location of sample plots (Brunet et al., 1998; Falkengren-Grerup,
1998). Results of the study, using multivariate methods, suggest that nitrogen deposition has
affected the field-layer vegetation directly by increased nitrogen availability and, indirectly, by
accelerating soil acidity. Time series studies indicate that 20 of the 30 field-layer species
(nonwoody plants) that were associated most closely with high nitrogen deposition increased in
frequency in areas with high nitrogen deposition during the past decades. Included in the field-
layer species were many species generally considered nitrophilous; however, there were also
several acid-tolerant species (Brunet et al, 1998). In an experimental study involving 15 herbs
and 13 grasses, Falkengren-Grerup (1998), observed that species with a high nitrogen demand
and a lesser demand for nutrients other than nitrogen were particularly competitive in areas with
acidic soils and high nitrogen deposition. The grasses grew better than herbs with the addition of
nitrogen. It was concluded that, at the highest nitrogen deposition, growth was limited for most
species by the supply of nutrients other than nitrogen; and, at the intermediate nitrogen
concentration, the grasses were more efficient than the herbs in utilizing nitrogen. Nihlgard
(1985) suggested that excessive nitrogen deposition may contribute to forest decline in other
specific regions of Europe. Additionally, Schulze (1989), Heinsdorf (1993), and Lamersdorf and
Meyer (1993) attributed magnesium deficiencies in German forests, in part, to excessive
nitrogen deposition.
The carbon to nitrogen (C:N) ratio of the forest floor can also be changed by nitrogen
deposition over time. This change appears to occur when the ecosystem becomes nitrogen
saturated (Gundersen et al., 1998a). Long-term changes in C:N status have been documented in
Central Europe and indicate that nitrogen deposition has changed the forest floor. In Europe,
low C:N ratios coincide with high deposition regions (Gundersen et al., 1998a). A strong
decrease in forest floor root biomass has been observed with increased nitrogen availability.
Roots and the associated mycorrhizae appear to be an important factor in the accumulation of
organic matter in the forest floor at nitrogen-limited sites. If root growth and mycorrhizal
4-106
-------
formation are impaired by excessive nitrogen deposition, the stability of the forest floor
vegetation may be affected by stimulating turnover and decreasing the root litter input to the
forest floor, thus decreasing the nitrogen that can be stored in the forest floor pool (Gundersen
et al., 1998b). Nitrogen-limited forests have a high capacity for deposited nitrogen to be retained
by plants and microorganisms competing for available nitrogen (Gundersen et al., 1998b).
Nitrate leaching has been correlated significantly with nitrate status but not with nitrate
depositions. The forest floor C:N ratio has been used as a rough indicator of ecosystem nitrogen
status in mature coniferous forests and the risk of nitrate leaching; analyses of European
databases indicate an empirical relationship between forest floor C:N ratio and nitrate leaching
(Gundersen et al., 1998a). Nitrate leaching was observed when the deposition received was
more than 10 kg N/ha. All of the data sets supported the threshold at which nitrate leaching
seems to increase at a C:N ratio of 25. Therefore, to predict the rate of changes in nitrate
leaching, it is necessary to be able to predict the rate of changes in the forest floor C:N ratio.
Decreased foliar and soil nitrogen and soil C:N ratios, as well as changes in nitrogen
mineralization rates, have been observed when comparing responses to nitrogen deposition in
forest stands east and west of the Continental Divide in the Colorado Front Range (Baron et al.,
2000; Rueth and Baron, 2002). Understanding the variability in forest ecosystem response to
nitrogen input is essential to assessing pollution risks (Gundersen et al., 1998a).
The plant root is an important region of nutrient dynamics. The rhizosphere includes the
soil that surrounds and is influenced by plant roots (Wall and Moore, 1999). The mutualistic
relationship between plant roots, fungi, and microbes is critical for the growth of the organisms
involved. The plant provides shelter and carbon; whereas the symbiont provides access to
limiting nutrients such as nitrogen and phosphorus. As indicated above, changes in soil nitrogen
influence the mycorrhizal-plant relationship. Mycorrhizal fungal diversity is associated with
above-ground plant biodiversity, ecosystem variability, and productivity (Wall and Moore,
1999). Aber et al. (1998) showed a close relationship between mycorrhizal fungi and the
conversion of dissolved inorganic nitrogen to soil nitrogen. During nitrogen saturation, soil
microbial communities change from being fungal, and probably being dominated by
mycorrhizae, to being dominated by bacteria. The loss of mycorrhizal function has been
hypothesized as the key process leading to increased nitrification and nitrate mobility. Increased
nitrate mobility leads to increased cation leaching and soil acidification (Aber et al., 1998).
4-107
-------
The interrelationship of above- and below-ground flora is illustrated by the natural
invasion of heathlands by oaks (Quercus robur). The soil-forming factors under the heath have
been vegetation typed during the last 2000 years; whereas the invasion by oaks has been taking
place for only a few decades. Clearly, changes in the ground floor and soil morphology takes
place when trees colonize heath (Nielsen et al., 1999). The distribution of roots also changed
under the three different vegetation types. Under both heather and the Sitka spruce plantation,
the majority of roots are confined to the uppermost horizons; whereas under oak, the roots are
distributed more homogeneously. There was also a change in the C:N ratio when heather was
replaced by oaks. Also, the spontaneous succession of the heath by oaks changed the biological
nutrient cycle into a deeper vertical cycle when compared to the heath where the cycle is
confined to the upper soil horizons. Soils similar to those described in this Jutland, Denmark
study with a mainly organic buffer system seem to respond quickly to changes in vegetation
(Nielsen et al., 1999).
The effects of changes in root to shoot relationships in plants were observed in studies of
the coastal sage scrub (CSS) community in southern California, which is composed of the
drought-deciduous shrub $ Artemisia californica, Enceliafarmosa, andEriogonumfasciculatum.
The CSS in California has been declining in land area and in shrub density over the past 60 years
and is being replaced in many areas by Mediterranean annual grasses (Allen et al., 1998; Padgett
et al., 1999; Padgett and Allen, 1999). Nitrogen deposition was considered as a possible cause.
Up to 45 kg/ha/year are deposited in the Los Angeles Air Basin (Bytnerowicz and Fenn, 1996).
Tracts of land set aside as reserves, which in many cases in southern California are surrounded
by urbanization, receive large amounts of nitrogenous compounds from polluted air. The CSS is
of particular interest, because some 200 sensitive plant species and several federally listed
animal species are found in the area (Allen et al., 1998). Because changes in plant community
structure often can be related to increases in the availability of a limiting soil nutrient or other
resource, experiments were conducted to determine whether increased nitrogen availability was
associated with the significant loss in native shrub cover. Studies indicated that the three native
perennial shrubs (Artemisia californica, Eriogonum fasciculatum, andEnceliafarinosa) tended
to be more nitrophilous than the two exotic annual grasses (Bromus rubens, Avenafatua) and the
weedy pod mustard (Brassica geniculata). These results contrast with most models dealing with
the adaptation of perennial species to stressful environments (Padgett and Allen, 1999).
4-108
-------
If nitrogen were the only variable between the invasive annuals and native shrubs, neither shrubs
nor grasses would have a particular advantage. Although CSS shrubs are able to take up
nitrogen at high rates, native grasses have a denser seedbank and earlier germination than native
species. The native seedlings are unable to compete with dense stands of exotic grasses, and
thus are gradually replaced by the grasses following disturbances such as frequent fire (Eliason
and Allen, 1997; Clone et al., 2002; Yoshida and Allen, 2001). In addition, nitrogen-induced
changes in arbuscular mycorrhizal fungi may also affect the growth of native seedlings.
Nitrogen enrichment of the soils has induced a shift in the arbuscular mycorrhizal community
composition. Larger-spored fungal species (Scutellospora and Gigaspora), due to a failure to
sporulate, decreased in number with a concomitant proliferation of small-spored species of
Glomus aggregation, G. leptotichum, and G. geospomm, indicating a strong selective pressure
for the smaller spored species of fungi (Edgerton-Warburton and Allen, 2000). These results
demonstrate that nitrogen enrichment of the soil significantly alters the arbuscular mycorrhizal
species composition and richness and markedly decreases the overall diversity of the arbuscular
mycorrhizal community. The decline in coastal sage scrub species can, therefore, be directly
linked to the decline of the arbuscular mycorrhizal community (Tidgerton-Warburton and Allen,
2000).
In addition to excess nitrogen deposition effects on terrestrial ecosystems of the types
noted above (e.g., dominant species shifts and other biodiversity impacts), direct atmospheric
nitrogen deposition and increased nitrogen inputs via runoff into streams, rivers, lakes, and
oceans can noticeably affect aquatic ecosystems as well (Figure 4-15). Estuaries are among the
most intensely fertilized ecosystems on Earth, receiving far greater nutrient inputs than other
systems. Chesapeake Bay is a prime example (Fenn et al., 1998). Another illustrative example
is found in recently reported research (Paerl et al., 2001) characterizing the effects of nitrogen
deposition on the Pamlico Sound, NC estuarine complex, which serves as a key fisheries nursery
supporting an estimated 80% of commercial and recreational finfish and shellfish catches in the
southeastern U.S. Atlantic coastal region. Such direct atmospheric nitrogen deposition onto
waterways feeding into the Pamlico Sound or onto the sound itself, combined with indirect
nitrogen inputs via runoff from upstream watersheds, contribute to conditions of severe water
oxygen depletion; formation of algae blooms in portions of the Pamlico Sound estuarine
complex; altered fish distributions, catches, and physiological states; and increases in the
4-109
-------
incidence of disease in fish. Under extreme conditions of especially high rainfall rate events
(e.g., hurricanes) affecting watershed areas feeding into the sound, the effects of nitrogen runoff
(in combination with excess loadings of metals or other nutrients) can be massive — e.g.,
creation of the widespread "dead-zone" affecting large areas of the Pamlico Sound for many
months after hurricane Fran in 1996 and hurricanes Dennis, Floyd, and Irene in 1999 impacted
eastern North Carolina.
The primary pathways of nitrogen loss from forest ecosystems are hydrological transport
beyond the rooting zone into groundwater or stream water, or surface flows of organic nitrogen
as nitrate and nitrogen loss associated with soil erosion (Fenn et al., 1998). Stream water nitrate
concentrations have been related to forest successional stage in the eastern United States.
Logging and fire histories of an area are major variables determining the capability of a forest
stand to retain nitrogen. Nitrogen concentrations were high in mature ecosystems after
disturbances such as clearcutting, but lower in mid-successional forests.
Nitrogen saturation of a high elevation watershed in the southern Appalachian Mountains
was observed to affect stream water chemistry. High nitrate concentrations have been observed
in streams draining undisturbed watersheds in the Great Smoky Mountains National Park in
Tennessee and North Carolina. Nitrate concentrations were highest at higher elevations and in
areas around old-growth forests that had never been logged (Fenn et al., 1998).
In the Northeast, nitrogen is the element most responsible for eutrophication in coastal
waters of the region (Jaworski et al., 1997). There has been a 3- to 8-fold increase in nitrogen
flux from 10 watersheds in the northeastern United States since the early 1900s. These increases
are associated with the deposition of nitrogen oxide emissions from combustion, which have
increased 5-fold. Riverine nitrogen fluxes have been correlated with atmospheric deposition
onto their landscapes and with nitrogen oxides emissions into their airsheds. Data from
10 benchmark watersheds with good historical records, indicate that -36 to 80% of the riverine
total nitrogen export, with an average of 64%, was derived directly or indirectly from nitrogen
oxide emissions (Jaworski et al., 1997).
Nitrogen saturation of a high-elevation watershed in the southern Appalachian Mountains
was observed to affect stream water chemistry. The Great Smoky Mountains in the southeastern
United States receive high total atmospheric deposition of sulfur and nitrogen (2,200 Eq/ha/year
of total sulfur and approximately 1,990 Eq/ha/year of total nitrogen). A major portion of the
4-110
-------
atmospheric loading is from dry and cloud deposition. Extensive surveys conducted in October
1993 and March 1994 indicated that stream pH values were near or below pH 5.5 and that the
acid neutralizing capacity (ANC) was below 50 jieq/L at high elevations. Analysis of stream
water indicated that nitrate was the dominant anion (Flum and Nodvin, 1995; Nodvin et al.,
1995). The study was expanded to the watershed scale with the monitoring of precipitation,
throughfall, stream hydrology, and stream chemistry. Nitrogen saturation of the watershed
resulted in extremely high exports of nitrate and promoted both chronic and episodic stream
acidification in which the nitrate was the dominant ion. Significant exports of base cation was
also observed. Nitrification of the watershed soils resulted in elevations of soil solution
aluminum concentrations to levels known to inhibit calcium uptake in red spruce (Nodvin
etal., 1995).
Excessive nitrogen loss is a symptom of terrestrial ecosystem dysfunction and results in the
degradation of water quality with potentially deleterious effects on terrestrial and aquatic
ecosystems (Fenn and Poth, 1999). Data from a number of hydrologic, edaphic, and plant
indicators indicate that the mixed conifer forests and chaparral systems directly exposed to air
pollution from greater Los Angeles are nitrogen saturated. Preliminary data suggests that
symptoms of nitrogen saturation are evident in mixed conifer or chaparral sites receiving
atmospheric deposition of 20 to 25 kg/N/ha/year (Fenn et al, 1996). Available data clearly
indicate that ecosystems with a Mediterranean climate have a limited capacity to retain nitrogen
within the terrestrial system (Fenn and Poth, 1999). A 3-year study of stream water NO3
concentrations along nitrogen deposition gradients in the San Bernardino Mountains in southern
California evaluated stream water quality and whether the stream water concentrations covaried
with nitrogen deposition across pollution gradients in the San Bernardino Mountains. Stream
water NO3 concentrations at Devil Canyon in the San Gabriel Mountains northeast of
Los Angeles are the highest reported in North America for forested watersheds (Fenn and Poth,
1999). Five of the six streams monitored maintained elevated NO3 throughout the year. Peak
nitrate concentrations ranged from 40 to 350 |imol/L. In the San Gorgonio Wilderness, an area
of low to moderate deposition where 12 streams were sampled, only the five that had the greatest
air pollution exposure had high NO3 concentrations. These results suggest a strong association
between NO3 levels exported in stream water and the severity of chronic nitrogen deposition to
the terrestrial watersheds. However, nitrogen processing within terrestrial and aquatic systems,
4-111
-------
even in areas with high nitrogen deposition, determine stream water NO3 concentrations
(Fenn and Poth, 1999). The Fernow Experimental Forest in West Virginia, the Great Smoky
Mountains National park in Tennessee, and watersheds in southwestern Pennsylvania are the
only undisturbed forested sites in North America known to have stream water NO3
concentrations within the range of values found at Devil Canyon (Fenn and Poth, 1999).
Effects of Sulfur Deposition. Sulfur is a major component of plant proteins and, as such,
is an essential plant nutrient. The most important source of sulfur is sulfate taken up from the
soil by plant roots, even though plants can use atmospheric SO2 (Marschner, 1995). The
availability of organically bound sulfur in soils depends largely on microbial decomposition, a
relatively slow process. The major factor controlling the movement of sulfur from the soil into
vegetation is the rate of release from the organic to the inorganic compartment (May et al., 1972;
U.S. Environmental Protection Agency, 1982; Marschner, 1995). Sulfur plays a critical role in
agriculture as an essential component of the balanced fertilizers needed to grow and increase
worldwide food production (Ceccotti and Messick, 1997). Atmospheric deposition is an
important component of the sulfur cycle. This is true not only in polluted areas where
atmospheric deposition is very high, but also in areas of low sulfur input. Additions of sulfur
into the soil in the form of SO42 could alter the important organic-sulfur/organic-nitrogen
relationship involved in protein formation in plants. The biochemical relationship between
sulfur and nitrogen in plant proteins and the regulatory coupling of sulfur and nitrogen
metabolism indicate that neither element can be assessed adequately without reference to the
other. A sulfur deficiency reduces nitrate reductase and, to a similar extent, glutamine
synthetase activity. Nitrogen uptake in forests, therefore, may be loosely regulated by sulfur
availability, but sulfate additions in excess of needs do not necessarily lead to injury (Turner and
Lambert, 1980; Hogan et al., 1998).
Only two decades ago, there was little information comparing sulfur cycling in forests with
other nutrients, especially nitrogen. With the discovery of deficiencies in some unpolluted
regions (Kelly and Lambert, 1972; Humphreys et al., 1975; Turner et al., 1977; Schnug, 1997)
and excesses associated with acidic deposition in other regions of the world (Meiwes and
Khanna, 1981; Shriner and Henderson, 1978; Johnson et al., 1982a,b), interest in sulfur nutrition
and cycling in forests has heightened. General reviews of sulfur cycling in forests have been
4-112
-------
written by Turner and Lambert (1980), Johnson (1984), Mitchell et al. (1992a,b), and Hogan
et al. (1998). The salient elements of the sulfur cycle as it may be affected by changing
atmospheric deposition are summarized by Johnson and Mitchell (1998). Sulfur has become the
most important limiting factor in European agriculture because of the desulfurization of
industrial emissions (Schnug, 1997).
Most studies dealing with the effects of sulfur deposition on plant communities have been
conducted in the vicinity of point sources and have investigated the above-ground effects of SO2
or the acidifying effects of sulfate on soils (Krupa and Legge, 1998; Dreisinger and McGovern,
1970; Legge, 1980; Winner and Bewley, 1978a,b; Laurenroth and Michunas, 1985; U.S.
Environmental Protection Agency, 1982). Krupa and Legge (1986), however, observed a
pronounced increase in foliar sulfur concentrations in all age classes of needles of the hybrid
pine lodgepole x jack pine (Pinus contorta x P. banksiand). This vegetation had been exposed
to chronic low concentrations of sulfur dioxide (SO2) and hydrogen sulfide (H2S) for more than
20 years and, then, to fugitive sulfur aerosol. Observations under the microscope showed no
sulfur deposits on the needle surfaces and led to the conclusion that the sulfur in the needles was
derived from the soil. The oxidation of elemental sulfur and the generation of protons is well
known for the soils of Alberta, Canada. This process is mediated by bacteria of the Thiobacillus
sp. As elemental sulfur is gradually converted to protonated SO4, it can be leached downward
and readily taken up by plant roots. The activity of Thiobacillus sp. is stimulated by elemental
sulfur additions (Krupa and Legge, 1986).
Effects of Acidic Deposition on Forest Soils. Acidic deposition over the past quarter of a
century has emerged as a critical environmental stress that affects forested landscapes and
aquatic ecosystems in North America, Europe, and Asia (Driscoll et al., 2001). Acidic
deposition can originate from transboundary air pollution and affect large geographic areas. It is
composed of ions, gases, particles derived from gaseous emissions of SO2, NOX, NH3, and
particulate emissions of acidifying and neutralizing compounds and is highly variable across
space and time. It links air pollution to diverse terrestrial and aquatic ecosystems and alters the
interactions of the H+ with many elements (e.g., S, N, Ca, Mg, Al, and Hg). Acidic deposition
contributes directly and indirectly to biological stress and the degradation of ecosystems and has
4-113
-------
played a major role in the recent acidification of soil in some areas of Europe and, to a more
limited extent, eastern North America (Driscoll et al., 2001).
Substantial and previously unsuspected changes in soils have been observed in polluted
areas of eastern North America, the United Kingdom, Sweden, and central Europe and in less
polluted regions of Australia and western North American (reviewed by Johnson et al., 1999 and
by Huntington, 2000). In some cases, trends were toward more acidic soils (e.g., Markewitz
et al., 1998); in others cases, there were no consistent trends, with some soils showing increases
and some showing decreases at different sampling times and some showing no change (e.g.,
Johnson and Todd, 1998; Trettin et al., 1999; Yanai et al., 1999).
Significant changes in soil chemistry have occurred at many sites in the eastern United
States during recent decades. Patterns of change in tree ring chemistry, principally at high
elevations sites in the eastern United States, reflect the changing inputs of regional pollutants to
forests. A temporal sequence of changes in uptake patterns, and possibly in tree growth, would
be expected if significant base cation mobilization and depletion of base cations from eastern
forest soils has occurred. Temporal changes in the chemistry of the tree rings of red spruce were
examined as indicators of historical changes in the red spruce's chemical environment.
Analysis of changes in wood chemistry from samples across several sites indicate that
there have been substantial departures from the expected linear decreases in calcium
accumulation patterns in wood. A region-wide calcium increase above expected levels followed
by decreasing changes in wood calcium suggest that calcium mobilization began possibly 30 to
40 years ago and has been followed by reduced accumulation rates in wood, presumably
associated with decreasing calcium availability in soil (Bondietti and McLaughlin, 1992).
The period of calcium mobilization coincides with a region-wide increase in the growth rate of
red spruce; whereas the period of decreasing levels of calcium in wood corresponds temporally
with patterns of decreasing radial growth at high elevation sites throughout the region during the
past 20 to 30 years. The decline in wood calcium suggests that calcium loss may have increased
to the point at which base saturation of soils has been reduced. Increases in aluminum and iron
typically occur as base cations are removed from the soils by tree uptake (Bondietti and
McLaughlin, 1992). The changes are spatially and temporally consistent with changes in the
emissions of SO2 and NO2 across the region and suggest that increased acidification of soils has
occurred.
4-114
-------
Studies by Shortle and Bondietti (1992) support the view that changes in soil chemistry in
eastern North America forest sites occurred many decades ago, "before anybody was looking."
Sulfur and nitrogen emissions began increasing in eastern North America in the 1920s and
continued to increase into the 1980s when sulfur began to decrease but nitrogen emissions did
not (Garner et al., 1989). Shortle and Bondietti (1992) present evidence that, from the late 1940s
into the 1960s, the mor humus (organic) layer of acid-sensitive forest sites in eastern North
America underwent a significant change that resulted in the loss of exchangeable essential base
cations and interrupted the critical base nutrient cycles between mature trees and the root-humus
complex. The timing of the effect appears to have coincided with the period when the SOX and
NOX emissions in eastern North America subject to long-range transport were increasing the
most rapidly (see above; Shortle and Bondietti, 1992). Although forest ecosystems other than
the high-elevation spruce-fir forests are not currently manifesting symptoms of injury directly
attributable to acid deposition, less sensitive forests throughout the United States are
experiencing gradual losses of base cation nutrients, which, in many cases, will reduce the
quality of forest nutrition over the long term (National Science and Technology Council, 1998).
In some cases it may not even take decades, because these forests already have been receiving
sulfur and nitrogen deposition for many years. The current status of forest ecosystems in
different U.S. geographic regions varies, as does their sensitivity to nitrogen and sulfur
deposition. Variation in potential future forest responses or sensitivity are caused, in part, by
differences in deposition of sulfur and nitrogen, ecosystem sensitivities to sulfur and nitrogen
additions, and the responses of soils to sulfur and nitrogen inputs (National Science and
Technology Council, 1998).
Acidic deposition has played a major role in recent soil acidification in some areas of
Europe and, to a more limited extent, eastern North America. Examples include the study by
Hauhs (1989) at Lange Bramke, Germany, which indicated that leaching was of major
importance in causing substantial reduction in soil-exchangeable base cations over a 10-year
period (1974 to 1984). Soil acidification and its effects result from the deposition of NO3" and
SO42 and associated H+. The effects of excessive nitrogen deposition on soil acidification and
nutrient imbalances have been well established in Dutch forests (Van Breemen et al., 1982;
Roelofs et al., 1985; Van Dijk and Roelofs, 1988). For example, Roelofs et al. (1987) proposed
that NH3/NH4+ deposition leads to heathland changes via two modes: acidification of the soil
4-115
-------
with the loss of cations K+, Ca2+, and Mg2+; and nitrogen enrichment that results in "abnormal"
plant growth rates and altered competitive relationships. Nihlgard (1985) suggested that
excessive nitrogen deposition may contribute to forest decline in other regions of Europe.
Falkengren-Grerup (1987) noted that, over about 50 years, unexpectedly large increases in the
growth of beech trees (Fagus sylvatica L.) were associated with decreases in pH and
exchangeable cations in some sites in southernmost Sweden.
Likens et al. (1996, 1998) suggested that soils are changing at the Hubbard Brook
Watershed, NH, because of a combination of acidic deposition and reduced base cation
deposition. They surmised, based on long-term trends in stream water data, that large amounts
of Ca and Mg have been lost from the soil-exchange complex over a 30-year period, from
approximately 1960 to 1990. The authors speculate that the declines in base cations in soils may
be the cause of recent slowdowns in forest growth at Hubbard Brook. In a follow-up study,
however, Yanai et al. (1999) found no significant decline in Ca and Mg in the forest floors at
Hubbard Brook over the period 1976 to 1997. They also found both gains and losses in forest
floor Ca and Mg between 1980 and 1990 in a regional survey. Thus, they concluded that "forest
floors in the region are not currently experiencing rapid losses of base cations, although losses
may have preceded the onset of these three studies." The biogeochemistry of calcium at
Hubbard Brook is discussed in detail by Likens et al. (1998).
Hydrogen ions entering a forest ecosystem first encounter the forest canopy, where they
are often exchanged for base cations that then appear in throughfall (Figure 4-18 depicts a model
of H+ sources and sinks). Base cations leached from the foliage must be replaced through uptake
from the soil, or foliage cations will be reduced by the amounts leached. In the former case, the
acidification effect is transferred to the soil where H+ is exchanged for a base cation at the
root-soil interface. The uptake of base cations or NH4+ by vegetation or soil microorganisms
causes the release of H+ in order to maintain charge balance; conversely, the uptake of nutrients
in anionic form (NO3 , SO42 , PO43 ) causes the release of Off in order to maintain charge
balance. Thus, the net acidifying effect of uptake is the difference between cation and anion
uptake. The form of ions taken up is known for all nutrients but nitrogen, because either NH4+ or
N(V can be used by the vegetation. In that nitrogen is the nutrient taken up in the greatest
quantities, the uncertainty in the ionic form of nitrogen utilized creates great uncertainty in the
overall H+ budget for soils (Johnson, 1992).
4-116
-------
Deposition
H+
000000000
000000000
000000000
Soil
Organism
Uptake
C02 + H20
Carbonic Acid Formation
R-COOH
Organic Acid Formation
2OI-T
Soil
Organism
Uptake
2OH"
Leaching
Figure 4-18. Schematic of sources and sinks of hydrogen ions in a forest.
Source: Taylor etal. (1994).
The cycles of base cations differ from those of N, P, and S in several respects. The fact
that calcium, potassium, and magnesium exist primarily as cations in solution, whereas N, P, and
S exist primarily as anions, has major implications for the cycling of the nutrients and the effects
of acid deposition on these cycles. The most commonly accepted model of base cation cycling
in soils is one in which base cations are released by the weathering of primary minerals to cation
exchange sites, making the cation available for either plant uptake or leaching (Figure 4-18).
The introduction of FT by atmospheric deposition or by internal processes will affect the fluxes
of Ca, K, and Mg via cation exchange or weathering processes. Therefore, soil leaching is often
4-117
-------
of major importance in cation cycles, and many forest ecosystems show a net loss of base cations
(Johnson, 1992).
Two basic types of soil change are involved: (1) a short-term intensity type change
resulting from the concentrations of chemicals in soil water and (2) a long-term capacity change
based on the total content of bases, aluminum, and iron stored in the soil (Reuss and Johnson,
1986; Van Breemen et al., 1983). Changes in intensity factors can have a rapid effect on the
chemistry of soil solutions. Rapid changes in intensity resulting from the addition of sulfur and
nitrogen in acidic deposition can cause immediate increases in acidity and the mobilization of
aluminum in soil solutions. Increased aluminum concentrations and an increase in the Ca: Al
ratio in soil solution have been linked to a significant reduction in the availability of essential
base cations to plants, an increase in plant respiration, and increased biochemical stress (National
Science and Technology Council, 1998).
Capacity changes are the result of many factors acting over long time periods. The content
of base cations (calcium, magnesium, sodium, and potassium) in soils results from additions via
atmospheric deposition, decomposition of vegetation, and geologic weathering; losses may occur
through plant uptake and leaching. Increased leaching of base cations may result in nutrient
deficiencies in soils, as has been happening in some sensitive forest ecosystems (National
Science and Technology Council, 1998).
Aluminum toxicity is a possibility in acidified soils. Atmospheric deposition (or any other
source of mineral anions) can increase the concentration of Al, especially A13+, in soil solution
without causing significant soil acidification (Johnson and Taylor, 1989). Aluminum can be
brought into soil solution in two ways: (1) by acidification of the soil and (2) by increasing the
total anion and cation concentration of the soil solution. The introduction of mobile, mineral
acid anions to an acid soil will cause increases in the concentration of aluminum in the soil
solution, but extremely acid soils in the absence of mineral acid anions will not produce a
solution high in aluminum. An excellent review of the relationships among the most widely
used cation-exchange equations and their implications for the mobilization of aluminum into soil
solution is provided by Reuss (1983).
A major concern has been that soil acidity would lead to nutrient deficiency. Calcium is
essential for root development and the formation of wood, and it plays a major role in cell
membrane integrity and cell wall structure. Aluminum concentrations in the soil can influence
4-118
-------
forest tree growth in regions where acidic deposition and natural acidifying processes increase
soil acidity. Acidic deposition mobilizes calcium and magnesium, which are essential for root
development and stem growth. Mobilized aluminum can also bind to fine root tips of red spruce,
further limiting calcium and magnesium uptake (Shortle and Smith, 1988; Shortle et al., 1997).
There is abundant evidence that aluminum is toxic to plants. Upon entering tree roots,
aluminum accumulates in root tissues (Thornton et al., 1987; Vogt et al., 1987a, b). Reductions
in calcium uptake have been associated with increased aluminum uptake (Clarkson and
Sanderson, 1971), and a number of studies suggest that the toxic effect of aluminum on forest
trees could be caused by a Ca2+ deficiency (Shortle and Smith, 1988; Smith, 1990a). Mature
trees have a high calcium requirement relative to agriculture crops (Rennie, 1955). Shortle and
Smith (1988) attributed the decline of red spruce in eight stands across northern New England
from Vermont to Maine to an imbalance of A13+ and Ca2+ in fine root development.
To be taken up from the soil by roots, Ca must be dissolved in soil water (Lawrence and
Huntington, 1999). Aluminum in soil solution reduces Ca uptake by competing for binding sites
in the cortex of fine roots. Tree species may be adversely affected if high aluminum to nutrient
ratios create a nutrient deficiency by limiting the uptake of calcium and magnesium (Shortle and
Smith, 1988; Garner, 1994). Acid deposition, by lowering the pH of Al-rich soil, can increase
Al concentrations in soil water through dissolution and ion exchange processes. Aluminum is
more readily taken up than is Ca because of its greater affinity for negatively charged surfaces
and, when present in the forest floor, Al tends to displace adsorbed Ca, causing it to be more
readily leached. The continued buildup of Al in the forest floor layer, where nutrient uptake is
greatest, can lower the efficiency of Ca uptake when the Ca:Al ratio in soil water is less than one
(Lawrence and Huntington, 1999). A reduction in Ca uptake suppresses cambial growth and
reduces the rate of wood (annual ring) formation, decreases the amount of functional sapwood
and live crown, and predisposes trees to disease and injury from stress agents when the
functional sapwood becomes less that 25% of cross-sectional stem area (Smith, 1990a). A 1968
Swedish report to the United Nations postulated a decrease in forest growth of ~1.5%/year when
the Ca:Al ratio in soil water is less than one (Lawrence and Huntington, 1999). The concern that
acidification and nutrient deficiency may result in forest decline remains today.
Acidic deposition has been firmly implicated as a causal factor in the northeastern high-
elevation decline of red spruce (DeHayes et al., 1999). The frequency of freezing injury of red
4-119
-------
spruce has increased over the past 40 years, a period that coincides with increase emissions of
sulfur and nitrogen oxides and acidic deposition (DeHayes et al., 1999). Studies indicate that
there is a significant positive association between cold tolerance and foliar Ca in trees exhibiting
deficiency in foliar Ca. Most of the calcium in conifer needles is insoluble calcium oxalate and
pectate crystals, which are of little physiological importance. It is the labile calcium ions in
equilibrium within the plasma membrane that are of major physiological importance (DeHayes
et al., 1999). The membrane-associated pool of calcium (mCa), although a relatively small
fraction of total foliar ion pools, strongly influences the response of cells to changing
environmental conditions. The plant plasma membrane plays a critical role in mediating cold
acclimation and low-temperature injury. The leaching of Ca associated with acidic deposition is
considered to be the result of cation exchange due to exposure to the H+ion. The studies of
DeHayes et al. (1999) demonstrated that the direct deposition of acidic deposition on needles
represents a unique environmental stress, in that it preferentially removes mCa, which is not
readily replaced in the autumn. They propose that direct acidic deposition on red spruce foliage
preferentially displaces those Ca ions specifically associated with the plasma membranes of
mesophyll cells resulting in the reduction of mCa and the destabilizing of plasma membranes
and depletion of messenger Ca ions. Further, DeHayes et al. (1999) state that their studies raise
the strong possibility that acid rain alteration of the mCa and membrane integrity is not unique to
red spruce but has been demonstrated in many other northern temperate forest tree species,
including: yellow birch (Betula alleghaniensis)., white spruce (Picea glaucus\ red maple {Acer
rubrum), eastern white pine (Pinus strobus), and sugar maple (Acer saccharum). Assessments of
mCa, membrane integrity, and the effects of other secondary stressors have not yet been made
for these species.
Seasonal and episodic acidification of surface waters have been observed in the eastern
United States, Canada, and Europe (Hyer et al., 1995). In the Northeast, the Shenandoah
National Park in Virginia, and the Great Smoky Mountains, episodic acidification has been
associated with the nitrate ion (Driscoll et al., 2001; Hyer et al., 1995; Eshleman et al., 1995).
The short-term acid episodes occur during spring snowmelts and large precipitation events
(Driscoll et al., 2001). Episodic acidification of surface waters has usually been considered to be
a transient loss of acid neutralizing capacity associated with snowmelt/rainfall runoff and, as
such, represents a short-term (hours to weeks) effect considered to be distinguishable from the
4-120
-------
chronic long-term (years to centuries) changes in acidity. Studies of both episodic and chronic
acidification of surface waters indicate that acidification can have long-term adverse effects on
fish populations and result in the decline of species richness and the abundance of zooplankton
and macroinvertebrates (Driscoll et al., 2001; Eshleman et al., 1995). Nitrogen saturation of
soils and the slow release of nitrates has been shown to inhibit the recovery of acid sensitive
systems (Driscoll et al., 2001). The acidification of aquatic ecosystems and effects on aquatic
biota are discussed in more detail in the EPA document Air Quality Criteria for Nitrogen Oxides
(U.S. Environmental Protection Agency, 1993).
Air pollution is not the sole cause of soil change. High rates of acidification are occurring
in less polluted regions of the western United States and Australia because of internal soil
processes, such as tree uptake of nitrate and nitrification associated with excessive nitrogen
fixation (Johnson et al., 1991b). Many studies have shown that acidic deposition is not a
necessary condition for the presence of extremely acidic soils, as evidenced by their presence in
unpolluted, even pristine, forests of the northwestern United States and Alaska (Johnson et al.,
1991b). Soil can become acidic when H+ ions attached to NH4+ or JdNO3 remain in the soil after
nitrogen is taken up by plants. For example, Johnson et al. (1982b) found significant reductions
in exchangeable K+ over a period of only 14 years in a relatively unpolluted Douglas fir
Integrated Forest Study (IFS) site in the Washington Cascades. The effects of acid deposition at
this site were negligible relative to the effects of natural leaching (primarily of carbonic acid)
and nitrogen tree uptake (Cole and Johnson, 1977). Even in polluted regions, numerous studies
have shown the importance of tree uptake of NH4+ and NO3 in soil acidification. Binkley et al.
(1989) attributed the marked acidification (pH decline of 0.3 to 0.8 units and base saturation
declines of 30 to 80%) of abandoned agricultural soil in South Carolina over a 20-year period to
NH4+ and NO3 uptake by a loblolly pine plantation.
An interesting example of uptake effects on soil acidification is that of Al uptake and
cycling (Johnson et al., 1991b). Aluminum accumulation in the leaves of coachwood
(Ceratopetalum apetalum) in Australia has been found to have a major effect on the distribution
and cycling of base cations (Turner and Kelly, 1981). The presence of C. apetalum as a
secondary tree layer beneath brush box (Lophostemon confertus) was found to lead to increased
soil exchangeable A13+ and decreased soil exchangeable Ca2+ (Turner and Kelly, 1981). The
4-121
-------
constant addition of aluminum-rich litter fall obviously has had a substantial effect on soil
acidification, even if base cation uptake is not involved directly.
Given the potential importance of particulate deposition to the base cation status of forest
ecosystems, the findings of Driscoll et al. (1989, 2001) and Hedin et al. (1994) are especially
relevant. Driscoll et al. (1989, 2001) noted a decline in both SO42 and base cations in both
atmospheric deposition and stream water over the past two decades at Hubbard Brook
Watershed, NH. The decline in SO42 deposition was attributed to a decline in emissions, and
the decline in stream water SO42 was attributed to the decline in sulfur deposition. Hedin et al.
(1994) reported a steep decline in atmospheric base cation concentrations in both Europe and
North America over the past 10 to 20 years. The reductions in SO2 emissions in Europe and
North America in recent years have not been accompanied by equivalent declines in net acidity
related to sulfate in precipitation. These current declines in sulfur deposition have, to varying
degrees, been offset by declines in base cations and may be contributing "to the increased
sensitivity of poorly buffered systems." Analysis of the data from the IPS supports the authors'
contention that atmospheric base cation inputs may seriously affect ecosystem processes.
Johnson et al. (1994) analyzed base cation cycles at the Whiteface Mountain IPS site in detail
and concluded that Ca losses from the forest floor were much greater than occurred historically,
based on historical changes in forest floor Ca observed in an earlier study. Further, the authors
suggested that the difference between historical and current net loss rates of forest floor Ca may
be caused by sharply reduced atmospheric inputs of Ca after about 1970 and may be exacerbated
by sulfate leaching (Johnson et al., 1994).
The Ca: Al molar ratio has been suggested to be a valuable ecological indicator of an
approximate threshold beyond which the risk of forest injury from Al stress and nutrient
imbalances increases (Cronan and Grigal, 1995). The Ca:Al ratio also can be used as an
indicator to assess forest ecosystem changes over time in response to acidic deposition, forest
harvesting, or other process that contribute to acid soil infertility. This ratio, however, may not
be a reliable indicator of stress in areas with both high atmospheric deposition of ammonium and
magnesium deficiency via antagonism involving ammonium rather than aluminum and in areas
with soil solutions with calcium concentrations greater than 500 |imol/L (National Science and
Technology Council, 1998). Cronan and Grigal (1995), based on a review of the literature, have
made the following estimates for determining the effects of acidic deposition on tree growth or
4-122
-------
nutrition: (a) forests have a 50% risk of adverse effects if the Ca:Al ration is 1.0; (b) the risk is
75% if the ratio is 0.5:1.0; and (c) the risk approaches 100% if the ratio is 0.2:1.0. The Ca:Al
ratio of soil solution provides only an index of the potential for Al stress. Cronan and Grigal
(1995) stated that the overall uncertainty of the Ca:Al ratio associated with a given probability
ratio is considered to be approximately ±50%. Determination of thresholds for potential forest
effects requires the use of four successive measurement endpoints in the soil, soil solution, and
plant tissue: (a) soil base saturation less than 15% of effective cation exchange capacity; (b) soil
solution Ca:Al molar ratio less than 1.0:1.0 for 50% risk; (c) fine roots tissue Ca:Al molar ratio
less than 0.2:1.0 for 50% risk; and (d) foliar tissue Ca:Al molar ratio less than 12.5:1.0 for 50%
risk. The application of the Ca:Al ratio indicator for assessment and monitoring of forest health
risks has been recommended for sites or in geographic regions where the soil base saturation is
< 15%.
Critical Loads. The critical loads framework for assessing the effects of atmospheric
deposition originated in Europe where the concept has generally been accepted as the basis for
abatement strategies to reduce or prevent injury to the functioning and vitality of forest
ecosystems caused by long-range transboundary chronic acidic deposition (L0kke et al., 1996).
The critical load has been defined as a "quantitative estimate of an exposure to one or more
pollutants below which significant harmful effects on specified sensitive elements of the
environment do not occur according to present knowledge" (L0kke et al., 1996). This concept is
useful for estimating the amounts of pollutants that sensitive ecosystems can absorb on a
sustained basis without experiencing measurable degradation.
The response to pollutant deposition of an ecosystem is a direct function of the level of
sensitivity of the ecosystem to the pollutant and its capability to ameliorate change. The
estimation of ecosystem critical loads requires an understanding of how an ecosystem will
respond to different loading rates in the long term. The approach can be of special value for
ecosystems receiving chronic deposition of pollutants such as nitrogen and sulfur. A program
was designed to develop and evaluate a framework for setting critical loads of nitrogen and
sulfur in the United States in 1989 (Strickland et al., 1993). A flexible six-step approach has
been outlined for use with the critical load framework (Figure 4-19). These steps are
(1) selection of ecosystem components, indicators, and characterization of the resource;
4-123
-------
Identify Regions
Subregions
Populations
Ecosystems
Characterize Resource
Issue
Water Quality?
Human Health?
Biodiversity?
*• Characterize Deposition
> Model Selection <
Response Hindcasting
Select Endpoints
Nitrate < 0.1 mg/L
^Species Richnessj
pH>5.3
>• Response Forecasting
Ecosystem Response Presentation (Mapping)
Figure 4-19. Key elements of proposed framework for determining critical loads for
nitrogen and sulfur in the United States.
Source: Strickland etal. (1993).
(2) definition of the functional subregions; (3) characterization of deposition within each of the
subregions; (4) definition of an assessment endpoint; (5) selection and application of models;
and (6) mapping of projected ecosystem responses. The approach permits comparability in
ecosystem characteristic and data availability (Strickland et al., 1993).
Ecological endpoints or indicators are measurable characteristics related to the structure,
composition, or functioning of ecological systems (i.e., indicators of condition). One or more
measurable endpoints are associated with each of the EEA elements listed in Table 4-1. These
assessment endpoints represent a formal expression of the environmental value that is to be
protected. If the assessment endpoint is to be used as a regulatory limit, it should be socially
relevant. Selection of a specific ecosystem for study will depend on the severity of the problem
of concern for a region. Time scales of response must be considered in selecting and evaluating
ecosystem response(s) to changes in atmospheric deposition. Responses of aquatic ecosystems
4-124
-------
to depositions can occur quickly. Surface water acidification associated with nitrate leaching
should respond to decreases in nitrogen loading in a short period of time. However, changes in
growth responses of vegetation resulting from soil nutrient imbalances may require years or
decades to detect. The focus of concern should be on the populations within an ecosystem that
are sensitive to nitrogen and sulfur deposition (Hunsaker et al., 1993).
Biogeochemical indicators for monitoring forest nitrogen status have been proposed by
Fenn and Poth (1998). Because nitrogen is a major constituent of all forms of life and is cycled
through a complex web of processes involving many biotic and abiotic mechanisms, evaluating
forest nitrogen status is a challenge. Indicators of ecosystems at risk of nitrogen saturation
should include those that can be identified when nitrogen availability exceeds biotic demand.
Such indicators typically should monitor parameters that are normally at background or low
levels in nitrogen-limited systems and should be those that commonly respond to excess nitrogen
in a wide range of ecosystems (Fenn and Poth, 1998). Such indicators include foliar nitrogen,
nutrient ratios (N:P, N:cation), foliar nitrate, foliar 615 N, arginine concentration, soil C:N ratio,
NO3 in soil extracts or in soil solution, and flux rates of nitrogenous trace gasses from soil
(Fenn et al., 1998). The cardinal indicator or manifestation of nitrogen saturation in all
ecosystem types, including California forests and chaparral, is increased and prolonged NO3
loss below the main rooting zone and in stream water. Seasonal patterns of stream water nitrate
concentrations are especially good indicators of watershed nitrogen status (Fenn and Poth,
1998).
In Europe, the elements used in the critical load concept include: a biological indicator,
a chemical criterion, and a critical value. The biological indicator is the organism used to
indicate the status of the receptor ecosystem; the chemical criterion is the parameter that results
in harm to the biological indicator; and the critical value is the value of the chemical criterion
below which no significant harmful response occurs to the biological indicator (L0kke et al.,
1996). Trees, and sometimes other plants, are used as biological indicators in the case of critical
loads for forests. The critical load calculation using the current methodology is essentially an
acidity/alkalinity mass balance calculation. The chemical criterion must be expressible in terms
of alkalinity. Initially, the Ca:Al ratio was used; but, recently, the (Ca+Mg+K):Al ratio has been
used (L0kke et al., 1996).
4-125
-------
Ideally, changes in acidic deposition should result in changes in the status of the biological
indicator used in the critical load calculation. However, the biological indicator is the integrated
response to a number of different stresses. Furthermore, some other organisms are more
sensitive to acid deposition than trees. At high concentrations, A13+ is known to be toxic to
plants, inhibiting root growth and, ultimately, plant growth and performance (L0kke et al., 1996;
National Science and Technology Council, 1998). Sensitivity to Al varies considerably between
species and within species because of changes in nutritional demands and physiological status
that are related to age and climate. Experiments have shown that there are large variations in Al
sensitivity, even among ecotypes.
Mycorrhizal fungi have been suggested as possible biological indicators by L0kke et al.
(1996) because they are intimately associated with tree roots, depend on plant assimilates, and
play an essential role in plant nutrient uptake, thereby influencing the ability of their host plants
to tolerate different anthropogenically generated stresses. Mycorrhizas and fine roots are an
extremely dynamic component of below-ground ecosystems and can respond rapidly to stress.
They have a relatively short lifespan, and their turnover appears to be strongly controlled by
environmental factors. Changes in mycorrhizal species composition, or the loss of dominant
mycorrhizal species in areas where diversity is already low, may lead to increased susceptibility
of plants to stress (L0kke et al., 1996). Stress affects the total amount of carbon fixed by plants
and modifies carbon allocation to biomass, symbionts, and secondary metabolites. The
physiology of carbon allocation has also been suggested as an indicator of anthropogenic stress
(Andersen and Rygiewicz, 1991). Because mycorrhizal fungi are dependent for their growth on
the supply of assimilates from the host plants, stresses that shift the allocation of carbon reserves
to the production of new leaves at the expense of supporting tissues will be reflected rapidly in
decreased fine root and mycorrhizal biomass (Winner and Atkinson, 1986). Decreased carbon
allocation to roots also affects soil carbon and rhizosphere organisms. Soil dwelling animals are
important for decomposition, soil aeration, and nutrient redistribution in the soil. They
contribute to decomposition and nutrient availability mainly by increasing the accessibility of
dead plant material to microorganisms. Earthworms decrease in abundance, and in species
number, in acidified soils (L0kke et al., 1996).
4-126
-------
Effects of Wet and Dry Deposition on Biogeochemical Cycling — the Integrated Forest
Study. The Integrated Forest Study (IPS; Johnson and Lindberg, 1992a) has provided the most
extensive data set available on wet and dry deposition, as well as on deposition effects on the
cycling of elements in forest ecosystems. The overall patterns of deposition and cycling have
been summarized by Johnson and Lindberg (1992a), and the reader is referred to that reference
for details. The following is a summary of particulate deposition, total deposition, and leaching
in the IPS sites.
Particulate deposition in the IPS was separated at the 2-jim level; however, the decision
was made to include total particulate deposition in this analysis and may include the deposition
of particles larger than 10 jim. Particulate deposition contributes considerably to the total impact
of base cations to most of the IPS sites. On average, particulate deposition contributes 47% to
total Ca deposition (range: 4 to 88%), 49% of total K deposition (range: 7 to 77%), 41% to total
Mg deposition (range: 20 to 88%), 36% to total Na deposition (range: 11 to 63%), and 43% to
total base cation deposition (range: 16 to 62%). Of total particulate deposition, the vast majority
(> 90%) is > 2 |im in size.
Figures 4-20 through 4-23 summarize the deposition and leaching of Ca, Mg, K, and total
base cations for the IPS sites. As noted in the original synthesis (Johnson and Lindberg, 1992a),
measurements indicated annual gains of base cations for some sites (i.e., total deposition >
leaching), some losses (total deposition < leaching) at others, and some approximately in
balance. Not all cations follow the same pattern at each site. For example, a net accumulation of
Ca occurs at the Coweeta, TN, Durham (Duke), NC and Bradford Forest, Fl sites (Figure 4-20),
whereas accumulation of K was noted at the Duke, Florida, Thompson, WA, Huntington Forest,
NY, and White Face Mountain, NY sites (Figure 4-22). Magnesium accumulated only at the
Florida site (Figure 4-21), and only at the Florida site is there a clear net accumulation of total
base cations (Figure 4-23).
As noted previously, the factors affecting net Ca accumulation or loss include the
soil-exchangeable cation composition; base cation deposition rate; the total leaching pressure
because of atmospheric sulfur and nitrogen inputs, as well as natural (carbonic and organic)
acids; and biological demand (especially for potassium). At the Florida site, which has a very
cation-poor, sandy soil (derived from marine sand), the combination of all these factors leads to
net base cation accumulation from atmospheric deposition (Johnson and Lindberg, 1992a).
4-127
-------
1 ,UUU
500-
i_
re
0) n
Equivalent/ha/y
° ~* ~* i
3 O1 "o Ui
D 0 0 0
0 0 0 0 C
1 1 1 1
63%
^^
Pe
41%
--
1
- -1
rcent
55%
..
^ Wet
^ Leaching
of To
65%
1 — 1
y/.
yy,
y/
ys,
Y/
-
tal De
40%
"
LL£
positi
26%
—
1
I
on as
25%
PI
V/,
y/,
i
y/,
///
v/
v/
Partk
49%
;ulate
58%
1
I
Matte
88%
?r:
4%
\
-
L .
r •
46°/
^
1
i
'o
CP DL GS LP FS
Warmer Sites
DF RA NS FF MS WF ST
H Colder Sites *•
Figure 4-20. Calcium deposition in > 2-um particles, < 2-um 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.
The site showing the greatest net base cation losses, the red alder stand in Washington state,
is under extreme leaching pressure by nitrate produced by excessive fixation by that species
(Van Miegroet and Cole, 1984). In the red spruce site in the Smokies, the combined effects of
SO42~ and NO3~ leaching are even greater than in the red alder site (Figure 4-24), but a
considerable proportion of the cations leached from this extremely acid soil consist of FT and
A13+ rather than of base cations (Johnson and Lindberg, 1992a). Thus, the red spruce site in the
Smokies is approximately in balance with respect to calcium and total base cations, despite the
very high leaching pressure at this site (Figures 4-20 and 4-23).
4-128
-------
1,000
400-
re 200
a>
re
0 -
c
"re -200
'5
iS" -4°° 4
-600 -
-800
40%
•1
Pe
49%
"1"
|< 2 |jm
• i! Wet
^ Leaching
srcenl
44%
-a-
of To
48%
-
1
i
tal De
36%
i
1
positi
27%
1
on as
27%
u _
1
1
I
Parti
39%
_
1
culate
46%
Matte
88%
f
2r:
20%
-
BH
I
26%
--
1
CP DL GS LP FS DF RA NS FF MS WF ST
Warmer Sites ^ 4 Colder Sites
Figure 4-21. Magnesium deposition in > 2-um particles, < 2-um particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-20 for site abbreviations.
4UU -
300-
,_ 200 -
CO
0>
.>• 100-
^ n -
+3 U
C
15 -100-
_>
0- -200 -
LLI
-300-
-400-
-500 -
77%
H
Percent of Total Deposition as Particulate Matter:
^ ^ ,__,,_
76%
^
I
_
,
> 2 pm
|< 2 |jm
H Wet
^ Leaching
61%
H
.1.
48%
Jii
J
CP DL GS LP
^ Wariripr *
54%
PI
40%
H
_____
40%
•\
68%
•
--
FS DF RA
>ite 2-um particles, < 2-um particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-20 for site abbreviations.
4-129
-------
0)
2,000-
1,000-
0-
-1,000-
-2,000-
-3,000-
53%
»
Pe
47%
1
~
- \ > 2 |jm
- H Wet
^ Leaching
srcent
48%
i
-
of To
62%
_
I
tal De
49%
!
positi
28%
-•
1
on as
28%
i
'///
v/
v/.
/^
--
Parti
47%
culate
44%
;Matt
62%
-i
er:
16%
|
31%
-
!
y//
-
CP DL GS LP FS DF RA NS FF MS WF ST
Warmer Sites H Colder Sites
Figure 4-23. Base cation deposition in > 2-um particles, < 2-um particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-20 for site abbreviation.
The relative importance of particulate-base-cation deposition varies widely with site and
cation and is not always related to the total deposition rate. The proportion of Ca deposition in
particulate form ranges from a low of 4% at the Whiteface Mountain site to a high of 88% at the
Maine site (Figure 4-20). The proportion of K deposition as particles ranges from 7% at the
Smokies site to 77% at the Coweeta site (Figure 4-22). The proportion of total base cation
deposition ranges from 16% at the Whiteface site to 62% at the Maine site (Figure 4-23).
Overall, parti culate deposition at the site in Maine accounted for the greatest proportion of Ca,
Mg, K, and base cation deposition (88, 88, 57, and 62%, respectively) even though total
deposition was relatively low. At some sites, the relative importance of parti culate deposition
varies considerably by cation. At the Whiteface Mountain site, paniculate deposition accounts
for 4, 20, and 40% of Ca, Mg, and K deposition, respectively. At the red spruce site in the
Smokies, parti culate deposition accounts for 46, 26%, 7% of Ca, Mg, and K deposition,
respectively.
4-130
-------
7,000-
6,000--
5,000--
re
4,000
^
"re 3,000-
P: 4%,
7%
Percent of Total Cation Leaching Balanced
by SO^'and NO^ from Particles (P) and Other (O) Sources
1% 8%, 9%, 10% 10% 8% 6% 18%
0:28%' 55%
6%
78%
30%' 42%
Other Anions
Participate Sulphur and Nitrogen
Other Sulphur and Nitrogen Sources
55%
73%
69%
1%
77%
CP DL GS LP FS DF RA NS FF MS WF ST
Figure 4-24. Total cation leaching (total height of bar) balanced by SO42 and NO3
estimated from particulate deposition (assuming no ecosystem retention,
particulate sulfur and nitrogen) and by other sources (both deposition and
internal) of SO42~ and NO3" (other sulfur and nitrogen sources) and by
other anions in the Integrated Forest Study sites. See Figure 4-20 for
site abbreviations.
As indicated in the IPS synthesis, SO42 and NO3 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 SO42 thus reducing SO42 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 4-24 shows the total cation leaching rates of the IPS sites
and the degree to which cation leaching is balanced by SO42 + NO3 deposition. The SO42 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 (i.e., wet and gaseous deposition, internal production).
4-131
-------
As noted in the IPS synthesis, total SO42 and NO3 inputs account for a large proportion
(28 to 88%) of total cation leaching in most sites. The exception is the Georgia loblolly pine site
where there were high rates of HCO3 and Cl" leaching (Johnson and Lindberg, 1992a). The role
of particulate sulfur and nitrogen deposition in this leaching is generally very small (< 10%),
however, even if it is assumed that there is no ecosystem sulfur or nitrogen retention.
It was previously noted in this chapter that the contribution of particles to total deposition
of nitrogen and sulfur at the IPS sites is lower than that for base cations. On average, particulate
deposition contributes 18% to total nitrogen deposition (range: 1 to 33%) and 17% to total
sulfur deposition (range: 1 to 30%). Particulate deposition contributes only a small amount to
total H+ deposition (average = 1%; range: 0 to 2%). (It should be noted, however,
that particulate H+ deposition in the > 2 jim fraction was not measured.)
Based on the IPS data, it appears that particulate deposition has a greater effect on base
cation inputs to soils than on base cation losses associated with the inputs of sulfur, nitrogen, and
H+. It cannot be determined what fraction of the mass of these particles are < 10 jim, but only a
very small fraction is < 2 jim. These inputs of base cations have considerable significance, not
only to the base cation status of these ecosystems, but also to the potential of incoming
precipitation to acidify or alkalize the soils in these ecosystems. As noted above, the potential of
precipitation to acidify or alkalize soils depends on the ratio of base cations to H+ in deposition,
rather than simply on the input of H+ alone. In the case of calcium, the term "lime potential" has
been applied to describe this ratio; the principle is the same with respect to Mg and K. Sodium
is a rather special case, in that it is a poorly absorbed cation and leaching tends to balance input
over a relatively short term.
Net balances of base cations tell only part of the story as to potential effects on soils; these
net losses or gains must be placed in the perspective of the soil pool size. One way to express
this perspective is to compare soil pool sizes with the net balances. This comparison is made for
exchangeable pools and net balances for a 25-year period as shown in Figures 4-25 to 4-27.
It is readily seen that the net leaching losses of cations pose no threat in terms of depleting
soil-exchangeable Ca2+, K+, or Mg+2 within 25 years at the Coweeta, Duke, Georgia, Oak Ridge,
or Douglas-fir sites. However, there is a potential for significant depletion at the red alder,
Whiteface Mountain, and Smokies red spruce sites.
4-132
-------
300,000
»_ 250,000
re
•jg1 200,000
.G
c 150,000
"re
.£ 100,000
3
LU
50,000
0
50 000
•D-
~
-
-
r ~
--
-
•
-
,fl.
-
CP DL GS LP FS
.4 \A/nrmnr Qitnc
-
,.y.
.=.
r —
| |Soil Exchangeable
H(Dep - Leaching) x 25
DF RA NS
k. ^ <
.0.
,-B-
.=..
HF MS WF ST
^nlrlnr 55itoc
Figure 4-25. Soil exchangeable Ca2+ pools and net annual export of Ca2+ (deposition minus
leaching times 25 years) in the Integrated Forest Study sites. See Figure 4-20
for site abbreviations.
80,000 -
re
>, 60,000 -
15
£
C 40,000 •
o>
re
=. 20,000 •
UJ
o •
_9n nnn .
fl.
_
i — i
-
-
u _
1 — |
--
1
£.
-U-
1
.=-
\^\ Soil Exchangeable
H(Dep - Leaching) x 25
-=.
*
....
B.
CP DL GS LP FS DF RA NS HF MS WF ST
Warmer Sites H Colder Sites
Figure 4-26. Soil exchangeable Mg2+ pools and net annual export of Mg2+ (deposition
minus leaching times 25 years) in the Integrated Forest Study sites.
See Figure 4-20 for site abbreviations.
4-133
-------
140,000-
120,000-
i_
(0
SJ, 100,000-
75
:£ 80,000 -
+j
c
U 60,000 -
CD
'= 40,000 -
CT
LU
20,000 -
0 -
.9n nnn .
:n:
:D:
r ~
I — I
^H
-
-
-
--=--
.D.
-B-
L _
Q Soil Exchangeable
H(Dep - Leaching) x 25
i
-a-
L.dL-
-fl-
CP DL
GS LP FS
Warmer Sites
DF RA NS HF MS WF ST
Colder Sites
Figure 4-27. Soil exchangeable K+ pools and net annual export of K+ (deposition minus
leaching times 25 years) in the Integrated Forest Study sites. See Figure 4-20
for site abbreviations.
The range of values for soil-exchangeable turnover is very large, reflecting variations in
both the size of the exchangeable pool and the net balance of the system. Soils with the highest
turnover rates are those most likely to experience changes in the shortest time interval, other
things being equal. Thus, the Whiteface Mountains, Smokies, and Maine red spruce sites, the
Thompson red alder site, and the Huntington Forest northern hardwood site appear to be most
sensitive to change. The actual rates, directions, and magnitudes of change that may occur in
these soils (if any) will depend on weathering inputs and vegetation outputs, in addition to
deposition and leaching. It is noteworthy that each of the sites listed above as sensitive has a
large store of weatherable minerals, whereas many of the other soils (with larger exchangeable
cation reserves) have a small store of weatherable minerals (e.g., Coweeta white pine,
Duke loblolly pine, Georgia loblolly pine, and Oak Ridge loblolly pine) (Johnson and Lindberg,
1992a; April and Newton, 1992).
Base cation inputs are especially important to the Smokies red spruce site because of
potential aluminum toxicity and calcium and magnesium deficiencies. Johnson et al. (199la)
4-134
-------
found that soil-solution Al concentrations occasionally reached levels found to inhibit Ca uptake
and cause changes in root morphology in solution culture studies of red spruce (Raynal et al.,
1990). In a follow-up study, Van Miegroet et al. (1993) found a slight, but significant, growth
response to Ca and Mg fertilizer in red spruce saplings near the Smokies red spruce site. Joslin
et al. (1992) reviewed soil and solution characteristics of red spruce in the southern
Appalachians, and it appears that the IPS site is rather typical of observations for that area.
Wesselink et al. (1995) reported on the complicated interactions among changing
deposition and soils at a beech and spruce forest in Solbing, Germany (including repeated
sampling of soil exchangeable base cation pools) from 1969 to 1991 and compared these results
with those of a simulation model. They identified three basic stages of change in this ecosystem.
During Stage I, there was increased deposition of sulfur and constant deposition of base cations,
causing increased base cation leaching and reduced base saturation in the soils. During Stage II,
sulfur deposition is reduced; soil solution SO42 and base cation leaching decline accordingly,
but base saturation continues to decrease. During Stage III, two alternative scenarios are
introduced: (a) sulfur deposition continues to decline, whereas base cation deposition says
constant; or (b) both sulfur and base cation deposition decline. Under Stage Ill-a, sulfate and
base cation leaching continue to decline, and base saturation begins to increase as base cations
displace exchangeable Al and cause it to transfer to the gibbsite pool. Under Stage Ill-b, this
recovery in base saturation is overridden by the reduction in base cation deposition.
The IPS project, for the first time, accurately quantified atmospheric deposition inputs to
nutrient cycles using state-of-the-art techniques to measure wet and dry deposition. The main
aim of the project was to determine effects of atmospheric deposition on nutrient status of a
variety of forest ecosystems and to determine if these effects are in any way related to current or
potential forest decline. Acidic deposition appears to be having a significant effect on nutrient
cycling in most of the forest ecosystems studied in the IPS project. The exceptions were the
relatively unpolluted Douglas fir, red alder, and Findley Lakes sites in Washington state. The
nature of the effects, however, varies from one location to another (Johnson, 1992). In all but the
relatively unpolluted Washington sites, atmospheric deposition was having a significant (often
overwhelming) effect on cation leaching from the soils. In general, nutrient budget data from
IPS and the literature suggest that the susceptibility of southeastern sites to base cation depletion
4-135
-------
from soils and the development of cation deficiencies by that mechanism appears to be greater
than in northern sites (Johnson, 1992).
Atmospheric deposition may have significantly affected the nutrient status of some US
sites through the mobilization of Al. Soil-solution Al levels in the Smokies sites approach and
sometimes exceed levels noted to impede cation uptake in solution culture studies. It is therefore
possible that the rates of base cation uptake and cycling in these sites have been reduced by soil
solution Al levels. To the extent that atmospheric deposition has contributed to these elevated
soil-solution Al levels, it has likely caused a reduction in base cation uptake and cycling rates at
these sites. Nitrate and sulfate are the dominant anions in the Smokies sites, and nitrate pulses
are the major cause of Al pulses in soil solution (Johnson, 1992). The connection between Al
mobilization and forest decline is not clear. The decline in red spruce has been more severe in
the Northeast than in the Southeast; yet all evidence indicates that Al mobilization is most
pronounced in the southern Appalachians. However, the Whiteface Mountain site which was
selected for study because it was in a state of decline had soil solution levels lower than in the
Smokies, which are not in a visibly obvious state of decline (there was no dieback other than the
fir killed by the balsam wooly adelgid and no needle yellowing). Thus, Al mobilization
continues to be worthy of further study (Johnson, 1992).
The simple calculations shown above give some idea of the importance of particulate
deposition in these forest ecosystems, but they cannot account for the numerous potential
feedbacks between vegetation and soils nor for the dynamics through time that can influence the
ultimate response. One way to examine some of these interactions and dynamics is to use
simulation modeling. The nutrient cycling model (NuCM) has been developed specifically for
this purpose and has been used to explore the effects of atmospheric deposition, fertilization, and
harvesting on some of the IPS sites (Johnson et al., 1993). The NuCM model is a stand-level
model that incorporates all major nutrient cycling processes (uptake, translocation, leaching,
weathering, organic matter decay, and accumulation).
Johnson et al. (1999) used the NuCM model to simulate the effects of reduced S, N, and
base cation deposition on nutrient pools, fluxes, soil, and soil solution chemistry in two
contrasting southern Appalachian forest ecosystems: the red spruce and Coweeta hardwood
sites from the IPS project. The scenarios chosen for these simulations included no change; 50%
4-136
-------
N and S deposition; 50% base cation deposition; and 50% N, S, and base cation deposition (50%
N, S, base cation). The NuCM simulations suggested that, for the extremely acid red spruce site,
S and N deposition is the major factor affecting soil-solution Al concentrations and base cation
deposition is the major factor affecting soil solution base cation concentrations. The effects of
S and N deposition were largely through changes in soil solution SO42 and NO3 and,
consequently, mineral acid anion (MAA) concentrations rather than through changes in soils.
This is illustrated in Figures 4-28a,b and 4-29a,b, which show simulated soil-solution mineral-
acid anions, base cations, Al, and soil base saturation in the B horizon in the red spruce site.
The 50% S and N scenario caused reductions in soil solution SO42 , NO3 and, therefore,
MAA concentrations, as expected. This, in turn, caused short-term reductions in base cation
concentrations. However, by the end of the 24-year simulation, base cations in the 50% S,
N scenario were nearly as high as in the no-change scenario because base saturation had
increased and the proportion of cations as Al decreased. The 50% base cation scenario had
virtually no effect on soil solution SO42 , NO3 and, therefore, MAA concentrations, as expected,
but did cause a long-term reduction in base cation concentrations. This was caused by a long-
term reduction in base saturation (Figure 4-29a,b). Thus, the effects of base cation deposition
were solely through changes in soils rather than through changes in soil solution MAA, as
postulated by Driscoll et al. (1989). In the less acidic Coweeta soil, base saturation was high and
little affected by the scenario cited above; Al was unimportant; and S and N deposition had a
much greater effect than base cation deposition in all respects (Figure 4-30a,b).
In summary, Johnson et al. (1999) found that the results of the red spruce simulations, in
part, supported the hypothesis of Driscoll et al. (1989) in that: base cation deposition can have a
major effect on base cation leaching through time in an extremely acid system. This effect
occurred through changes in the soil exchanger and not through changes in soil solution MAA
concentration. On the other hand, S and N deposition had a major effect on Al leaching at the
Noland Divide site. This occurred primarily because of changes in soil solution MAA
concentration. At the less acidic Coweeta site, base cation deposition had a minor effect on soils
and soil solutions; whereas S and N deposition had delayed but major effects on base cation
leaching because of changes in SO42 and MAA concentrations.
4-137
-------
oo
Red Spruce
_o
O
E
0)
D)
ro
O
M-
o
500 —
400 —
300
.O 200 —
"+->
2
+j
c
g 100-
o
O
0-
Mineral Acid Anions
i A»
V i v \A i *V\k I ^Ul
V H j V\, | s
— No Change
...... 50% N,S
-- 50% BC
i
4
\
6
i
8
i i i i i i i
10 12 14 16 18 20 22 24
Year
Figure 4-28a. Simulated soil solution mineral acid anions in the red spruce site with no change, 50% N and
S deposition, and 50% base cation deposition.
Source: Redrawn from Johnson et al. (1999).
-------
-j^
•^o
200-
j*
o
O
£ 160.
O
!i 120^
o
o
c
o
O
c
o
o
80-
40-
Q.
Red Spruce
Base Cations
|! §1 fit J*t li Ift ft.
j :• " \ j 1 J; ,\ fi/J .ll^ *i|j fy ! \/| 1 \ jfV I Vfl "•% f* I |,*;^ *^Jlsj I \ J|l . I /\ |F ^ ^
\M • i I ^ i \^ I ^ li \ e sH /VB § \ N 1^ / ^ 9 \ P-1 /I w ^ i «I / 4 * f =/lI r^
•:" ^ V^ ^ " ^> / \\tf M ^^ iil 4 ^ ^ ^« " 1 ;J i V p i 1 i *
— No Change
50% N,S
— 50% BC
i
2
n^
6
i^
8
10
12
Year
14
16
18
20 22
24
Figure 4-28b. Simulated soil solution base cations in the red spruce site with no change, 50% N and S deposition,
and 50% base cation deposition.
Source: Redrawn from Johnson et al. (1999).
-------
Red Spruce
4^
O
^ 500
_p
O
E
•3 400
Q)
E1
ii
o
M-
o
300 —
_O 200
"J
(0
o
c
o
O
Aluminum
i
«
11 s~ ~ „ „. „
I \ f| f™ | fl^fSF ii I I
ii
r *
II ill; H *#L *M^ fl 9wI
"^J II || .| ||§fjR 1 i| a|l?/^g. 1 L! L'| ifH^/1^! tl ftft^^l'MI L! Ijil^'l'll I'i 1
-W
^v
— No Change
50% N,S
~ 50% BC
I
2
I
6
I
8
10
12 14
Year
16 18 20 22 24
Figure 4-29a. Simulated soil solution Al in the red spruce site with no change, 50% N and S deposition, and
50% base cation deposition.
Source: Redrawn from Johnson et al. (1999).
-------
_o
O 5.
E 5
D)
TO
O
M-
o
O
c
O
O
— No Change
• 50% N,S
— 50% BC
i
2
i
4
Red Spruce
A I A 1 I A 1 I A /
Is $ al T A 1 I A ,V
^i IAI ls i t \\/f i h
I I!A?**: * i I' i ' j?
11IM // ; / \ / vf
'I I/ i '/ t'vj I I V
f J v V
Base Saturation
II *• it I"- ; '
I;«. /A//
4 \! I V /i */ f/I I/
* l i * B £
' V \/ '' /> Ay A ;' l / '* A '
w \ i ' i / ' i / v /
,1 <^ /\ / V/ l i VJ »J ' / I/
A /V 'A\'/l A1' I/
V
i i i i i i i i
8 10 12 14 16 18 20 22 24
Year
Figure 4-29b. Simulated soil solution soil base saturation in the red spruce site with no change, 50% N
and S deposition, and 50% base cation deposition.
Source: Redrawn from Johnson et al. (1999).
-------
-^
to
Coweeta
QO-
O
E
O)
(0
O
s—
O
c
O
*3 40 H
O
U
C
O
O
Mineral Acid Anions
« »
I 1 I 3 il Si 11 il 111*1
I 1 1 II li II. i i 1 1 II
i
»
20
Ii
I
»*
i
II
*1
1
I « il «l I' »i i « « 1 4 Pf IS »B il
i ii ii ji ii i!*iilsp|H is 5*1 t
1 Jl il il il lillirlltia 111
I It * I I ' ' I J • I i e* i I Si • IP * 11 ! ?' '' 5
1 I I Is '. i ii i f I I * ! «4 1 IS i 11 S 7 f « V I li
IliiNn /
I Hi iSfs| ilfii :| § : I : ; i ; : ; : i : IJ « • «i i ig : ::
Iff ill W i * ! M ii !,! hi !Si ill S I/! I!' i/'
ft '11 i 1 i ; / i ,i Ui ii« ifl : ? ui u: u u •/
if ii n 11 " i,i u Hi i* is in u1 y v v v v •
llflhlHimn ? ' ' '
i/i i/! y v * f
— No Change
50% N,S
-50%BC
8
10
12
Year
14
16
18 20 22 24
Figure 4-30a. Simulated soil solution mineral acid anions in the Coweeta site with no change, 50% N and
S deposition, and 50% base cation deposition.
Source: Redrawn from Johnson et al. (1999).
-------
100
_o
o
E
CC
O
*o
c
o
+3
03
O
E
O
O
80-
60-
40
Coweeta
Base Cations
i
i
1! |
i (I
'\ P I
1 i IS iv
',! I si. i v
. ,-B uy
^h' i/l VH n
US Iff :
R- *1. il *l I * il ™ f I * I 1 J| 1 I 5. $ i I" l»f llaii IHiff If if I i*p* I
! I 11 It i 1 M * I ft i ii i~rt MfH !«M iri^ iwfc i iji * S/i * ;i $! fi -r: ^. !i/ p,
il fi 1 PI *= 1 Pi €;: 1 «*;: liSlii! ilflll:"; pafl-i |[ S'll ' l/ll flili /:! I*/-! \X*:s\/l'{! U '1
•• I P * P1 i1^ ft «11 ly^i^'* pifl^i "^M • - &•' '-i Sn i Wlv I f 111 ' n, \n fv. W |x YII ll "tf r* H i\
1«! \1 ,j v:| A{ iLfl jjStj let I Ivi! l^li 1 pi i i! I i! 1 If Ml i ^ I M ! I j l M M M
IHJ jV|j ri| i ji j if PI i 'i M i i i M /1 11 M M M M / \
\ || i |i ! J| j / ! / I / i I ! j I j i j i j! 1 ji \ || I!/ \ sj I j/ ill i ^j |
— No Change
•••••• 50% N,S
-50%BC
1
2
i
4
6
8
10
12 14
Year
16
18
20
22 24
Figure 4-30b. Simulated soil solution base cations in the Coweeta site with no change, 50% N and S deposition,
and 50% base cation deposition.
Source: Redrawn from Johnson et al. (1999).
-------
Effects of Trace Elements
Trace metals are natural elements that are ubiquitous in small (trace) amounts in soils,
ground water, and vegetation. Many are essential elements required for growth by plants and
animals as micronutrients. Naturally occurring surface mineralizations can produce metal
concentrations in soils and vegetation that are as high, or higher, than those in the air and
deposited near man-made sources (Freedman and Hutchinson, 1981). The occurrence and
concentration of trace metals in any ecosystem component depend on the sources of the metal
(i.e., via the soil or as a particulate). Even when air pollution is the primary source, continued
deposition can result in the accumulation of trace metals in the soil (Martin and Coughtrey,
1981). Many metals are deposited and bound into soils by chemical processes and are not
available to plants (Saunders and Godzik, 1986).
When aerial deposition is the primary source of metal particles, both the chemical form and
particle size deposited determine the heavy metal concentration in the various ecosystem
components (Martin and Coughtrey, 1981). Human activities introduce heavy metals into the
atmosphere and have resulted in the deposition of antimony, cadmium, chromium, copper, lead,
molybdenum, nickel, silver, tin, vanadium, and zinc (Smith, 1990c). Extensive evidence
indicates that heavy metals deposited from the atmosphere to forests accumulate either in the
richly organic forest floor, or in the soil layers immediately below the floor, areas where the
interaction of roots and soil is greatest. The greater the depth of soil, the lower the metal
concentration. The accumulation of metal in the soil layers where the biological activity is
greatest, therefore, has the potential to be toxic to roots and soil organisms and to interfere with
nutrient cycling (Smith, 1990c). The shallow-rooted plant species are most likely to take up
metals from the soil (Martin and Coughtrey, 1981). Though all metals can be directly toxic at
high levels, only copper, nickel, and zinc have been documented as frequently being toxic.
Toxicity due to cadmium, cobalt, and lead has been seen only under unusual conditions (Smith,
1990c). Exposures to the above metals at lower concentrations have the potential, over the long
term, to interfere with the nutrient-cycling processes when they affect mycorrhizal function.
Biological accumulation of metals through the plant-herbivore and litter-detrivore chains
can occur. A study of Cd, Pb, and Zn concentrations in earthworms suggested that Cd and Zn
were bioaccumulated in earthworms, but Pb was not. Studies indicate that heavy metal
deposition onto the soil, via food chain accumulation, can lead to excessive levels and toxic
4-144
-------
effects in certain animals. Cadmium appears to be relatively mobile within terrestrial food
chains; however, the subsequent mobility of any metal after it is ingested by a herbivorous
animal depends on the site of accumulation within body tissues. Although food chain
accumulation of a metal may not in itself cause death, it can reduce the breeding potential in a
population (Martin and Coughtrey, 1981).
Case studies have revealed that the deposition of Cu and Zn particles around a brassworks
can result in an accumulation of incompletely decomposed litter. In one study, litter
accumulation was reported up to 7.4 km from the stack of a primary smelter in southeastern
Missouri. Similar results were reported around a metal smelter at Avonmouth, England. In the
latter case, litter accumulation was associated closely with concentrations specifically of Cd, as
well as with those of Pb, Cu, and Zn (Martin and Coughtrey, 1981). Accumulations of metals
emitted in PM were also reported in soil litter close to a metal smelter at Palmerton, PA in 1975
and 1978. The continued presence of Cd, Pb, Zn, and Cu in the upper soil horizons (layers) were
observed 6 years after the smelter terminated operation in 1980. Metal levels were highest near
the smelter. Experimental data (using mesh bags containing litter) supports the hypothesis that
reduced decomposition occurs close to heavy metal sources.
Accumulation of heavy metals in litter presents the greatest potential for interference with
nutrient cycling. Microorganisms are essential for the decomposition of organic matter and soil
fertility. Toxic effects on the microflora can be caused by Zn, Cd, and Cu. Addition of a few
mg/kg of soil of Zn can inhibit the more sensitive microbial processes (van Beelen and Doelman,
1997). Experiments by Kandeler et al. (1996) indicated that microbial biomass and enzyme
activities decreased with increasing heavy metal concentrations. The amount of decrease varied
among the enzymes, with those involved in carbon cycling being least affected, the activities of
the enzymes involved in the cycling of N, P, and S (especially arylsulfatase and phosphatase)
were dramatically affected.
Metal accumulation in litter can be found mainly around brass works and Pb and Zn
smelters. Invertebrates inhabiting soil litter may also accumulate metals. Earthworms from
roadsides were shown to contain elevated concentrations of Cd, Ni, Pb, and Zn; but, interference
with earthworm activity was not cited (Martin and Coughtrey, 1981). It has been shown,
however, that when soils are acidic, earthworm abundance decreases, and bioaccumulation of
metals from soil may increase exponentially with decreasing pH (L0kke et al, 1996). Organisms
4-145
-------
that feed on earthworms living in soils with elevated levels of Cd, Ni, Pb, and Zn for extended
periods may accumulate Pb and Zn to toxic levels (Martin and Coughtrey, 1981). Increased
concentrations of heavy metals have been found in a variety of small mammals living in areas
with elevated heavy metal concentrations in the soils. Furthermore, increased amounts of metal
in body tissues were seen in both amphibians and mammals. Levels of Cd in the kidneys and
liver of white-tailed deer (Odocoileus virginaus) were five times higher at Palmerton, PA than in
those from 180 km southwest (downwind). The abnormal amounts of metal in the tissues of
terrestrial vertebrates and the absence or low abundance of wildlife at Palmerton indicated that
ecological processes within 5 km of the Zn smelter continued to be markedly influenced even
6 years after its closing (Storm et al., 1994).
Studies by Babich and Stotsky (1978) support the concept that increased accumulation of
litter in metal-contaminated areas is due to the effects on the microorganismal populations.
Cadmium toxicity to microbial populations was observed to decrease and prolong logarithmic
rates of microbial population increase, to reduce microbial respiration and fungal spore
formation and germination, to inhibit bacterial transformation, and to induce abnormal
morphologies. Smith (1991) reported the effects of Cd, Cu, Ni, and Zn on the symbiotic activity
of fungi, bacteria, and actinomycetes. The formation of mycorrhizae by Glomus mosseae with
onions was reduced when Zn, Cu, Ni, and Cd was added to the soil. The relationship of the
fungus with white clover, however, was not changed. It was suggested that the effect of heavy
metals on vesicular-arbuscular mycorrhizal fungi will vary from host to host (Gildon and Tinker,
1983). Studies with ericoid plants indicated that, in addition to Calluna vulgaris, mycorrhizae
also protect Vaccinium macrocarpa and Rhodendronponticum from heavy metals (Bradley
et al., 1981). Heavy metals tend to accumulate in the roots, thereby lessening shoot toxicity.
The effects of sulfur deposition on litter decomposition in the vicinity of smelters must also
be considered. Metal smelters emit SO2 as well as heavy metals. Altered litter decomposition
rates have been well documented near SO2 sources (Prescott and Parkinson, 1985). The
presence of sulfur in litter has been associated with reduced microbial activity (Bewley and
Parkinson, 1984). In addition, the effects of sulfur on the symbiotic activity of fungi, bacteria,
and actinomycetes were reported by Smith (1990d).
The potential pathways of accumulation of trace metals in terrestrial ecosystems, as well as
the possible consequences of trace metal deposition on ecosystem functions, is summarized in
4-146
-------
Figure 4-31. The generalized trophic levels found in an ecosystem and the various physiological
and biological processes that could be affected by trace metals are shown in the figure.
Reduction in physiological processes can affect productivity, fecundity, and mortality (Martin
and Coughtrey, 1981). Therefore, any effects on the structure and function of an ecosystem are
likely to occur through the soil and litter (Tyler, 1972).
9. Retranslocation
1. Wet/Dry Deposition
3. Litterfall, Resuspension,
Deposition, Leaching,
Stem Flow
Above-
Ground
Storage,
Metabolism
Biologically
Unavailable
Figure 4-31. Relationship of plant nutrients and trace metals with vegetation.
Compartments (roman numerals) represent potential storage sites;
whereas arrows (arabic numerals) represent potential transfer routes.
Certain species of plants are tolerant of metal-contaminated soils (e.g., soils from mining
activities) (Antonovics et al., 1971). Certain species of plants also have been used as
bioindicators of metals (e.g., Astragalus is an accumulator of selenium). The sources of both
4-147
-------
macroelements and trace metals in the soil of the Botanical Garden of the town of Wroclow,
Poland, were determined by measuring the concentrations of the metals in Rhododendron
catawbiense, Ilex aquifolium, andMahonia aquifolium growing in a garden and comparing the
results with the same plant species growing in two botanical gardens in nonpolluted areas.
Air pollution deposition was determined to be the source of metals in plants rather than soils
contamination (Samecka-Cymerman and Kempers, 1999).
The effects of Pb in ecosystems are discussed in the EPA document Air Quality Criteria
for Lead (U.S. Environmental Protection Agency, 1986). Studies have shown that there is cause
for concern in three areas in which ecosystems may be extremely sensitive to Pb: (1) delay of
decomposition because of inhibition by Pb of the activity of some decomposer microorganisms
and invertebrates, (2) subtle shifts toward Pb-tolerant plant populations, and (3) Pb in the soil
and on the surfaces of vegetation where it may circumvent the processes of biopurification. The
problems cited above arise because Pb is deposited on the surface of vegetation, accumulates in
the soil, and is not removed by the surface and ground water of the ecosystem (U.S.
Environmental Protection Agency, 1986).
4.2.4 Urban Ecosystems
Humans dominate Earth's ecosystems. Evidence accumulating from anthropological and
archeological research indicates that human influence has been pervasive for thousands of years
(Grimm et al., 2000). Major human effects on the environment probably began as early as
12,000 to 15,000 years ago and continue to be a major influence on all natural ecosystems.
Human activities have been more intense in cities, suburbs, exurbs and in supporting hinterlands
(Grimm et al., 2000). Interest has increased recently in the study of urban ecological systems.
Vitousek et al. (1997) pointed out that understanding a human-dominated planet requires
that the human dimensions of global change; that is, the social, cultural, and other drivers of
human actions need to be included in ecological analyses. Therefore, humans must be integrated
into models for more complete understanding of extant ecological systems.
In the past, ecological plant or animal studies conducted in urban settings used traditional
ecological approaches and considered humans as agents of disturbance. Although the term
"urban ecosystem" has been used to describe human-dominated ecosystems, it does not
adequately take into account the developmental history, sphere of influence, and potential
4-148
-------
impacts required to understand the true nature of an urban ecosystem (Mclntyre, et al., 2000).
Because urbanization is both an ecological and a social phenomenon, urban ecology implicitly
recognizes the role humans play in developing unique systems. Therefore, if urban ecology is to
be a truly interdisciplinary field, both social and natural sciences must be integrated into the
study of urban ecosystems (Mclntyre, et al., 2000).
Although the study of ecological phenomena in urban environments is not a new area of
science, the concept of the city as an ecosystem is relatively new for the field of ecology (Grimm
et al., 2000). There is a wealth of information on the terrestrial components of urban ecological
systems. However, much of it is organized from the perspective of ecology in cities while the
more comprehensive perspective identified as ecology of cities is needed (Pickett et al., 2001).
The basic questions addressed by the literature of ecology in cities relate to how ecological
patterns and processes differ in cities as compared with other environments. What is the effect
of the city (i.e., a concentration of human population and activities) on the ecology of organisms
inside and outside its boundary and influence? The concept of ecology of cities has to do with
how aggregated parts make up the whole, i.e., how cities process energy or matter relative to
their surroundings (Grimm et al., 2000). The latter concept includes primary production, species
richness, biogeophysical budgets, ecosystem patterns and processes, and an open definition of
urban ecosystems that incorporates the exchanges of materials and influence between cities and
surrounding landscapes (Pickett et al., 2001). If ecosystems are to be understood, there is a need
for a new integrative ecology that explicitly incorporates human decisions, culture, institutions,
and economic systems (Grimm et al., 2000). This makes an ecological approach to land use
planning essential to maintain long-term sustainability of ecosystem benefits, services, and
resources (Zipperer et al., 2000). The ecological and social effects of "edge city" need to be
studied as they may be greater than the previous patterns of suburbanization. The classical
ecosystem approach and a patch dynamic approach are needed to understand and manage the
dynamics of urban and urbanizing ecosystems (Zipperer et al., 2000).
There has been little work on the rates of atmospheric deposition to urban ecosystems
despite extensive data on concentrations and chemical reactions of air pollutants in cities.
A search of the literature produced no references that dealt with the effects of PM deposition.
Lovett et al. (2000), however, reported that urban ecosystems are likely to be subjected to large
rates of deposition of anthropogenic pollutants. Decades of research on urban air quality
4-149
-------
indicate that cities are often sources of NOX, SOX, and dust, among many other pollutants. Some
of these air pollutants are major plant nutrients (e.g., nitrogen) and may be affecting nutrient
cycles in plant-dominated areas in and around cities, but have not been studied. The gases and
particles in urban air can increase atmospheric deposition within and downwind of the city.
Studying deposition rates of atmospheric pollutants in urban areas can provide quantitative
estimates of amounts of gaseous and particulate air pollutants removed by urban vegetation.
To determine the patterns of atmospheric deposition and throughfall in the vicinity of a
large city, Lovett et al. (2000) measured bulk deposition, oak forest throughfall, and particulate
dust at sites along a transect within and to the north of New York City. They observed that
concentrations and fluxes of NO3 , NH4+, Ca2+, Mg2+, SO42 , and Cl" in throughfall all declined
significantly with increasing distance from the city, while H+ concentration and flux increased
significantly with increasing distance from the city. Most of the change in concentrations and
fluxes occurred within 45 km of the city. Additionally, it was observed that throughfall nitrogen
was twice as high in the urban areas compared with suburban and rural areas. Most of the dry
deposition of nitrate was from gaseous nitrogen oxides. As mentioned earlier, the effects of the
atmospheric deposition of the particulate pollutants was not mentioned.
McDonnell et al. (1997) in a 10-year study of ecosystem processes along an urban-rural
gradient included plant litter dynamics and nitrogen cycling of two key components of a forest
ecosystem: litter decomposition and heavy metal levels in soil and foliar litter. Foliar litter
decomposition integrates many features of the abiotic and biotic environment. It is an important
site of heavy metal incorporation into ecosystems and provides both a habitat and a resource for
fungi, bacteria, and invertebrates. Litter decomposition integrates the effects of resource quality,
environmental factors, and activities of decomposer organisms on nutrient cycling and serves as
an easily measured indicator of the effect of urbanization on an important ecosystem function.
McDonnell et al. (1997) noted that levels of heavy metals were higher in the foliar litter of urban
forest soils than in rural forest soils. The levels in urban forest stands approached or exceeded
the levels reported to affect soil invertebrates, macrofungi, and soil microbial processes. The
urban forests exhibited reduced fungal biomass and microarthropod densities when compared to
rural stands. These results supported the concept that urban forests have depauperate
communities because of anthropogenic stress resulting from poor air quality due to high levels of
SO2, SO42 , O3 and NO3 ; elevated levels of soil- and forest-floor-heavy metals; and low water
4-150
-------
availability such as those caused by hydrophobic soils (McDonnell et al., 1997). Thus, forests at
the urban end of the gradient exhibited reduced fungal and microarthropod populations and
poorer leaf quality than the more rural forests. The potential effect of these conditions on the
ecosystem processes of decomposition and nitrogen cycling in urban forests appeared to be
ameliorated by two other anthropogenic factors: increased average temperatures caused by the
heat island effect and the introduction and successful colonization of earthworms in the urban
forests (McDonnell et al., 1997).
McDonnell et al. (1997) observed that the changes in forest nitrogen dynamics were related
to increased anthropogenic nitrogen deposition in an urban environment. The studies of Aber
et al. (1989) in the northeastern United States on forest nitrogen dynamics demonstrated that
elevated nitrogen deposition over many years results in increased nitrification and the
mineralization of more nitrogen than can be taken up by plants and microorganisms.
Nitrification can lead to decreases in fine root biomass and increases in nitrate leaching below
the root zone. These effects of nitrogen deposition were not related to inputs from a specific
source such as PM.
While studies of heavy metal deposition in or near cities have been performed, they do not
cite the effects of metals in the soil. Pouyat and McDonnell (1991) discussed heavy metal
accumulations in forest soils along an urban-rural gradient in southeastern New York.
Variations in the amounts of Zn, Cu, Ni, and Cd appeared to be indicative of a pattern of
atmospheric deposition near point sources (Section 4.3.2.6). The concentrations of heavy metals
in forest floor and soils corresponded closely with the urban-rural land use gradient. Again, as in
the study by Lovett et al. (2000), the pollutants were highest near the urban end of the gradient
and declined toward rural sites, with Pb, Ni and Cu highest near the urban end.
The air quality of the region around East St. Louis has been of concern due to industries in
the area (Kaminski and Landsberger, 2000a), which include ferrous and nonferrous metal
smelters (Pb, Zn, Cu, and Al), coal-fired power plants, producers of organic and inorganic
chemicals, municipal waste incinerators, and petroleum refineries. The city is also in the path of
diverse plumes from refineries to the north, coal-fired power plants to the west, and nonferrous
smelters to the south. Concentrations of heavy metals and metalloids (As, Cd, Cu, Hg, Pb, Sb,
Zn) in the soil provided a basis for analysis (Kaminski and Landsberger, 2000b). These studies
of the extent of long-term metal deposition on the soil surface and depth of soil contamination,
4-151
-------
as well as the leaching dynamics of heavy metals were made to determine possible effects on
biota uptake or groundwater contamination. The effects on biota are not mentioned; however,
the soils in the area acted as a sink and there was little groundwater mobility (Kaminski and
Landsberger, 2000b).
4.3 AIRBORNE PARTICLE EFFECTS ON VISIBILITY
4.3.1 Introduction
Visibility may be thought of as the degree to which the atmosphere is transparent to visible
light (National Research Council, 1993). The beauty of scenic vistas in many parts of the United
States is often diminished by haze that reduces contrast, washes out colors, and renders distant
landscape features indistinct or invisible. This degradation of visibility is due primarily to the
scattering and absorption of light by fine particles suspended in the atmosphere. One
quantitative measure of visibility, used traditionally by meteorologists, is the visual range,
defined as the farthest distance at which a large black object can be distinguished against the
horizon sky (U.S. Environmental Protection Agency, 1979).
In August 1977, Congress amended the Clean Air Act (CAA) to establish as a national goal
"the prevention of any future and remedying of any existing impairment of visibility in
mandatory Class I Federal areas (many national parks and wilderness areas), which impairment
results from manmade air pollution" (Title I Part C Section 169A, U.S. Code [1990]). The 1977
Amendments also included provisions requiring applicants for new major source permits to
assess the potential for their projects to cause adverse effects on air quality-related values (e.g.,
visibility) in nearby Class I areas. In 1980, the EPA established regulatory requirements under
Section 169A to address Class I protection from "reasonably attributable" visibility impairment,
i.e., visibility impairment attributable to a single source or small group of sources.
The CAA, as amended in 1990 (Section 169B), required EPA both to conduct research on
regional visibility impairment and to establish the Grand Canyon Visibility Transport
Commission (GCVTC). The GCVTC was charged with assessing and providing advice to help
preserve clear days and to improve visibility in the 16 national parks and wilderness areas
located on the Colorado Plateau. The GCVTC also was mandated to provide recommendations
to the U.S. EPA for reduction of visibility impairment due to regional haze, described as any
4-152
-------
perceivable change in visibility (light extinction, visual range, contrast, or coloration) from that
which would have existed under natural conditions and is caused predominantly by a
combination of many anthropogenic sources over a wide geographical area (U.S. EPA, 1999a).
In July 1999, EPA published the Regional Haze Rule (Federal Register, 1999). The regulation
established a program for the improvement and protection of visibility in the 156 protected Class
I parks and wilderness areas and included the establishment of baseline and current visibility
conditions and the tracking of changes in visibility conditions over time. Implementation of the
regional haze regulations is supported by EPA's PM25 monitoring network and an expanded
Interagency Monitoring of Protected Visual Environments (IMPROVE) network. The PM25
monitoring network and the IMPROVE network are described briefly later in this section and in
more detail elsewhere (National Park Service, 1998; Evans and Pitchford, 1991; U.S.
Environmental Protection Agency, 2000b; U.S. Environmental Protection Agency, 2001a).
The objective of the visibility discussion in this section is to provide a brief description of
the fundamentals of atmospheric visibility and to summarize the linkage between PM and
visibility. Visibility is affected by air quality and, unlike the PM concentration, is not a property
of an element of volume in the atmosphere. Visibility can be quantified only for a sight path and
depends on the illumination of the atmosphere and the direction of view. However, the
concentration of particles in the atmosphere plays a key role in determining visibility.
Therefore, visibility impairment may be controlled by control of particle concentrations. The
relationships between particles, other factors, and visibility impairment are described in this
section. For a more detailed discussion on visibility, the reader is referred to the 1996 Air
Quality Criteria for Paniculate Matter (U.S. Environmental Protection Agency, 1996a); the
Recommendations of the Grand Canyon Visibility Transport Commission (Grand Canyon
Visibility Transport Commission, 1996); the National Research Council (National Research
Council, 1993); the National Acid Precipitation Assessment Program (Trijonis et al., 1991);
Interim Findings on the Status of Visibility Research (U.S. Environmental Protection Agency,
1995a); Visibility: Science and Regulation (Watson, 2002), and reports summarizing visibility
science and data from the IMPROVE visibility monitoring network (Malm, 2000; Sisler, 1996;
Sisleretal., 1993).
4-153
-------
4.3.2 Factors Affecting Atmospheric Visibility
The visual perception of a distant object is influenced by a large number of factors
including human vision (the eye), the brain's response to signals received from the eye, the
interaction of light with the atmosphere (e.g., atmospheric illumination, path and transmitted
radiance, contrast, and optical properties), and atmospheric pollution from natural and
anthropogenic sources. Detailed discussion of this full range of topics can be found in the 1996
PM AQCD (U.S. Environmental Protection Agency, 1996a) and other general references
(Malm, 1999; Watson, 2002). This section focuses only on those topics that have been
addressed by more recent research, including atmospheric illumination, the optical properties of
gases and particles in the atmosphere, and the effects of relative humidity on the optical
properties of particles.
4.3.2.1 Optical Properties of the Atmosphere and Atmospheric Particles
Atmospheric particles and gases attenuate image-forming light as it travels from a viewed
object to an observer. The fractional attenuation of light per unit distance is known as the light
extinction coefficient. The light extinction coefficient, bext, is expressed in units of one over
length, for example inverse kilometers (km"1) or inverse megameters (Mm"1). The light
extinction coefficient can be expressed as the sum of the light scattering and light absorption
coefficients of particles and gases:
where the subscripts/? and g signify particles and gases, and s and a signify scattering and
absorption.
The light extinction coefficient can be measured with a reasonable degree of accuracy or
can be calculated with the size, composition, shape, and the orientation of the particles. The
light extinction coefficient is influenced by meteorological conditions and optical properties
along the sight path.
4-154
-------
Relationship Between Light, Targets, and Objects in a Sight Path
The appearance of a distant object is determined by light from two sources: the light
reflected from the object itself (initial radiance) and the light reflected by the intervening
atmosphere (path radiance). Human vision and the brain's response to signals received from the
eye can distinguish between objects by contrast or differences in the radiance of adjacent objects.
Light reflected by objects is attenuated by scattering and absorption as it travels through the
atmosphere toward the observer. The portion that reaches the observer is the transmitted
radiance.
During the daytime, the sight path is illuminated by the direct rays of the sun, diffuse
skylight, light that has been reflected from the surface of the Earth, etc. Some of this
illumination is scattered toward the observer by the air molecules and PM in the sight path.
The accumulation of the light scattered into the sight path is the path radiance or air light. The
path radiance significantly influences the light transmitted by the object being viewed. As the
path radiance increases, the light transmitted by an object decreases.
The transmitted radiance carries the information about the object; the path radiance only
carries information about the intervening atmosphere and is often quite featureless. When the
transmitted radiance is dominant, visibility is good. Conversely, when the path radiance is
dominant, visibility is poor. In a dense fog, the transmitted radiance from nearby objects can be
seen, but the transmitted radiance from more distant objects is completely overwhelmed by the
path radiance (i.e., the light scattered by the fog). Distant objects are lost in the white (or gray)
of the fog (Gazzi et al., 2001).
Figure 4-32 illustrates the radiance seen by an observer looking at a hillside or through the
aperture of a measurement instrument. The radiance that enters the eye of the observer (or the
aperture of a measurement instrument) is known as the apparent radiance (i.e., the sum of the
transmitted and path radiance). The competition between the transmitted radiance and the path
radiance determines visibility.
Light Absorption and Scattering by Gases
In the ambient atmosphere, the only visible-light-absorbing gas of any consequence is
nitrogen dioxide (NO2), which primarily absorbs blue light and, if present in sufficient
concentration across a sight path, contributes to the yellow or brown color seen in urban hazes.
4-155
-------
Observer
Figure 4-32. 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 can no longer be distinguished from the light scattered
into the sight path.
Source: Watson and Chow (1994).
Usually the absorption by NO2 is much smaller than the scattering by particles that are typically
present in polluted environments, such as urban areas. The most common exception to this
situation of relatively small NO2 absorption is in effluent plumes from combustion facilities
where the particles are effectively removed but the 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.
Light scattering by gases in the atmosphere is described by the Rayleigh scattering theory
(van de Hulst, 1981) and is referred to as Rayleigh scattering. The magnitude of Rayleigh
scattering depends on the gas density of the atmosphere and varies from about 9 Mm"1
4-156
-------
to 11 Mm"1 for most locations of interest, depending primarily on site elevation. To simplify
comparisons of light extinction coefficient values among sites at a variety of elevations, a
standard value of 10 Mm"1 is often used for the Rayleigh scattering component (Malm, 2000).
Light Absorption and Scattering by Particles
Light scattering by particles tends to dominate light extinction except under pristine
atmospheric conditions when Rayleigh scattering by gas molecules is the largest contributor.
If the particle size, refractive index, and shape are known, the extinction coefficient can be
calculated. For particles of sizes similar to the wavelength of visible light, Mie equations for
homogeneous spheres can be used to calculate the scattering and absorption of individual
particles.
Absorption by particles is primarily caused by elemental carbon (also referred to as soot or
light-absorbing carbon) generated by the incomplete combustion of fossil fuels. Some minerals
in crustal particles also absorb light and can be a significant factor during fugitive dust episodes.
Most particle absorption data are determined by measuring light transmission or reflection
of particles captured on filter media. Absorption estimates made in this way are sensitive to the
filter substrate used, the optical configuration of the transmission measurement, particle loading
on the filter, and particle scattering albedo with the result that there are significant uncertainties
for measurements of filtered particles (Horvath, 1993). Another approach to estimating aerosol
light absorption is by subtracting concurrent light scattering measurements made with a
nephelometer from light extinction measurements made with a transmissometer. Substantial
uncertainty in this difference approach results from the assumption that the point measurement
of scattering is representative of the scattering over a long path (1 to 10 km) that is typically
required for transmissometer measurements. A recently field-tested prototype photoacoustic
spectrometer designed to determine absorption of suspended aerosol and an enclosed-folded path
transmissometer offer hope for resolving the problems of the filter-based and difference
approaches to the measurement of light absorption by particles (Arnott et al., 1999).
The relationship between elemental carbon concentration and particle absorption can be
calculated using Mie equations for particles with known size distribution, particle density, index
of refraction, shapes, and for various internal mixtures with non-absorbing aerosol materials
(Fuller et al., 1999). Mie equations are used to determine the efficiency factors for extinction
4-157
-------
Qext, Qscat, and Qabs- The efficiency factors represent the fraction of light falling on a circle with
the same diameter and index of refraction as the particle. While such application of this theory
can provide a range of absorption efficiencies for various model aerosol distributions, it is rare
that sufficiently detailed particle characterization data for ambient aerosols are available. Also,
although elemental carbon is the strongest and most common of the absorbing particles, light
absorption by elemental carbon particles can be reduced when the particle is covered by other
chemical species (Dobbins et al., 1994) or may be enhanced when coated with a non-absorbing
refractive material such as ammonium sulfate (Fuller et al., 1999).
More commonly, estimates of elemental carbon absorption efficiency are empirically
determined from the ratios of, or the slopes of regression analysis fits to, absorption coefficient
and corresponding elemental carbon concentration measurements. Use of the regression
approach permits the inclusion of crustal component concentrations as a second dependent
parameter, so that crustal absorption can also be estimated. Uncertainties in the absorption
efficiency determined empirically are a combination of the measurement uncertainties for the
absorption coefficients, elemental carbon concentrations, and, where used, the crustal
concentrations. In reviews of estimates of elemental carbon light absorption mass efficiency
(i.e., the absorption coefficient per carbon mass concentration), Horvath (1993) and Liousse
et al. (1993) found values ranging from 2 to 17 m2/g. Moosmiiller et al. (1998) showed that by
limiting the absorption coefficient estimates to those using photoacoustic methods, the
absorption efficiency shows a wavelength dependence with highest values (17 m2/g) at the
shortest wavelength used (A = 0.42 jim) and lowest values (3 m2/g) at the longest wavelengths
used (A = 0.8 jim). The center of the visible light wavelength (A = 0.53 jim) yielded elemental
carbon absorption efficiencies values of-10 m2/g, a commonly used value for elemental carbon
absorption efficiency. Fuller et al. (1999) suggested that isolated spheres of light absorbing
carbon have a specific absorption of less than 10 m2/g. Light absorption by carbon particles will
be greater than 10 m2/g only if the particles are internally mixed and the occluding particles are
sufficiently large. Absorption values for graphitic and amorphous carbon for primary sizes
typical of diesel soot are ~5 m2/g.
Particle scattering tends to dominate light extinction except under pristine atmospheric
conditions when Rayleigh scattering by gas molecules is the largest contributor. Light-scattering
by particles has been reported to account for 68 to 86% of the total extinction coefficient in
4-158
-------
several cities in California (Eldering et al., 1994). When light-scattering increases, visibility is
impaired because of a decrease in the transmitted radiance and an increase in the path radiance.
The single most important factor that determines the amount of light scattered by a particle is its
size, as shown in Figure 4-33 (based on Mie calculations). The maximum single-particle
scattering efficiency (i.e., scattering per cross-sectional area of a particle) is associated with
particles with diameters of about the wavelength of visible light (centered at 0.53 jim).
For particles that are small compared to the wavelength of light, the single-particle scattering
efficiency is low. For particles larger than the wavelength, the single-particle scattering
efficiency initially decreases with diameter and then fluctuates around a value of two as size
increases. However, a larger particle always scatters more light than a smaller particle because
particle cross-sectional area increases faster with diameter than does the decrease in single-
particle scattering efficiency at any point on the scattering efficiency curve. The mass scattering
efficiency (i.e., the scattering per mass concentration) peaks for particles that are about 0.5 jim to
0.8 |im in diameter. Smaller particles are much less efficient at scattering light, but the mass of
particles increases with particle size faster than the increase in the amount of light they scatter.
o
Total MIE Scattering Coefficient
For R= 1.50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Size Parameter, a
Figure 4-33. 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 =
Source: Penndorf (1958).
4-159
-------
Use of the Mie equation to calculate light scattering or the light scattering efficiency of
particles in the atmosphere is severely limited by the general lack of sufficiently detailed particle
characterization data. At a minimum, size-resolved particle composition data (e.g., aerosol
collected on an 8-stage impactor) are needed to permit meaningful Mie scattering calculations.
The chemical composition provides clues to the appropriate particle density and index of
refraction, while the size distribution is inferred by fitting a distribution function to the
concentration for each stage. Assumptions are still necessary to address the particle component
mixture characteristics of the aerosol. Resulting scattering calculations can be compared to
directly measured particle extinction to assess the reasonableness of the Mie calculations.
Reported calculated dry scattering efficiencies for sulfates range from 1.2 to 5.6 m2/g.
Sulfate scattering efficiencies have been reported to increase by a factor of two when the size
distribution increased from 0.15 to 0.5 jim (McMurry et al., 1996). Calculated scattering
efficiencies for carbon particles ranged from 0.9 to 8.1 m2/g. A scattering efficiency of 1.0 and
0.6 m2/g was reported for soil and coarse mass, respectively (U.S. Environmental Protection
Agency, 1996a; Sisler and Malm, 2000).
Figure 4-34 gives the volume-specific light scattering efficiency in units of jim"1 as a
function of particle diameter. The light scattering coefficient is derived by multiplying the
volume-specific light scattering efficiency factor by the volume concentration. The mass-
specific light scattering efficiency can be obtained by dividing the values for the curves by the
density of the PM.
Similar results have been produced in field nephelometer measurements of ambient particle
scattering. A variety of nephelometer configurations, unrestricted or size-selective inlets, and
the control of sample air temperature and relative humidity, permit the composite scattering
properties of ambient aerosol to be directly observed (Day et al., 1997). When sample-
controlled nephelometer data are combined with collocated particle speciation data, composite
particle scattering efficiency values for ambient aerosol can be empirically derived (Malm et al.,
2000).
The scattering efficiency for particles has been reported by White et al. (1994) for dry
particles < 2.5 jim (2.4 and 2.5 m2/g) and coarse particles (0.34 to 0.45 m2/g). Other reported
values for coarse particles include 0.4 and 0.6 m2/g (White and Macias, 1990; Trijonis et al.,
1987). Nephelometer measurements for light scattering by coarse particles is underestimated
4-160
-------
Particle Diameter (|jm)
Figure 4-34. Volume-specific light-scattering efficiency as a function of particle diameter
Dp. The calculations were performed for the indicated indices of refraction
and a wavelength of 550 nm. For large particle diameters, the scattering
efficiencies tend toward a value of 3/Dp. Mass-specific light-scattering
efficiencies (in units of m2/g) can be obtained by dividing the values of the
curves by the particle density (in units of g/cm3).
Source: U.S. Environmental Protection Agency (1996a).
(White et al., 1994). Chow et al. (2002a) reported scattering efficiencies of 2 to 3 m2/g but in
some cases > 5 m2/g for dry particles < 2.5 jim.
4.3.2.2 Relative Humidity Effects on Particle Size and Light-Scattering Properties
The ability of some commonly occurring chemical components of atmospheric aerosol to
absorb water from the vapor phase has a significant effect on particle light scattering.
Hygroscopic particles, which typically include sulfuric acid, the various ammonium sulfate salts,
ammonium nitrate, and sodium chloride, change size by the accumulation and loss of water as
4-161
-------
they maintain equilibrium with the vapor phase as a function of changes in relative humidity.
For some materials (e.g., sulfuric acid), the growth is continuous and reversible over the entire
range of relative humidity. For other materials, water absorption begins abruptly for a dry
particle at a specific relative humidity known as the deliquescent point (e.g., -80% for
ammonium sulfate) and continues as relative humidity increases. There is a hysteresis effect
with these materials in that, once wet, the relative humidity can be reduced below the
deliquescent point until crystallization occurs at a substantially lower relative humidity (e.g.,
-30% for ammonium sulfate). Figure 4-35 shows the water vapor growth curve for ammonium
sulfate.
o
Q
_0
*j
(B
1.5-
0
1.0
y'
^^^ A
f Hysteresis Loop
/ for (NH4)2 SO4
k
B A
I I I I I
30 50 70
9
o
o
Q.
- 5 ^
-4 £
(0
OL
o | 1 1 1
O -* M CO
Volume Growth
Relative Humidity (%)
Figure 4-35. Particle growth curve as a function of relative humidity (RH) showing
deliquescent growth of ammonium sulfate [(NH4)2 SO4] particles at the
deliquescent point (A, about 80% RH), reversible hygroscopic growth of
ammonium sulfate solution droplets at RH > 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).
4-162
-------
The water growth behavior for hygroscopic materials commonly found in atmospheric
aerosol in pure form or in some mixtures is generally well known as a result of laboratory
measurements (Tang and Munkelwitz, 1994; Tang, 1997). Models that calculate water growth
of mixtures from known solubility properties of many common water-soluble chemicals have
long been available (Zdanovskii, 1948) and have been successfully applied to determine growth
for particles with known composition (Saxena and Peterson, 1981; Pilinis et al., 1995; Saxena
etal., 1993).
The water growth of individual ambient particles can be directly measured using a
humidity-controlled tandem differential mobility analyzer (TDMA) (McMurry and Stolzenburg,
1989; Zhang et al., 1993). Inferences can be made about the mixtures of soluble and insoluble
particle components by comparing TDMA-measured growth and size-resolved aerosol
composition data with water growth model predictions (Pitchford and McMurry, 1994; Zhang
et al., 1993; Saxena et al., 1995). A practical limitation of TDMA measurements in studying
aerosol optical properties is that particles > 0.5 jim are not well measured by this approach.
Accounting for water growth of atmospheric aerosols is important in determining visibility,
because particles containing hygroscopic or deliquescent materials change size, index of
refraction, and, hence, scattering efficiency, with changing relative humidity. The nonlinear
nature of particle growth curves for hygroscopic aerosols means that substantial light scattering
changes result from modest relative humidity changes under humid conditions (relative humidity
> 90%). The magnitude of the water growth effect on light scattering for ambient aerosols can
be directly measured with humidity-controlled nephelometer measurements (Day et al., 1997).
Measurements of water growth effects on light scattering are compared to the results of water
growth and Mie scattering models applied to size-resolved composition data using various
mixture assumptions to infer average mixture and other aerosol characteristics (Malm et al.,
2000a,b).
While the importance of inorganic hygroscopic particles is well understood, the role of
organic compounds in particle water growth has been the subject of recent investigations.
In their interpretation of TDMA and particle composition data from two locations, Saxena et al.
(1995) made the case that organic components of the aerosol enhanced water absorption by
particles at a remote desert location and retarded water absorption at an urban location. They
4-163
-------
speculated that the latter might be due to hydrophobic organic material coatings on inorganic
hygroscopic particles.
Whereas some of the thousands of organic compounds that are in atmospheric aerosols are
known to be hygroscopic and a significant fraction of the organic aerosol material is known to be
water soluble, there is a lack of water absorption data for most organic compounds. The
incomplete water solubility data, combined with incomplete data on the abundance of the
numerous organic compounds in ambient aerosols, means that organic water growth model
calculations are not a reasonable approach for assessing the importance of water growth by
organic aerosol components in the atmosphere. To overcome this constraint, Saxena et al.
(1995) compared organic concentration to the difference between total aerosol water measured
by TDMA and model-estimated water for the inorganic hygroscopic aerosol components. In
contrast, Pitchford and McMurry (1994) used the same remote location data set and showed that,
on 6 of the 8 sampling days, water uptake by the sulfates and nitrates could account for all of the
measured water absorption.
Swietlicki et al. (1999) made TDMA measurements in northern England and found that
growth takes place in two modes, one mode being less hygroscopic that the other. They
concluded that growth could be attributed to the inorganic content of the aerosol. Cocker et al.
(2001) measured hygroscopic properties of Pasadena, CA aerosol and concluded that growth
factors increased when forest fires were present. McDow et al. (1995) measured water uptake by
diesel soot, automobile exhaust, and wood smoke particles. They found all three emission types
absorbed water: the wood smoke sample weight increased by about 10% as sample relative
humidity was increased; whereas diesel soot sample weight increased by only 2 to 3%. Chughtai
et al. (1999) examined the hydration characteristics of a number of anthropogenic and natural
organic materials. They found surface water adsorption increased with age and surface
oxidation. Hemming and Seinfeld (2001) evaluated the relative hygroscopicity of different
organics, the differences in the amount of water taken up by mixtures, and the individual
components of the mixtures in their pure state using the UNIFAC. They found that mixtures
take up less water than the individual components in the pure state. The relative hygroscopicity
of atmospheric organics was diacids > monoacids > alcohols > carbonyls. Analysis of humidity-
controlled and size-resolved chemistry data from the Great Smoky Mountains and Grand Canyon
National Parks (Malm et al., 1997; Malm and Kreidenweis, 1997; Malm et al., 2000a) showed
4-164
-------
that ambient organic aerosol are, at most, weakly hygroscopic to within the measurement
uncertainty and modeling assumptions.
A more detailed discussion of the effects of relative humidity on the size distribution of
ambient particles appears in Chapter 2 of this document.
4.3.3 Relationships Between Particles and Visibility
Visibility, referring to the appearance of scenic elements in an observer's line of sight,
depends on more than the optical characteristics of the atmosphere. Numerous scene and
lighting characteristics are important to this broad definition of visibility. However, under a
variety of viewing conditions, visibility reduction or haziness is directly related to the extinction
coefficient.
Light extinction, the sum of the light scattered and absorbed by particles and gases, is
frequently used to estimate the effect of air pollution on visibility. Light extinction is usually
quantified using the light extinction coefficient, i.e., the sum of the light scattering and
absorption coefficients for gases and particles (see Section 4.3.2.1).
The influence of particles on visibility degradation is dependent on the particle size,
composition, and solubility (Pryor and Steyn, 1994). Fine particles (particles with mass mean
diameters < 2.5 jim) scatter more light than coarse particles. Fine particle species include
sulfates (assumed to be ammonium sulfate), nitrates (assumed to be ammonium nitrate),
organics, light-absorbing carbon, and soil (Malm et al., 1994). Of the fine particle species,
sulfates and nitrates are the most hygroscopic and require the use of a relative humidity
adjustment factor. The effect of particle light extinction can be determined by totaling the
scattering and absorption of light by multiplying the mass-specific efficiency values and the
mass concentration for each of the particle species. The effect of relative humidity and the
relative humidity adjustment factors are discussed in Section 4.3.2.2.
Visibility is measured by human observation, the light extinction coefficient (light
scattering and absorption by particles and gases), and parameters related to the light extinction
coefficient (visual range, deciview) and fine particle mass concentrations. The equation for the
light extinction coefficient, light extinction by particles is stated in the discussion of the
IMPROVE Program.
4-165
-------
The visual range method of visibility measurement, commonly taken to be the greatest
distance that a large dark object (e.g., a mountain in shadow) can be seen against the background
sky (Middleton, 1952), was developed for and continues to function well as an aid in military
operations and transportation safety. Visual range is inversely proportional to the light
extinction. Visual range can be calculated from a point measurement of light, assuming that the
atmosphere and the illumination over the sight path is uniform and the threshold contrast is 2%
of the extinction coefficient
Visual Range = Klbext (4-8)
where visual range is in kilometers, &ext is in km"1, and a threshold contrast of 2% is assumed.
If &ext is in Mm"1, the Koschmieder constant becomes 3,912.
An index of haziness, expressed in deciview (dv) units, is also very simply related to the
light extinction coefficient (Pitchford and Malm, 1994).
Haziness (dv) = 10 In (bext /10 Mm -1) (4-9)
An important characteristic of this visibility index is that it is more nearly linearly related to
perceived changes in haze level than either visual range or light extinction. A change of 1 or
2 dv in uniform haze under many viewing conditions will be seen as a small but noticeable
change in the appearance of a scene, regardless of the initial haze condition.
Figure 4-36 illustrates the relationship of light extinction in Mm"1, deciview index, and
visual range in kilometers. Although the deciview is related to extinction, it is scaled in such a
way that is perceptually correct (Fox et al., 1999).
The amount of light that an aerosol of a given mass concentration scatters depends on the
particle size distribution. Several studies and reports (Chow et al., 2002a; Samuels et al., 1973;
Waggoner and Weiss, 1980; Waggoner et al., 1981) indicate that the mass concentration of
particles of < 3 jim in diameter correlate well with light scattering and visibility when the
measurements are made under dry conditions. The question of whether the relationship between
4-166
-------
Extinction (Mm"1)
Deciviews (dv)
10 20 30 40 50 70 100 200 300 400 500 700 1000
i
0
i
i
7
i
i
11
i
i
14
i
i
16
i
I I III
19 23
I I III
I I
30 34
I 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 4-36. Comparison of extinction (Mm"1) and visual range (km).
Source: Fox etal. (1999).
light scattering and fine particle mass varies too greatly between geographical locations to allow
the use of fine particle mass as a metric for visibility effects has been considered. McMurry
(2000), citing a study by Charlson et al. (1968), stated that the ratio of dry scattering coefficient
to the dry fine particle mass concentration from different geographical locations did not vary a
great deal. However, the use of either fine particle mass or the individual major components to
determine particle-related light scattering requires measurements, estimates, or assumptions
about concurrent ambient relative humidity in order to accurately estimate the relationship
between ambient fine particles and light extinction.
Figure 4-37 shows the relationship between fine particle mass and calculated light
extinction. The figure, as reported in Chow et al (2002b), was generated using data reported by
Samuels et al. (1973). According to Samuels et al. (1973), there was a direct correlation
between particle mass concentration, light scattering, and visibility. However, there were large
standard errors in the scattering coefficient.
Most routine aerosol monitoring programs and many special study visibility
characterization programs were designed to measure the major aerosol components (Malm et al.,
1994; Tombach and Thurston, 1994; Watson et al., 1990); they were not designed to determine
the microphysical and chemical characteristics of these species. However, the inherent
limitations of estimating aerosol optical properties from bulk aerosol measurements have been
addressed, at least in part, by a number of authors. For instance, Ouimette and Flagan (1982)
have shown using basic theoretical considerations that if an aerosol is mixed externally (i.e.,
separate particles contain the major aerosol components) or, if internally mixed, the index of
4-167
-------
1600-,
£ 1400-
1200-
O)
C
(0
± 1000-1
(0
.52 800-
G)
're
600-
400-
O) 200-
M
x
50 100 150 200 250 300 350
PM3
Figure 4-37. Proportionality of observed daytime haziness to fine particle mass
concentration in Los Angeles. Visual ranges are 8-h averages of hourly
human observations, plotted as extinction according to Koschmieder
formula. Mass concentrations are from 8-h samples collected behind
a cyclone with 3-um cut point. Relative humidities were <• 70%.
Source: Chow et al. (2002b).
refraction is not a function of composition or size and the aerosol density is independent of
volume, then
where a{ is the specific mass scattering efficiency and m{ is the mass of the individual aerosol
species.
Sloane (1983, 1984, 1986), Sloane and Wolff (1985), and more recently, Lowenthal et al.
(1995) and Malm and Kreidenweiss (1997) have shown that differences in estimated specific
scattering between external and internal model assumptions are usually less than about 10%.
In the absence of detailed microphysical and chemical information of ambient particles, the
4-168
-------
above studies demonstrate that a reasonable estimate of aerosol extinction can be achieved by
assuming each species is externally mixed.
The latest IMPROVE network report (Malm, 2000) included calculated aerosol light
extinction for each of the five major fine fraction particle (PM25) components, coarse fraction
mass (PM10_2 5), and Rayleigh scattering by gases and summed them for an estimate of total light
extinction in Mm"1 using the following algorithm:
bext = (3)f(RH) [SULFATE]
+(3)f(RH) [NITRATE]
+(1.4) [ORGANIC CARBONJ
+(10) [LIGHT ABSORBING CARBON] (4-11)
+(1)1SOIL/
+(0.6) [COARSE PMJ
+10 (for Rayleigh scattering by gases)
where each PM term is the product of a constant dry extinction efficiency for that species, the
mass concentration of the species, and, for sulfate and nitrate, an adjustment factor that is a
function of relative humidity to account for their hygroscopic behavior. The relative humidity
adjustment term for sulfate and nitrate, shown in Figure 4-38, is based upon the ammonium
sulfate growth curve, shown in Figure 4-28, smoothed between the upper and lower curves of the
hysteresis loop for the relative humidity range of 30 to 80%. The extinction efficiencies for soil
and coarse mass used in this algorithm are taken from a literature review by Trijonis et al.
(1987). The extinction efficiency for light absorbing (elemental) carbon of 10 m2/g is consistent
with the value reported by Moosmiiller et al. (1998) corresponding to A = 0.53 in the middle of
the visible light spectrum. The dry extinction efficiencies of 3 m2/g for sulfate and nitrate
species and 4 m2/g for organic species are based on literature reviews by Trijonis et al. (1991)
and by White (1991). Trijonis' best estimate for sulfates is 2.5 m2/g with an uncertainty of a
factor of 2, while White's average low and high estimates for the rural West are 3.0 and
3.7 m2/g. For organics, Trijonis estimated a dry extinction efficiency of 3.75 m2/g with an
uncertainty of a factor of 2, and White's range for the rural West is 1.8 to 4.1 m2/g. Malm et al.
4-169
-------
u
re
LL
O)
c
IE
O)
'53
20 30 40 50 60 70 80
Relative Humidity (percent)
90 100
Figure 4-38. Relative humidity adjustment factor, f(RH), for ammonium sulfate as a
function of relative humidity.
Source: Malm (2000).
(1996) and Malm (2000) used this algorithm to successfully reconstruct light scattering at a total
of eleven IMPROVE monitoring sites.
Malm (2000) used additional sophisticated aerosol size, composition, and microphysical
data from a special study at the Great Smoky Mountains National Park to compare the
performance of a number of models for calculating light extinction. He found that the simplist
approach adequately predicted for periods of low light scattering but under-predicted by about
30% during periods of high sulfate concentration. The greatest improvement over the simple
model was obtained by including the degree of sulfate ammoniation in the model. This
produced better estimates of the extinction coefficient over the entire range. Table 4-15 lists
various visibility metrics and methods for visibility measurement.
4-170
-------
TABLE 4-15. VISIBILITY MEASUREMENT TECHNIQUES
Visibility Metric and Method
Measurement Principle
Visual Range
Light extinction (bea)
Long-path transmissometer
Short-path transmissometer
Contrast transmittance:
Teleradiometer
Photographs and time-lapse
film
Particle scattering (bscatp):
Integrating nephelometer
Human observation of prevailing visibility. Targets are selected at known distances from an observer. Nighttime targets
require lights and may differ from daytime targets. Each hour, the observer records the distance (i.e., visual range)
corresponding to the furthest target that is visible. This method provides the longest history of visibility measurements
in the United States, as it was used at most U.S. airports from 1948 to 1995.
Directly measures the radiance of a constant light source (transmitter) after the light travels over a finite atmospheric path.
The transmittance of the path is calculated by dividing the measured radiance by the calibrated initial intensity of the light
source. The average extinction of the path is calculated from the transmittance and length of the path.
Starting in early 1990s, many airports replaced human observations with automated sensors (i.e., the Automated Surface
Observing System [ASOS], Automated Weather Observing System [AWOS]) to measure 1-min-average light extinction.
The visibility sensor measures forward scattering using a xenon flash lamp source. Instruments can provide measurements
up to 48 km, but they are not recorded beyond 16 km.
A telescope is focused on a distant target and the background. Changes in radiance are measured by photodiode detectors.
Measurements can be made at different wavelengths (e.g., 450, 550, and 630 nm) using narrow band filters. Measures
effects of particles of all sizes. Sensitive to variations in inherent contrast (e.g., bright or dark cloud behind the target)
and nonuniform illumination conditions. Because path radiance depends on how the atmosphere is illuminated, contrast
transmittance represents human perception better than air quality.
A densitometer measures light transmitted through different portions of a color slide. The film's light-response function
(the gamma curve) determines target-sky contrast. Computerized photographic simulations can change contrast
transmittance for different meteorological and atmospheric conditions, and can be used to judge how people react to
these changes.
Air is drawn into a nephelometer chamber that is illuminated with white or filtered (typically 500-550 nm) light. Light is
detected at 90° to the direction of illumination to measure the amount scattered out of the light path. Chamber dimensions
limit the integrated arc to -10-170° instead of a full 0-180°, which results in the underestimation of some forward scattering
from coarse particles. Nephelometers are calibrated with gases of known indices of refraction. Particles (especially
hygroscopic and volatile species) may be modified as they pass through the chamber, which is inadvertently heated by
the illumination source.
-------
TABLE 4-15 (cont'd). VISIBILITY MEASUREMENT TECHNIQUES
Visibility Metric and Method
Measurement Principle
-^
to
Particle absorption (&„,.,„):
Aethalometer or particle
soot absorption photometer
Photoacoustic spectroscopy
Filter transmittance,
reflectance
Suspension of insoluble
elemental carbon
Component extinction
(beasumofbscatp,bscatg,
babs,P, and babS:g)
Chemical extinction (bmtY
Filter measurements for
SO42~, NO3 , organics, EC,
fine soil, and coarse mass,
plus clear air scattering
Particles are collected on a quartz-fiber filter tape. The change in transmittance at selected wavelengths (e.g., 880 nm)
across the filter before and after sampling is measured or compared with the reference area. When the filter spot darkens,
a new portion of the tape is moved into the sampling position. Assumes a constant relationship between black carbon (BC)
mass and quartz filter transmittance of ~19m2/g. Assumes a relationship of 10 nf/g between BC absorption and
concentration.
Particles absorb energy from a modulated laser (-514.5 nm) and transfer heat to the surrounding air. Expansion of the
heated gas produces sound waves (acoustic signals) that are proportional to the amount of absorbed energy. These are
detected by a high-sensitivity microphone. Absorption appears to vary with illumination wavelength.
Uses densitometer, integrating plate, or integrating sphere spectrophotometer to measure how much light is transmitted or
reflected. The difference in the logarithms of light transmission through the filter before and after sampling is proportional
to the particle deposit. Light transmission of particles collected on a filter may overestimate light absorption because some
of the incident light is scattered within the filter and by other particles in the deposit. There are no absolute calibration
standards for densitometry, integrating plate, or sphere methods.
Particles collected on a Nuclepore polycarbonate-membrane filter are extracted in 30% isopropanol/70% distilled deionized
water to form a suspension of insoluble EC particles. Using a spectrophotometer, light transmission (e.g., 400-650 nm,
peaking at 575 nm) is measured through the liquid extract.
The sum of clean air scattering estimated from temperature and pressure, NO2 absorption estimated from NO2
concentrations, particle scattering measured by nephelometer, and particle absorption (6^) measured by one of the babsp
methods. Measurements are at a single location rather than along a sight path.
Six aerosol chemical components are used to calculate chemical extinction.
Adapted from: Watson (2002).
-------
4.3.4 Photographic Modeling of Visibility Impairment
None of the visibility indices communicate visibility associated with various aerosol
conditions as well as directly seeing their effects on a scene. Photographic modeling for the
representation of haze can be useful in portraying changes in visibility specifically due to
changes in air pollutant concentrations. Photographic modeling holds constant the effects of sun
angle, cloud cover, and relative humidity and is a cost-effective method of evaluating various air
quality scenarios. Photographic modeling is difficult to do with actual photographs because of
the range of possible conditions in the same scene over multiple days; and, over time,
photographs can be expensive to produce. Another limitation in using photographic models for
representation of haze is that haze is assumed to be uniformly distributed throughout the scene
and selected conditions are idealized, so the full range of conditions that occur in a scene are not
represented.
Eldering et al. (1996) proposed the use of a model that uses simulated photographs from
satellite and topographic images to evaluate the effect of atmospheric aerosols and gases on
visibility. Use of this model requires ground-based photography and data concerning the size
distribution and chemical composition of atmospheric aerosols, NO2 concentration, temperature,
and relative humidity for a clear day. Light extinction and sky color are then calculated based on
differences in aerosol size distribution, NO2 concentration, temperature, and relative humidity.
The images created represent natural landscape elements.
Molenar et al. (1994) provides a discussion of existing visual air quality simulation
methods based on techniques under development for the past 20 years. A photograph taken on a
very clean, cloud-free day serves as the base image. The photograph is taken during the season
and at the same time of day as the scene to be modeled. The light extinction represented by the
scene is derived from aerosol and optical data associated with the day the image was taken, or it
is estimated from contrast measurements of features in the image. The image is then digitized to
assign an optical density to each picture element (pixel) for the wavelength bands of interest.
A detailed topographic map and an interactive image-processing display system is used to
determine the specific distance, elevation angle, and azimuth angle for each element in the
picture with respect to the observer's position.
Various models are employed to allow the presentation of different air quality scenarios.
The output from atmospheric aerosol models (e.g., extinction, scattering coefficients, single-
4-173
-------
scattering albedo, and scattering phase matrix) is incorporated into radiative transfer models to
calculate the changes in radiant energy (path radiance, image radiance, sky radiance, terrain
radiance) caused by scattering and absorption by gases and particles as it passes through the
atmosphere. Atmospheric aerosol models are also used to model the effect of relative humidity
on the visual air quality (Molenar et al., 1994).
Molenar et al. (1994) has developed a system call WinHaze that permits the viewing of
computer-generated uniform hazes superimposed on digitized scenic photographs of both remote
and urban scenes. The program simulates changes in visual air quality imagery from user-
specified changes in optical parameters (e.g., aext, visual range, or deciview values) or aerosol
species concentrations. WinHaze includes imaging for various Class I national parks and
wilderness areas and for Boston, MA; Dallas, TX; Denver, CO; Fort Collins, CO; Phoenix, AZ;
and Tucson, AZ. The computer software is available through the IMPROVE website
(http ://vista. cira. colostate.edu/ improve/).
4.3.5 Visibility Monitoring Methods and Networks
Visibility monitoring studies measure the properties of the atmosphere either at the sampler
inlets (point measurements), as is the case with air quality measurements, or by determining the
optical properties of a sight path through the atmosphere (path measurements). Instrumental
methods for measuring visibility are generally of three types: (1) direct measurement of light
extinction of a sight path using a transmissometer, (2) measurement of light scattering at one
location using an integrating nephelometer, and (3) measurement of ambient aerosol mass
concentration and composition (Mathai, 1995).
The largest instrumental visibility monitoring network in the United States is the
Automated Surface Observing System (ASOS), commissioned by the National Weather Service,
Federal Aviation Administration, and Department of Defense for use at more than 900 airports.
The system is designed to objectively measure the clarity of the air versus the more subjective
evaluations of human observations. The system provides real-time data for airport visibility.
The visibility sensor, instead of measuring how far one can see, measures the clarity of the
air using a forward-scatter visibility meter. The clarity is then converted to what would be
perceived by the human eye using a value called Sensor Equivalent Visibility (SEV). Values
derived from the sensor are not affected by terrain, location, buildings, trees, lights, or cloud
4-174
-------
layers near the surface. The amount of moisture, dust, snow, rain, and particles in the light beam
will affect the amount of light scattered. The sensor transmits 1-minute values based on rolling
10-minute periods. Hourly visibility range data are available only at a quantized resolution of
18 binned ranges with a visual range of up to 10 miles. The value provides a generally accurate
and representative visibility measurement within a 2 to 3 mile radius of the site. The forward
scatter meter was found to correlate fairly well with extinction coefficient measurements from
the Optec Transmissometer (National Weather Service, 1998).
Visibility data from the ASOS network is reported in terms of visual range in increments of
1/4 to 1 statute mile. Visual range conditions exceeding 10 miles are truncated to 10 miles for
real-time reporting purposes. Data is not extensively archived at ASOS locations; however,
researchers are able to download the raw data directly from certain sites. In addition, since 1998,
the raw visibility data (including light extinction measurements corresponding to visual ranges
exceeding 10 miles) have been archived for a number of sites. Visual range measurements
beyond 10 miles may be used to derived PM concentrations except in clean environments.
These data are available from the National Climatic Data Center.
The largest monitoring network that includes both visibility and aerosol conditions is the
IMPROVE network. This network was formed in 1987 as a collaborative effort between
Federal, regional, and state organizations responsible for protection of visibility in the 156
mandatory Class I Federal areas (national parks and wilderness areas) and other areas of interest
to land management agencies, states, tribes, and other organizations (National Park Service,
1998; U.S. Environmental Protection Agency, 1995b, 1996a, 1999b; Eldred et al., 1997; Perry
et al., 1997; Sisler and Malm, 2000). It is predominantly a rural-based network with more than
140 sites across the country. The primary monitoring objectives of the IMPROVE program are
to document current visibility conditions in the mandatory Class I areas, to identify
anthropogenic chemical species and emission sources of visibility impairment through the
collection of speciated PM2 5 data, and to document long-term trends for assessing progress
towards elimination of anthropogenic visibility impairment. The IMPROVE network has also
been involved in visibility related research, including the advancement of visibility monitoring
instrumentation and analysis techniques and visibility monitoring and source attribution field
studies (National Park Service, 1998; Evans and Pitchford, 1991).
4-175
-------
Visibility monitoring under the IMPROVE network can be divided into three categories:
aerosol, optical, and scene. Twenty-four hour PM2 5 and PM10 aerosol samples are collected by
filters at least every third day. The PM2 5 samples are analyzed to determined the mass
concentration of the major parti culate constituents (sulfates, nitrates, organic carbon compounds,
elemental carbon, chlorides, and crustal elements) and for elements that indicate sources of
visibility-impairing particles (trace elements and ions). Optical monitoring provides a direct
measurement of light scattering and absorption. Color photographic imaging documents the
appearance of the scene under a variety of air quality and illumination conditions (U.S.
Environmental Protection Agency, 1999b). It is anticipated that all data generated by the
IMPROVE network will be added to the AIRS database.
The EPA has deployed a new national monitoring network (Federal Reference Method
Monitoring network) designed to assess PM2 5 concentrations and composition. There are over
1,000 monitoring sites in operation and many sites report data to the AIRS. Analyses of these
data are expected to provide a more complete understanding of visibility conditions, in particular
urban visibility, across the country. The PM2 5 monitoring effort has been coordinated with
visibility monitoring efforts currently in place to maximize the benefits of all of the monitoring
programs (U.S. Environmental Protection Agency, 1997b, 2000b, 200la).
The Northeast States for Coordinated Air Use Management (NESCAUM) has established a
real-time visibility monitoring network (CAMNET) using digital photographic imaging. There
is currently digital photographic imaging for five urban locations (Boston, MA; Burlington, VT;
Hartford, CT; Newark, NJ; and New York City, NY), and two rural locations (Acadia National
Park, ME and Mt. Washington, NH). The visibility images are updated every 15 minutes. Near
real-time air pollution and meteorological data are updated every hour. Archived images will be
available for studies of the visual effects of PM air pollution in the Northeast. CAMNET may be
accessed at www.hazecam.net (Northeast States for Coordinated Air Use Management, 2002;
Leslie, 2001).
The Midwest Regional Planning Organization, in cooperation with a number of other
groups, has also developed a real-time visibility camera network (hazecam). The camera
network includes several urban (Chicago, IL; Indianapolis, IN; and Cincinnati, OH) and rural
locations (Seney NWR, MI; Mayville, WI; and Isle Royale National Park, Ml/Grand Portage,
MN). The Midwest hazecam can be found at http://www.mwhazecam.net.
4-176
-------
4.3.6 Visibility Impairment: Trends and Current Conditions
In the United States, visibility impairment is caused by particles primarily composed of
sulfates, nitrates, organic compounds, carbon soot, and crustal dust. Visibility is best in Alaska
and the western Great Basin. Moderate levels of light extinction are common on the Pacific
Coast, including the western slopes of the Sierra Nevadas in California and the Cascade Range in
Oregon and Washington. Visibility is most impaired in the areas encompassing and adjacent to
the Ohio and Tennessee River Valleys. Visibility gradually improves along the Atlantic
seaboard northeast of New York City (Watson, 2002). Natural visibility in the East and West is
about 75 to 150 km (45 to 90 miles) and 200 to 300 km (120 to 180 miles) (U.S. Environmental
Protection Agency, 200Ib).
4.3.6.1 Trends in Visibility Impairment
Trends in visibility impairment or haziness often are used as indicators of trends in fine
particles mass. Observations of visual range, obtained by the National Weather Service and
available through the National Climatic Data Center of the National Oceanic and Atmospheric
Administration provide one of the few truly long-term, daily records of impairment related to air
pollution. After some manipulation including correction for relative humidity effects, the visual
range data can be used as an indicator of fine mode particle pollution. The data reduction
process and analyses of resulting trends have been reported by Schichtel et al. (2001), Husar
et al. (1994), Husar and Wilson (1993), and Husar et al. (1981).
There are many statistical approaches to estimating trends. These approaches include
simple correlation and regression analyses, time-series analyses, and methods based on
nonparametric statistics. A discussion and comparison of the methods for the detection of linear
trends is provided in Hess et al. (2001). Schimek (1981) provides a discussion of nonlinear
trends. In its annual air quality trends report, the EPA characterized trends using a
nonparametric regression analysis approach commonly referred to as the Theil test (U.S.
Environmental Protection Agency, 1998; Hollander and Wolfe, 1973).
Regional Trends and Class I Areas
The two largest contributors to visibility impairment are sulfates and carbon-based
particles. In the East, sulfates are responsible for 60 to 86% of the visibility impairment. The
4-177
-------
sulfate contribution decreases further west but is still responsible for between 25 to 50% of the
visibility impairment. Carbon-based particles are responsible for 10 to 18% of the visibility
impairment in the East and 25 to 40% in the West. Nitrates account for only 7 to 16% of the
light extinction in the East but are responsible for between 5 to 45% of the light extinction in the
West. Crustal material can be a major contributor in the West, accounting for 5 to 25% of the
light extinction. Elemental carbon also is contributor to light extinction, but to a lesser degree
(U.S. Environmental Protection Agency, 2001a,b).
The EPA designated five regional groupings as part of the regional haze program. The
regions are Northeast (Mid-Atlantic/Northeast Visibility Union), Southeast (Visibility
Improvement State and Tribal Association of the Southeast), Central (Central States Regional
and Air Partnership), Midwest (Midwest Regional Planning Organization), and West (Western
Regional Air Partnership). The regional groupings serve as consensus organizations comprised
of states, tribes, and federal agencies coordinating the implementation of the regional haze rule.
Using hourly prevailing daytime visibility data from human observations at weather
stations, Schichtel et al. (2001) observed that haziness declined approximately 10% across the
United States between 1980 and 1995. The decrease in haziness was highest in the southeastern
United States with a 20% decrease in the 90th percentile light extinction and a 12% decrease in
the 75th percentile over the 15-year period. There was a 17% decrease in the 90th percentile and
a 9% decrease in the 75th percentile over the eastern United States. Over the eastern United
States, haziness was greatest during the summer months. The greatest visibility impairment was
adjacent to the Appalachian Mountains in Tennessee and the Carolinas (extinction coefficient of
> 0.2 km"1 equivalent to 6 miles). During the cold season, elevated haze (extinction coefficient
of > 0.2 km"1) was seen between the Great Lakes and the Ohio River Valley, over the gulf states
between Texas and Florida, along the coast from North Carolina to New Jersey, and along the
Pacific coast, particularly central and south California.
Visibility impairment or haziness in the southeastern United States from sulfate emissions
is greatest in the humid summer months because of the ability of sulfate to absorb atmospheric
water vapor. Summer haziness increased in the southeastern United States from the 1950s to
1980 along with increasing SO2 emissions. A statistically significant increase in summer sulfate
concentrations was noted in two Class I areas in the eastern United States (Shenandoah and the
Great Smoky Mountains) from 1982 to 1992 (Eldred et al., 1993; Cahill et al., 1996). During
4-178
-------
that time, the majority of the Southwest showed decreasing sulfur concentrations (Eldred et al.,
1993; Eldred and Cahill, 1994). Increasing summer sulfate concentrations were later shown at
those two locations by Iyer et al. (2000).
Limited visibility measurements are available for the upper Midwest region (Illinois,
Indiana, Michigan, Ohio, and Wisconsin and the tribal lands located in those states). The
Midwest Regional Planning Organization conducted an initial assessment of the regional haze
problem in this region using existing reports and available air quality data for four major urban
areas (St. Louis, Chicago, Detroit, Cincinnati) and Class I areas. The "worst" and "best"
visibility days occur throughout the year. Particulate sulfates were the major contributors to
light extinction during the summer months, and nitrates dominated on the worst visibility days
during the winter and fall in both urban and Class I areas. Organics also were significant
contributors to light extinction in urban areas. Higher PM2 5 concentrations were correlated with
poorer visibility in the southern portion of the region (Midwest Regional Planning Organization,
2001).
The EPA's National Air Quality and Emission Trends Report summarized an estimate of
the regional trends and current conditions in 35 Class I areas and one urban area (Washington,
DC), using chemical concentrations data from the IMPROVE network (U.S. Environmental
Protection Agency, 200la). The visibility trends analysis is an aggregate of 10 eastern Class I
areas and 26 western Class I areas. Trends were presented for annual average values for the
clearest ("best") 20% , middle ("typical") 20%, and haziest ("worst") 20% of the days monitored
each year. The visibility trends, given in changes in deciview values, for the eastern and western
sites are illustrated in Figures 4-39a and 4-39b. From the figures, it can be seen that the haziest
days in the West are equivalent to the best days in the East. In the East, there was a 16%
(1.5 deciview) improvement in haziness on the clearest days since 1992. Improvements in
visibility were noted in the East for the haziest days. However, based on monitoring data for
1999, visibility remains significantly impaired, with a visual range of 23 km for the haziest days
compared to a mean visual range of 84 km for the clearest days. A 25% and a 14% improvement
in visibility impairment were seen for the clearest and middle days in the West, respectively;
whereas conditions for the haziest days degraded by 18.5% (1.7 deciviews) between 1997 and
1999, but were relatively unchanged compared to 1990 conditions (U.S. Environmental
Protection Agency, 200la).
4-179
-------
35
30
'5
3 25
S 20
i_
g. 15
E
.-s1 10
!5
!<2 5
0
Haziest 2p-percent
Typical 20-percent
Clearest 20-percent
92 93 94 95 96 97 98 99
Year
Figure 4-39a. Aggregate visibility trends (in deciviews) for 10 eastern Class 1 areas.
"3T
'>
'o
0)
2,
4-1
c
0)
E
'S
u.
&
!5
'(/>
>
oa
•^n
25
20
1 E:
10
c
n
Haziest 20-percent
Typical 20-percent
_
Clearest 20-percent
90 91 92 93 94 95 96 97 98 99
Year
Figure 4-39b. Aggregate visibility trends (in deciviews) for 26 western Class 1 areas.
Source: U.S. Environmental Protection Agency (2001a).
4-180
-------
Figures 4-40a and 4-40b illustrate aggregate trends in aerosol light extinction, including
trends by major aerosol component for the haziest 20% of days monitored for the 10 eastern
Class I areas from 1992 to 1999 and the haziest 20% of days monitored for the 26 western
Class I areas from 1990 to 1999. The National Ambient Air Trends Report also includes a
number of maps characterizing aerosol light extinction and key components at 36 IMPROVE
sites (all rural except Washington, DC) for 1997 through 1999 (U.S. Environmental Protection
Agency, 200la).
Urban Trends.
Most of the available visibility measurements, with the exception of the airport visual
range measurements, focus on the impact of pollution on visibility in scenic vistas and regional
haze (Class I areas). Many urban metropolitan areas monitor daily visibility conditions. These
findings are generally not available in a published form and may not distinguish between
pollution- and weather-related effects. Although the EPA Regional Haze Rule addresses
visibility impairment in Class I areas and calls for states to establish goals for improving
visibility in these areas and to develop long-term strategies for reducing emissions of air
pollutants that cause visibility impairment, the steps states take to implement the regulation will
also improve visibility and health in broad areas across the country.
Kleeman et al. (2001), citing previously published studies, provided an historical
description of visibility conditions in Southern California from the early 1930s. Based on airport
observation data for 1932 to 1949, visibility conditions began to decrease in Los Angeles with
the advent of industrialization and population growth. Visibility conditions were worse during
the 1940s than the 1930s, with the lowest visibility conditions occurring between 1944 and 1947.
During this period, there was nearly a complete loss of extremely good visibility days. Between
1943 and 1947, the number of extremely good visibility days during the summer season dropped
from 21% to 0.2%. Between 1950 and 1961, deteriorating visibility conditions extended
eastward from Los Angeles along the corridor adjacent to the foothills of the San Gabriel and
San Bernardino Mountains. The visual range in the areas nearest Los Angeles was < 3 miles for
more than 140 days per year when the relative humidity was < 70%. Further east of Los
Angeles, past Ontario and San Bernardino, the visibility was < 3 miles for 110 days per year
during the same time period. Improvements in visibility conditions have been made since the
4-181
-------
200
150-
100-
c
o
~ o
X
LLI
o
(A
-------
1970s. The greatest improvements occurred in the western Los Angeles Basin. In Ontario, the
average number of days per year when the visual range was > 10 miles was 99 between 1976 and
1978 and increased to 113 between 1988 and 1990 (Kleeman et al., 2001).
In contrast to the Los Angeles Basin and Ontario area, Denver and surrounding areas have
experienced high pollution episodes since the 1970s. Climate changes contributed to the high
pollution episodes by trapping cooler air under a cover of warm air, causing the pollution to
remain stagnant over the area and producing a brown cloud comprising a variety pollutants
including nitrogen and sulfur oxides, as well as grit and dust. Debate over the cause of the
increasing pollution in the Denver area and the controversy over converting coal-fired power
plants to a cleaner natural gas system led to the initiation of the 1973 Denver Air Pollution
Study, 1978 Denver Winter Haze Study, 1987-1988 Metro Denver Brown Cloud Study, and the
1993 Denver Brown Cloud Modeling Study. In 1990, in an effort to improve air quality, Denver
adopted a visibility standard of 0.076 km"1 (units of atmospheric extinction per kilometer;
20.1 deciviews) averaged over 4 hours. While this is a step towards reducing air pollution, the
Denver region still exceeds the visibility standard 50 to 80 times per year (Lloyd, 2002).
During the fall and winter of 1988 to 1990, a major air quality study was conducted in
Phoenix, AZ to address degrading visibility conditions in Phoenix and other urban areas. The
objectives were to (1) develop a data base of visibility, air quality, and meteorological
measurements; (2) establish quantitative relationships between light extinction and emission
sources; and (3) evaluate measurement systems for short-term and long-term monitoring in
Phoenix. The major contributors to light extinction in Phoenix were residual wood burning,
primary mobile source emissions, and secondary ammonium nitrate (Chow et al., 1990).
The Arizona Department of Environmental Quality has conducted optical measurements of
visibility in Tucson since 1993 and in Phoenix since 1994. The measurements are divided into
the mean of the "dirtiest" 20% of all hours, the mean of all hours, and the mean of the "cleanest"
20% of all hours for the entire day and for the 5:00 to 11:00 a.m. period. Figures 4-41a,b
represent the trends in visibility conditions from 1993 to 2001 for Tucson and from 1994 to 2001
for Phoenix. Visibility on the dirtiest days in the Phoenix metropolitan area has not changed
since visibility monitoring started; but, visibility on the best days has significantly degraded.
There is a seasonally related effect on visibility in Phoenix: the dirtiest 20% of all hourly light
extinction and the mean of all hours are more pronounced during the winter and fall months
4-183
-------
140
DO
\ \ \ \
1993 1994 1995 1996
- dirtiest 20%-all ---
H- dirtiest 20%-5-11 -*-
\\
1997 1998
mean-all
mean-5-11
\ \ \
1999 2000 2001
s- cleanest 20%-all
*- cleanest 20%-5-11
Figure 4-41a. Light extinction trends in Tucson, Arizona from 1993 to 2002.
Source: Arizona Department of Environmental Quality (2002).
140
n r
1994 1995
T
T
T
1996
dirtiest 20%-all
dirtiest 20%-5-11
1997 1998
E- mean-all
*- mean-5-11
1999 2000 2001
-£s- cleanest 20%-all
-*?- cleanest 20%-5-11
Figure 4-41b. Light extinction trends in Phoenix, Arizona from 1994 to 2001.
Source: Arizona Department of Environmental Quality (2002).
4-184
-------
(Arizona Department of Environmental Quality, 2002). Daytime visibility is worst during the
morning. Samples taken from 1994 to 1996 showed that organic and elemental carbon
dominated visibility impairment during the dirtiest and cleanest days, whereas ammonium nitrate
was an important contributor during the 20% dirtiest days (Neuroth and Heisler, 2000).
The State of Virginia compared visibility trends at five locations both with and without
pre-1948 airport visual range data. Average annual airport visual-range-data from Roanoke
(1936 to 1998), DC Reagan National (1930 to 1998), Richmond (1942 to 1998), Lynchburg
(1935 to 1998), and Elkins, WV (1936 to 1994) were used in the analysis. Trends in visibility
conditions were dependent on the baseline year used for visibility measurements. When airport
visual range data from 1948 are used as the baseline data for visibility conditions, the trend in
visual range shows declining conditions for all sites. When pre-1948 data are included in the
analysis, visibility conditions basically are unchanged or improved (except for Elkins). When
1948 data are used as the baseline for visibility trends measurements, visual range decreases at
all locations. Average annual visual ranged varied between 7+ miles and 8+ miles for all sites
(Virginia Climate Advisory, 2000).
4.3.6.2 Current Conditions
Current visibility conditions have been well-characterized for Class I areas using updated
data from the IMPROVE network (U.S. Environmental Protection Agency, 200la; Malm, 2000;
IMPROVE, 1998). During recent decades, daytime visibility conditions at all major airports
throughout the United States were recorded hourly by human observation. These data were used
to determine current visibility conditions and visibility trends in the United States, as well as the
spatial distribution of visibility conditions (Trijonis et al., 1991). The use of human observation
is being replaced by an automated observation system, the Automated Surface Observing System
(ASOS). More than 900 airports are currently commissioned. In addition, the EPA has
deployed a new national monitoring network to assess PM2 5 concentrations and composition.
More detailed information on visibility conditions for urban and suburban areas will
become more widely available as data from the national PM2 5 speciation monitoring network
and the ASOS airport visibility network are further analyzed. Efforts are currently underway to
develop a web-based system to allow the use of the high resolution ASOS data in air quality
4-185
-------
monitoring and assessment programs. The objectives are to collect and quality control an
archive of the ASOS visibility data, to deliver processed hourly visibility data to public and air
quality communities, and to use the web-based system to support the acquisition and
dissemination of visibility data (Falke, 2001).
4.3.7 Societal Impacts of Particulate Matter Visibility Effects
Society recognizes the need to impose remedies for repairing and preventing further
anthropogenic pollutant-related effects on visibility conditions (Ely et al., 1991; Davidson
et al., 2000a). Information about how the public perceives and values improvements in visibility
comes from both economic studies and from local or state initiatives to adopt local visibility
goals and strategies.
4.3.7.1 Economic Studies
Various methods have been utilized to help determine the economic valuation of changes
in visibility. Where possible, direct economic valuation can be determined using marketplace
cost estimates. Avoided-cost methods estimate the costs of pollution by using the expenditures
that are made necessary by pollution damage. As an example, if ambient levels of PM result in
increased frequency of building cleaning or repainting, the appropriately calculated increase in
these costs is a reasonable estimate of true economic damage. Benefits associated with
reductions in the pollution levels are then represented by the avoided costs of these damages.
Estimating the benefits of clear skies is a more difficult and less precise exercise, because,
although the public values aesthetic views, they are not directly bought and sold in the
marketplace. However, there are several methods available to economists to estimate the
economic impact of these kinds of changes in environmental conditions (Freeman, 1993). These
methods include hedonic valuation or pricing, contingent valuation and contingent choice, and
travel cost (Johnson and Desvousges, 1997; Hanley and Spash, 1993). The primary methods
used to date for the valuation of visibility have been the hedonic price and contingent valuation
methods (Hanley and Spash, 1993). However, this is not an exact science, and there are still
issues and limitations associated with each of these methods.
4-186
-------
Hedonic pricing can be used to estimate economic valuations for environmental effects that
have a direct effect on market values. It relies on the measurement of differentials in property
values under various environmental quality conditions including air pollution and environmental
amenities, such as aesthetic views. The hedonic method works by analyzing the way that market
prices change with changes in environmental quality. Part of the economic costs imposed by
PM-related reductions in visibility can be estimated by looking at the differences in sales price
between otherwise identical houses that have different degrees of visibility impairment.
The contingent valuation method (CVM) is the most widely used method for estimating
value changes in both visibility and ecosystem functions (Hanley and Spash, 1993; Chestnut,
1997; Watson and Chow, 1993). The CVM creates hypothetical markets for goods and services
that have no market-determined price. As part of the evaluation, individuals are shown
photographs with perceivable differences in visibility levels. Carefully structured surveys are
administered to estimate the amount of compensation equivalent to a given change in
environmental quality or, equivalently, how much an individual would be willing to pay (WTP)
for improvements in environmental quality or willing to accept (WTA) existing conditions
without further deterioration. There is an extensive scientific literature and body of practice on
both this theory and technique; however, there are still concerns about the use of this technique
for quantitative purposes.
The travel-cost method estimates can be used to estimate the value of recreational benefits
of an ecosystem based on the environmental quality at the site. The travel-cost method uses
information on actual behavior rather than responses to hypothetical scenarios. The time and
travel expenses incurred to visit a site represents the price of access to the site. The willingness
to pay to maintain the site is determined by the number of times the individual visits the site at
different travel costs.
The effects of PM on visibility may differ widely between urban residential and
recreational areas. Therefore, separate estimates are needed to account for impacts associated
with changes in visibility in residential and recreational (Class I) areas. Chestnut and Dennis
(1997) compared the findings of the more recent studies on the economic impact of changes in
regional haze using the contingent valuation method in residential areas in several eastern cities
and in Los Angeles and San Francisco and using the hedonic value method in Los Angeles and
4-187
-------
San Francisco. The findings of the contingent valuation studies are shown in Table 4-16.
Findings using the contingent and hedonic methods of valuation for Los Angeles and San
Francisco are compared in Table 4-17.
Using the contingent valuation method, Chestnut and Rowe (1990) found that 83% of those
individuals responding to a survey on visibility were willing to pay to improve visibility in the
national parks. Survey participants were selected from California, Arizona, Missouri,
New York, and Virginia. The national parks from three regions (California, southwestern United
States, and southeastern United States) were considered in different versions of the survey.
The survey included questions on past and future visitations to national parks, potential pollution
effects from human activities outside of the park, three hypothetical visibility scenarios,
socioeconomic characteristics, and various photographic presentations of visibility conditions
within the parks. Higher responses were noted for residents residing in the state or region where
the national park was located; responses for males and the elderly were generally lower and there
was a direct correlation between household income and the response.
Using the results from the Chestnut and Rowe (1990) study, Chestnut and Dennis (1997)
calculated an extinction coefficient of 85 for in-state residents and 50 for out-of-state residents.
These extinction coefficients were suggested to represent an annual willingness to pay per
household of $15 and $9, in 1994 dollars, for a 20% improvement in visual range.
4.3.7.2 Public Perception and Attitude Studies
An initiative in Denver to address public concerns about visibility impairment began with a
series of visibility-related studies in the 1970s through the 1980s, leading to the adoption of a
visibility standard for the city of Denver in 1990. This standard is based on a light extinction
level of 0.076 km"1, averaged over four daylight hours, reflecting the short-term nature of the
perception of changes in visibility conditions. This standard is equivalent to a visual range of
approximately 50 km. The approach used to develop this standard relied upon citizen judgments
about acceptable and unacceptable levels of visual air quality (Ely et al., 1991). In Phoenix,
a study was conducted between 1988 and 1990 to address degrading visibility conditions (Chow
et al., 1990). This work led the Arizona Department of Environmental Quality to establish a
Blue Sky Index, which focuses on days in which the visual range, averaged over six daylight
hours, is 40 km or more. This target is based on a method very similar to that used in Denver for
4-188
-------
TABLE 4-16. RESIDENTIAL VISIBILITY CONTINGENT VALUATION
STUDY RESULTS
City
Atlanta
Chicago
Chicago
Atlanta
Boston
Mobile
Washington, DC
Cincinnati
Miami
Cincinnati
Mean WTP
($1994)
Unadj. $44
Partial $28
Full $20
-$361
$346
$430
-$301
$289
$432
-$222
$212
$262
-$240
$257
$302
-$356
$366
$465
-$88
$87
$98
-$152
$136
$160
$198
Starting-Ending
Visual Range (VR)
(miles)
17.6-20
38233
38247
38259
38327
38342
12-32
18-13
18-28
18-38
38264
38279
38289
15-10
15-25
15-35
38233
38248
38258
38359
38370
38380
11.4-16.4
Extinction
Coefficient
346
222
159
416
469
422
312
635
120
256
602
WTP for 20%
Change in VR
($) Reference
$63 McClelland
etal. (1993)
$76 Tolley et al.
(1986)
$86
$77
$57
$116
$22
$47
$110 Rae(1983)
Source: Chestnut and Dennis (1997).
4-189
-------
TABLE 4-17. RESIDENTIAL VISIBILITY VALUATION STUDY RESULTS FOR
LOS ANGELES AND SAN FRANCISCO
City
Los Angeles
San Francisco
Method
CVM
Property
Value
CVM
Property
Value
Mean
WTP
$130
$333
$183
$211
$124
Starting-Ending
Visual Range Extinction
(VR) (miles) Coefficient
2-12 1191328
2-28
12-28
18.6-16.3
16.3-18.3
WTP for
20% Change
inVR
$22
$242
$245-647
$496-552
Reference
Brookshire
et al. (1979)
Trijonis et al.
(1985)
Loehman et al.
(1985)
Trijonis et al.
(1985)
Adapted from: Chestnut and Dennis (1997).
obtaining citizens'judgments as to acceptable levels of visual air quality (BBC Research and
Consulting, 2003). While in practice these standard target values are exceeded many times per
year in these areas, they reflect a reasonable degree of consistency in the outcome of the
approach used to characterize the value that citizens in these two urban areas place on visual
air quality.
Studies conducted in other locations have resulted in similar visibility threshold
determinations, convergent on a minimal visual range of 40 to 60 km. For example, using a
methodology similar to that used by Ely et al. (1991) in Denver, Pry or (1996) conducted a study
of the perception of acceptable visibility conditions in two suburban locations in the lower Fraser
Valley in British Columbia, CN. This study reflected public judgments about the acceptability
of visual ranges of approximately 40 km to 60 km. Other notable visibility protection initiatives
have resulted in state visibility standards that are within this range, including standards for the
Lake Tahoe area (Molenar, 2000) and the state of Vermont.
4-190
-------
4.4 PARTICULATE MATTER EFFECTS ON MATERIALS
Effects of air pollution on materials are related to both aesthetic appeal and physical
damage. Studies have demonstrated that particles, primarily carbonaceous compounds, cause
soiling of commonly used building materials and culturally important items, such as statutes and
works of art. Physical damage from the dry deposition of air pollutants, such as PM (especially
sulfates and nitrates) and SO2, and the absorption or adsorption of corrosive agents on deposited
particles also can result in the acceleration of naturally occurring weathering processes of
man-made building and cultural materials.
In the atmosphere, PM may be "primary," existing in the same form in which it was
emitted, or "secondary," formed by the chemical reactions of free, absorbed, or dissolved gases.
The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen ions;
particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
material. A substantial fraction of the fine particle mass, particularly during the warmer months,
is secondary sulfate and nitrate. Sulfates may be formed by the gas-phase conversion of SO2 to
H2SO4 by OH radicals and aqueous-phase reactions of SO2 with H2O2, O3, or O2. During the day,
NO2 may be converted to nitric acid (HNO3) by reacting with OH radicals. Nitrogen dioxide
also can be oxidized to HNO3 by a sequence of reactions initiated by O3. A more detailed
discussion of the atmospheric chemistry of PM appears in Chapter 2 of this document.
4.4.1 Corrosive Effects of Particles and Sulfur Dioxide on
Man-Made Surfaces
This section (a) summarizes information on exposure-related effects on materials
associated with sulfur-containing pollutants (formed by chemical reactions of SO2 with other
atmospheric pollutants) as addressed in the 1996 PM AQCD (U.S. Environmental Protection
Agency, 1996a) and (b) presents relevant information derived from very limited research
conducted and published since completion of that document. The effects of nitrates on
man-made building materials and naturally occurring cultural materials were discussed in an
earlier EPA Nitrogen Oxides Criteria Document (U.S. Environmental Protection Agency, 1993).
4-191
-------
4.4.1.1 Metals
Metals undergo natural weathering processes in the absence of environmental pollutants.
The additive effects of pollutants on the natural weathering processes depend on the nature of the
pollutant, the deposition rate (the uptake of a pollutant by the material's surface), and the
presence of moisture. The influence of the metal-protective corrosion film, the presence of other
surface electrolytes, the orientation of the metal surface, the presence of surface moisture, and
the variability in the electrochemical reactions also contributes to pollutant exposure effects on
metal surfaces.
Several studies demonstrate the importance of the duration of surface wetness (caused by
dew and fog condensation and rain) on metals. Surface moisture facilitates the deposition of
pollutants, especially SO2, and promotes corrosive electrochemical reactions on metals (Haynie
and Upham, 1974; Sydberger and Ericsson, 1977). Of critical importance is the formation of
hygroscopic salts on the metal, increasing the duration of surface wetness and, thereby,
enhancing the corrosion process.
The effect of temperature on the rate of corrosion is complex. Under normal temperature
conditions, temperature does not have an effect on the rate of corrosion; but when the
temperature decreases, the relative humidity increases and the diffusivity decreases. Pitchford
and McMurry (1994) and Zhang et al. (1993) demonstrated particle-size-related effects of
relative humidity. The corrosion rate decreases as the temperature approaches freezing, because
ice inhibits the diffusion of SO2 to the metal surface and minimizes electrochemical processes
(Haynie, 1980; Biefer, 1981; Sereda, 1974).
The metal-protective corrosion film (i.e., the rust layer on metal surfaces) provides some
protection against further corrosion. The effectiveness of the corrosion film in slowing down the
corrosion process is affected by the solubility of the corrosion layer and the concentration and
deposition rate of pollutants. If the metal-protective corrosion film is insoluble, it may add some
protection against acidic pollutants. An atmospheric corrosion model that considers the
formation and dissolution of the corrosion film on galvanized steel was proposed by Spence
et al. (1992). The model considers the effects of SO2, rain acidity, and duration of wetness on
the rate of corrosion. Although the model does not specifically characterize particle effects, the
contribution of particulate sulfate was considered in model development.
4-192
-------
Whether suspended particles actually increase metal corrosion is not clear. Several studies
have suggested that suspended particles promote the corrosion of metals (Goodwin et al., 1969;
Barton, 1958; Sanyal and Singhania, 1956; Baedecker et al., 1991); however, other studies have
not demonstrated a correlation between particle exposure and metal corrosion (Mansfeld, 1980;
Edney et al., 1989). Walton et al. (1982) suggested that catalytic species within several species
in fly ash promote the oxidation of SOX to a corrosive state. Still other researchers indicate that
the catalytic effect of particles is not significant and that the corrosion rate is dependent on the
conductance of the thin-film surface electrolytes during periods of wetness. Soluble particles
likely increase the solution conductance (Skerry et al., 1988; Askey et al., 1993).
The corrosion of most ferrous metals (iron, steel, and steel alloys) increases with increasing
SO2 exposure. Steels are susceptible to corrosion when exposed to SO2 in the absence of
protective organic or metallic coatings. Studies on the corrosive effects of SO2 on steel indicate
that the rate of corrosion increases with increasing SO2 and is dependent on the deposition rate of
the SO2 (Baedecker et al., 1991; Butlin et al., 1992a). The corrosive effects of SO2 on aluminum
is exposure-dependent, but appears to be insignificant (Haynie, 1976; Fink et al., 1971; Butlin
et al., 1992a). The rate of formation of the patina (protective covering) on copper can take as
long as five years and is dependent on the SO2 concentration, deposition rate, temperature, and
relative humidity (Simpson and Horrobin, 1970). Further corrosion is controlled by the
availability of copper to react with deposited pollutants (Graedel et al., 1987). Butlin et al.
(1992a), Baedecker et al. (1991), and Cramer et al. (1989) reported an average corrosion rate of
1 (im/year for copper; however, less than a third of the corrosion was attributed to SO2 exposure,
suggesting that the rate of patina formation was more dependent on factors other than SO2.
A report by Strandberg and Johansson (1997) showed relative humidity to be the primary factor
in copper corrosion and patina formation. The results of the studies on particles and SO2
corrosion of metals are summarized in Table 4-18.
4.4.1.2 Painted Finishes
Exposure to air pollutants affects the durability of paint finishes by promoting
discoloration, chalking, loss of gloss, erosion, blistering, and peeling. Evidence exists that
particles can damage painted finishes by serving as carriers for corrosive pollutants (Cowling
4-193
-------
TABLE 4-18. CORROSIVE EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON METALS
Metal
Exposure Conditions
Comments
Source
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 2.1-60
ug/m3. Annual average concentrations were about
20 ug/m3. Meteorological conditions were
unaltered. Specimens exposed at 29 sites for 2 years
for mild steel and 1 year 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 1-5 years.
Average SO2 concentrations ranged from 2 ± 4 to
15 ± 17 ppb (5.2 ± 10.4 to 39.3 ± 44.5 ug/m3).
PM concentrations ranged from 14 to 60 ug/m3.
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 urn/year
for mild steel specimens for most industrial sites,
but ranged from 21 to 71 urn/year. The corrosion rate
ranged from 1.45 to 4.25 urn/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 urn/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 0.63-1.33 um per
year. 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 6.6-12.8 urn/year for carbon steel
and 3.7-5.0 urn/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 0.036-0.106 urn/year.
Corrosion greater on the underside of specimens,
possibly because of lack of washoff and increased
PM in area. Maximum corrosion rate was
0.85 urn/year. Pit depths of up to 72 um were
noted after 2 years of exposure.
Butlin etal. (1992a)
Showak and Dunbar
(1982)
Baedecker et al. (1991)
Cramer etal. (1989)
Baedecker et al. (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
Metal
Exposure Conditions
Comments
Source
Copper
Copper
Copper
Copper
Iron
See Baedecker et al. (1991) above for exposure
conditions.
See Butlin et al. (1992a) above for exposure
conditions.
Specimens exposed to 4-69 ppb (10.4-180.7 ug/m3)
and 1.0 ppm (2,618.7 ug/m3) SO2 for 20 h at
various relative humidities.
Specimens exposed artificially to 0.49 ± 0.01 ppm
(187 ± 3.8 ug/m3) SO2 for 4 weeks at 70 and 90%
relative humidity.
Specimens from restorations of Acropolis
monuments over many years. The oldest specimens
were 142 years old. Other specimens used for
monument restoration after 1950.
Average corrosion rate for 3- and 5-year exposures
was about 1 urn/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 urn/year. The corrosion rate was 1.48 urn/year
at the site receiving the most rainfall. The lowest
corrosion rate, 0.66 urn/year, was associated with
low rainfall, low SO2.
SO2 had no effect on copper when relative humidity
was < 75%. 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.
Specimens used after 1950s had an oxidation rate
25% greater than those specimens before 1950.
The accelerated oxidation had a negative impact
on the structure, producing a quicker formation of
cracks inside the marble.
Baedecker et al.
(1991)
Butlin etal. (1992a)
Strandberg and
Johansson (1997)
Eriksson etal. (1993)
Zuburtikudis and
Triantafyllou(2001)
-------
and Roberts, 1954) or by staining and pitting of the painted surfaces (Fochtman and Langer,
1957; Wolff etal., 1990).
The erosion rate of oil-based house paint has been reported to be enhanced by exposure to
SO2 and high humidity. An erosion rate of 36.71 ± 8.03 jim/year was noted for oil-based house
paint samples exposed to SO2 (78.6 |ig/m3), O3 (156.8 |ig/m3), andNO2 (94 |ig/m3), and low
humidity (50%) (Spence et al., 1975). The erosion rate increased with increased SO2 and
humidity. The authors concluded that SO2 and humidity accounted for 61% of the erosion.
Acrylic and vinyl coil coatings show less pollutant-related erosion. Erosion rates range from
0.7 to 1.3 jam/year and 1.4 to 5.3 jam/year, respectively. Similar findings on SO2-related erosion
of oil-based house paints and coil coatings have been reported by other researchers (Davis et al.,
1990; Yocom and Grappone, 1976; Yocom and Upham, 1977; Campbell et al., 1974). Several
studies have suggested that the effect of SO2 is caused by its reaction with extender pigments
such as calcium carbonate and zinc oxide (Campbell et al., 1974; Xu and Balik, 1989; Edney,
1989; Edney et al., 1988, 1989). However, Miller et al. (1992) suggested that calcium carbonate
acts to protect paint substrates. Another study indicated that exposure to SO2 can increase the
drying time of some paints by reacting with certain drying oils and will compete with the
auto-oxidative curing mechanism responsible for crosslinking the binder (Holbrow, 1962).
4.4.1.3 Stone and Concrete
Numerous studies suggest that air pollutants can enhance the natural weathering processes
on building stone. The development of crusts on stone monuments has been attributed to the
interaction of the stone's surface with sulfur-containing pollutants, wet or dry deposition of
atmospheric particles, and dry deposition of gypsum particles from the atmosphere. Because of
a greater porosity and specific surface, mortars have a greater potential for reacting with
environmental pollutants (Zappia et al., 1998). Details on these studies are discussed in
Table 4-19. The stones most susceptible to the deteriorating effects of sulfur-containing
pollutants are the calcareous stones (limestone, marble, and carbonated cement). Exposure-
related damage to building stones result from the formation of salts in the stone that are
subsequently washed away during rain events, leaving the stone surface more susceptible to the
effects of pollutants. Dry deposition of sulfur-containing pollutants promotes the formation of
gypsum on the stone's surface. Gypsum is a gray to black crusty material comprised mainly
4-196
-------
TABLE 4-19. CORROSIVE EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON STONE
Stone
Vermont marble
Marble sandstone
Exposure Conditions
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.
Comments
Between 10 and 50% of calcium in runoff water estimated from
gypsum formation from dry deposition of SO2.
Pollutant exposure related erosion was primarily caused by dry
deposition of SO2 and nitric acid between rain events and wet deposition
Source
Schuster et al.
(1994)
Baedecker
etal. (1992)
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
Ambient air conditions. Exposure ranged from
70-1,065 days. Averaged pollutant exposure
ranged from 1.4-20.4 ppb (3.7-53.4 ug/m3) SO2;
4.1-41.1 ppbNOx; 2.4-17.4 ppb (4.5-32.7 ug/m3)
NO2; 10.1-25.6 ppb (19.8-50.2 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.
of hydrogen ion. Recession estimates ranged from 15-30 urn/year for
marble and 25-45 um/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 etal.
significantly affect stone degradation. Stone loss associated with (1992)
SO2 exposure estimated to be 24 um/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 (1992b)
the weight changes.
Insignificant differences in erosion rate found between calcite Petuskey et al.
and 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 um thick. et al. (1998)
Innermost layer, ranging from brown to orange in color, primarily
consisted of calcite, between 10 and 600 um thick. Gypsum-rich
layer thought to be the result of sulfation of the calcitic layer by
atmospheric pollutants or dry or wet deposition of atmospheric dust.
In the absence of moisture, little reaction is seen. Sulfur dioxide is Haneef et al.
oxidized to sulfates in the presence of moisture. The effect is enhanced (1993)
in the 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
vo
oo
Monk's Park limestone
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 months under both sheltered and
unsheltered conditions. Mean daily atmospheric SO2
concentration was 68.7 |ig/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 months.
Samples exposed for 6 months (cold and hot conditions)
in ambient environment. PM concentrations ranged from
57.3 to 116.7 ng/m3 (site 1) and 88 to 189.8 ng/m3 (site 2).
Some exposures also were associated with high SO2, NO,
andNO2.
Samples artificially exposed to fly-ash containing
1,309.3 ng/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 the study.
Samples exposed to 7,856 |ig/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 stone. Viles (1990)
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 content (1996)
of particles.
Carrara marble found to be more reactive with SO2 than Georgia Yerrapragada
marble, possibly because of the compactness of the Georgia et al. (1994)
marble. Greater effects noted when samples were also exposed
to N02.
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 PM.
Exposure to fly-ash did not enhance oxidation of SO2 to sulfates. Hutchinson
Mineral oxides in fly ash contributed to sulphation of CaCO3. et al. (1992)
Exposure to SO2 produced significant quantities of calcium sulfite Zappia et al.
and calcium sulfate on specimens; however, the amount produced (1994 )
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 PARTICIPATE MATTER AND SULFUR DIOXIDE ON STONE
Stone
Exposure Conditions
Comments
Source
vo
vo
Limestone
Quartz-cemented
sandstone
Calcite-cemented
sandstone
Granite
Brick
Carrara marble
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 layer on surface.
Samples from structures taken from monuments in
Venice, Italy. Samples included those shielded
from running water (unwashed), areas at the
interface of the washed and unwashed areas, and
washed areas. Samples of fog and rain near the
monument were also taken for analysis.
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
Palazzo 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 months 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, Nord and
silicate, soot, and dirt. Ericsson
(1993)
Sulfate was higher in fog compared to other ions. The average concentration Fassina et al.
of chlorides and sulfates were 7-16 times higher in fog than rain. The degree (2001)
of sulphonation on stone samples from areas shielded from running water was
< 40%. At the interface between washed and unwashed surfaces and at washed
areas, sulphonation was > 40%. Dendrite shaped crust from the transformation
of calcium carbonate into gypsum was found on samples shielded from rain.
Samples from the washed areas displayed superficial granular disaggregation,
a natural process of deterioration ascribed to natural agents; however, the
process was accelerated due to the ambient air sulfates.
Crust samples contained calcite, soil dust, carbonaceous particles, and gypsum Ausset et al.
crystals. (1998)
Mortars were more reactive than the stones. Of the mortars, cement and Zappia et al.
pozzolan mortar were more reactive than the lime mortar. Carrara marble (1998)
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 Sabbioni et al.
layers on the mortar: an outer thin surface black crust composed of gypsum (1998)
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 Bugini et al.
oxides. Estimated rate of crust formation was 2-5 urn/year. Total amount of (2000)
gypsum formed over the lifetime of exposure was 5-13 mg/cm2, an estimated
0.2 mg/cm2/year.
-------
of calcium sulfate dihydrate from the reaction of calcium carbonate (calcite) in the stone with
atmospheric SO2 and moisture (relative humidities exceeding 65%) according to the following
reaction.
CaCO3 (marble) + SO2 + 2H2O -» CaSO4 2H2O + CO2 (4-12)
The sulfate anions formed in the moist air reacts with the Ca2+ through diffusion processes
forming the gypsum (Zuburtikudis and Triantafyllou, 2001). Approximately 99% of the sulfur
in gypsum is sulfate because of the sulfonation process caused by the deposition of SO2 aerosol.
Sulfites also are present in the gypsum layer as an intermediate product (Sabbioni et al., 1996;
Ghedini et al., 2000; Gobbi et al., 1998; Zappia et al., 1998). Gypsum is more soluble than
calcite and is known to form on limestone, sandstones, and marble when exposed to SO2.
Gypsum also has been reported to form on granite stone by replacing silicate minerals with
calcite (Schiavon et al., 1995). Gypsum occupies a larger volume than the original stone,
causing the stone's surface to become cracked and pitted. The rough surface serves as a site for
deposition of airborne particles. As the gypsum grows, it loosens and falls apart (Zuburtikudis
and Triantafyllou, 2001).
The dark colored gypsum is created by surface deposition of carbonaceous particles
(noncarbonate carbon) from combustion processes occurring in the area (Sabbioni, 1995;
Saiz-Jimenez, 1993; Ausset et al., 1998; Hermosin and Saiz-Jimenez, 2000), trace metals
contained in the stone, dust, and numerous other anthropogenic pollutants. After analyzing
damaged layers of several stone monuments, Zappia et al. (1993) found that the dark-colored
damaged surfaces contained 70% gypsum and 20% noncarbonate carbon. The lighter colored,
damaged layers were exposed to rain and contained 1% gypsum and 4% noncarbonate carbon.
It is assumed that rain removes reaction products, permitting further pollutant attack on the stone
monument and likely redeposits some of the reaction products at rain runoff sites on the stone.
After sulfur compounds, carbon was reported to be the next highest element in the dark crust on
historical monuments in Rome. Elemental carbon and organic carbon accounted for 8 and 39%
of the total carbon in the black crust samples. The highest percentage of carbon, carbonate
carbon, was derived from the carbonate matrix in the stones. The high ratio of organic carbon to
elemental carbon indicates the presence of a carbon source other than combustion processes
4-200
-------
(Ghedini et al., 2000). Cooke and Gibbs (1994) suggested that stones damaged during times of
higher ambient pollution exposure likely would continue to exhibit a higher rate of decay,
termed the "memory effect," than newer stones exposed under lower pollution conditions.
Increased stone damage also has been associated with the presence of sulfur-oxidizing bacteria
and fungi on stone surfaces (Garcia-Valles et al., 1998; Young, 1996; Saiz-Jimenez, 1993;
Diakumaku et al., 1995).
Dissolution of gypsum on the stone's surface initiates structural changes in the crust layer.
Garica-Valles et al. (1998) proposed a double mechanism: the dissolution of the gypsum, in the
presence of sufficient moisture, followed by recrystallization inside fissures or pores. In the
event of limited moisture, the gypsum is dissolved and recrystallizes at its original location.
According to the authors, this would explain the gypsum-rich crustal materials on stone surfaces
sheltered from precipitation.
Moisture was found to be the dominant factor in stone deterioration for several sandstones
(Petuskey et al., 1995). Dolske (1995) reported that the deteriorative effects of sulfur-containing
rain events, sulfates, and SO2 on marble were largely dependent on the shape of the monument
or structure rather than the type of marble. The author attributed the increased fluid turbulence
over a non-flat vertical surface versus a flat surface to the increased erosion. Sulfur-containing
particles also have been reported to enhance the reactivity of Carrara marble and Travertine and
Trani stone to SO2 (Sabbioni et al., 1992). Particles with the highest carbon content had the
lowest reactivity.
The rate of stone deterioration is determined by the pollutant and the pollutant
concentration, the stone's permeability and moisture content, and the pollutant deposition
velocity. Dry deposition of SO2 between rain events has also been reported to be a major
causative factor in pollutant-related erosion of calcareous stones (Baedecker et al., 1991; Dolske,
1995; Cooke and Gibbs, 1994; Schuster et al., 1994; Hamilton et al., 1995; Webb et al., 1992).
Sulfur dioxide deposition increases with increasing relative humidity (Spiker et al., 1992), but
the pollutant deposition velocity is dependent on the stone type (Wittenburg and Dannecker,
1992), the porosity of the stone, and the presence of hygroscopic contaminants.
Although it is clear from the available information that gaseous pollutants (dry deposition
of SO2 in particular) will promote the decay of some types of stones under specific conditions,
carbonaceous particles (noncarbonate carbon) may help to promote the decay process by aiding
4-201
-------
in SO2 transformation to a more acidic species (Del Monte and Vittori, 1985). Several authors
have reported enhanced sulfation of calcareous material by SO2 in the presence of particles
containing metal oxides (Sabbioni et al., 1996; Hutchinson et al., 1992).
4.4.2 Soiling and Discoloration of Man-Made Surfaces
Ambient particles can cause soiling of man-made surfaces. Soiling has been defined as the
deposition of particles of less than 10 jim on surfaces by impingement. Soiling generally is
considered an optical effect, that is, soiling changes the reflectance from opaque materials and
reduces the transmission of light through transparent materials. Soiling can represent a
significant detrimental effect, requiring increased frequency of cleaning of glass windows and
concrete structures, washing and repainting of structures, and, in some cases, reducing the useful
life of the object. Particles, especially carbon, may also help catalyze chemical reactions that
result in the deterioration of materials during exposure.
It is difficult to determine the accumulated particle levels that cause an increase in soiling.
Soiling is dependent on the particle concentration in the ambient environment, particle size
distribution, the deposition rate, and the horizontal or vertical orientation and texture of the
surface being exposed (Haynie, 1986). The chemical composition and morphology of the
particles and the optical properties of the surface being soiled will determine the time at which
soiling is perceived (Nazaroff and Cass, 1991). Carey (1959) reported that the average observer
could observe a 0.2% surface coverage of black particles on a white background. Work by
Bellan et al. (2000) indicated that it would take a 12% surface coverage by black particles before
there is 100% accuracy in identifying soiling. Sharpies et al. (2001) studied the effect of air
pollution, moisture, and the function of the room/building on glazing daylight transmittance for a
number of building windows. They found that the direct soiling load to a window was
dependent on the immediate external and internal environment. For instance, there was only a
10% reduction in daylight transmittance for windows from an office building that had not been
cleaned for 5 years compared to clean windows. The reduction in transmittance for windows in
a swimming pool complex was in excess of 20% due to soiling of the interior surface. For most
office buildings, there was a reduction of glazing transmittance ranging from 3 to 10%, with
most windows showing about a 3% reduction. The rate at which an object is soiled increases
linearly with time; however, as the soiling level increases, the rate of soiling decreases. The
4-202
-------
buildup of particles on a horizontal surface is counterbalanced by an equal and opposite
depletion process, which is based on the scouring and washing effects of wind and rain (Schwar,
1998).
4.4.2.1 Stones and Concrete
Most of the research evaluating the effects of air pollutants on stone structures has
concentrated on gaseous pollutants. The deposition of the sulfur-containing pollutants is
associated with the formation of gypsum on the stone (see Section 4.5.1.3). The dark color of
gypsum is attributed to soiling by carbonaceous particles from nearby combustion processes.
A lighter gray colored crust is attributed to soil dust and metal deposits (Ausset et al., 1998;
Camuffo, 1995; Moropoulou et al., 1998). Realini et al. (1995) recorded the formation of a dark
gypsum layer and a loss of luminous reflection in Carrara marble structures exposed for 1 year
under ambient air conditions. Dark areas of gypsum were found by McGee and Mossitti (1992)
on limestone and marble specimens exposed under ambient air conditions for several years. The
black layers of gypsum were located in areas shielded from rainfall; whereas particles of dirt
were concentrated around the edges of the gypsum formations. Lorusso et al. (1997) attributed
the need for frequent cleaning and restoration of historic monuments in Rome to exposure to
total suspended particulates. They also concluded that, based on a decrease in brightness
(graying), surfaces are soiled proportionately over time; however, graying is higher on horizontal
surfaces because of sedimented particles. Davidson et al. (2000b) evaluated the effects of air
pollution exposure on a limestone structure on the University of Pittsburgh campus, using
estimated average TSP levels in the 1930s and 1940s and actual values for the years 1957 to
1997. Monitored levels of SO2 were available for the years 1980 to 1998. Based on the
available data concerning pollutant levels and photographs, it was thought that soiling began
while the structure was under construction. With decreasing levels of pollution, the soiled areas
have been slowly washed away, the process taking several decades, leaving a white, eroded
surface.
4.4.2.2 Household and Industrial Paints
Few studies are available that evaluate the soiling effects of particles on painted surfaces.
Particles composed of elemental carbon, tarry acids, and various other constituents are
4-203
-------
responsible for the soiling of structural painted surfaces. Coarse-mode particles (> 2.5 jim)
initially contribute more soiling of horizontal and vertical painted surfaces than do fine-mode
particles (< 2.5 |im), but are more easily removed by rain (Haynie and Lemmons, 1990). The
accumulation of fine particles likely promotes remedial action (i.e., cleaning of the painted
surfaces); whereas coarse-mode particles are primarily responsible for soiling of horizontal
surfaces. Rain interacts with coarse particles, dissolving the particle and leaving stains on the
painted surface (Creighton et al., 1990; Haynie and Lemmons, 1990). Haynie and Lemmons
(1990) proposed empirical predictive equations for changes in surface reflectance of
gloss-painted surfaces that were exposed protected and unprotected from rain while oriented
horizontally or vertically.
Early studies by Parker (1955) and Spence and Haynie (1972) demonstrated an association
between particle exposure and the increased frequency of cleaning of painted surfaces. Particle
exposures also caused physical damage to the painted surface (Parker, 1955). Unsheltered
painted surfaces are initially more soiled by particles than sheltered surfaces, but the effect is
reduced by rain washing. Reflectivity is decreased more rapidly on glossy paint than on flat
paint (Haynie and Lemmons, 1990). However, surface chalking of the flat paint was reported
during the exposure. The chalking interfered with the reflectance measurements for particle
soiling. Particle composition measurements that were taken during exposure of the painted
surfaces indicated sulfates to be a large fraction of the fine mode and only a small fraction of the
coarse mode. Although no direct measurements were taken, fine-mode particles likely also
contained large amounts of carbon and possibly nitrogen or hydrogen (Haynie and Lemmons,
1990).
4.5 ATMOSPHERIC PARTICULATE MATTER, CLIMATE CHANGE,
AND EFFECTS ON SOLAR UV-B RADIATION TRANSMISSION
Atmospheric particles alter the amount of solar radiation transmitted through the Earth's
atmosphere. The absorption of solar radiation by atmospheric particles, together with the
trapping of infrared radiation emitted by the Earth's surface by certain gases, enhances the
heating of the Earth's surface and lower atmosphere i.e., the widely-known "greenhouse effect."
Increases in the atmospheric concentration of these gases due to human activities may lead to
4-204
-------
impacts, due to climate change, on human health and the environment. Lesser consequences of
airborne particles include alterations in the amount of ultraviolet solar radiation (especially
UV-B, 290 to 315 nm) penetrating through the Earth's atmosphere and reaching its surface
where UV radiation can exert various effects on human health, plant and animal biota, and other
environmental components.
The effects of atmospheric PM on the transmission of electromagnetic radiation emitted by
the sun at ultraviolet and visible wavelengths, and by the Earth at infrared wavelengths, depend
on the radiative properties (extinction efficiency, single-scattering albedo, and asymmetry
parameter) of the particles, which depend, in turn, on the size and shape of the particles, the
composition of the particles, and the distribution of components within individual particles.
In general, the radiative properties of particles are size- and wavelength-dependent, with the
extinction cross section tending toward its maximum when the particle radius is similar to the
wavelength of the incident radiation. This means that fine particles present mainly in the
accumulation mode would be expected to exert a greater influence on the transmission of
electromagnetic radiation than would coarse particles. The chemical components of particles
can be crudely summarized in terms of the broad classes identified in Chapter 2 of this
document. These classes include nitrate, sulfate, mineral dust, elemental carbon, organic carbon
compounds (e.g., PAHs), and metals derived from high-temperature combustion or smelting
processes. The major sources of these components are shown in Table 3-9 of Chapter 3 in this
document.
Knowledge of the effects of PM on the transfer of radiation in the visible and infrared
spectral regions is needed for assessing relationships between particles and climate change
processes, as well as environmental and biological effects. Knowledge of the factors controlling
the transfer of solar radiation in the ultraviolet spectral range is needed to assess potential the
biological and environmental effects associated with exposure to UV-B radiation. Climate
change processes, their potential to affect human and environmental health, and their potential
relationships to atmospheric PM are discussed below. Solar ultraviolet radiation processes and
the related effects of atmospheric PM are then summarized and discussed in the next section.
4-205
-------
4.5.1 Atmospheric Particle Interactions with Solar and Terrestrial Radiation
Related to Climate Change
4.5.1.1 The Projected Impacts of Global Climate Change
The study of atmospheric processes involved in mediating global climate change and its
potential consequences for human health and global ecosystems is an area of active research.
The most thorough evaluation of currently available science regarding climate change is the
Third Assessment Report (TAR) of the Intergovernmental Panel on Climate Change (IPCC,
200la). Earlier assessments include those conducted by the United Nations Environment
Program (UNEP, 1986), the World Meteorological Organization (WMO, 1988), the EPA (U.S.
Environmental Protection Agency, 1987), and others (e.g., Patz et al., 2000 a,b). The reader
is referred to these documents for more complete discussion of climate change science.
An abbreviated list of the IPCC conclusions, to date, and a short discussion of the potential
impacts of climate change on human health and welfare is provided here to serve as the context
for the discussion of the role of PM in climate.
The IPCC TAR (200la) notes that the increasing body of observations indicates that the
Earth is warming and that other climate changes are underway. These observations include the
global surface temperature record for the time period beginning in the year 1860, the satellite
temperature record begun in 1979, changes in snow and ice cover recorded since the 1950s, sea
level measurements taken throughout the 20th century, and sea surface temperature observations
recorded since the 1950s. Other evidence includes a marked increase over the past 100 years in
the frequency, intensity and persistence of the zonal atmospheric circulation shifts known as the
El Nino-Southern Oscillation (ENSO). ENSO events occur when the tropical ocean has
accumulated a large, localized mass of warm water that interrupts cold surface currents along
South America, altering precipitation and temperature patterns in the tropics, subtropics and the
midlatitudes.
Atmospheric concentrations of greenhouse gases (GHGs), which trap solar energy within
the climate system, are continuing to increase due to human activities. These activities will
continue to change the composition of the atmosphere throughout the 21st century. The IPCC
TAR describes the scientific evidence that ties the observed increase in GHGs over the past
50 years to human activities (IPCC, 200Ib).
4-206
-------
The IPCC (1998, 2001b) reports also describe the results of general circulation model
(GCM) studies that predict that human activities will alter the climate system in a manner that is
likely to lead to marked global and regional changes in temperature, precipitation, and other
climate properties. Global mean sea level is expected to increase, as will the number of extreme
weather events including floods and droughts, thus inducing changes in soil moisture. These
changes will directly impact human health, ecosystems, and global economic sectors, e.g.,
hydrology and water resources, food and fiber production, etc., (IPCC 1998, 200Ib). Table 4-20
summarizes these projected impacts. Wide variations in the course and net impacts of climate
change in different geographic areas can be expected. In general, projected climate change
impacts can be expected to represent additional stressors on those natural ecosystems and human
societal systems already impacted by increasing resource demands, unsustainable resource
management practices, and pollution — with wide variation likely across regions and nations in
their ability to cope with the consequent alterations in ecological balances, in the availability of
adequate food, water, and clean air, and in human health and safety. However, although many
regions are likely to experience severe adverse impacts of climate change (some possibly
irreversible), certain climate change impacts may be beneficial in some regions. For example,
ecosystems that may benefit from warmer temperatures or increased CO2 fertilization include
west coast coniferous forests and some western rangelands. In addition, reduced energy
requirements for heating, reduced road salting and snow-clearance costs, and longer open-water
seasons in northern channels and ports may benefit communities in the northern latitudes. The
IPCC report, "The Regional Impacts of Climate Change" (IPCC, 1998), describes the projected
effects of human-induced climate change on the different regions of the globe, including Africa,
the Arctic and Antarctic, the Middle East and arid Asia, Australasia, Europe, Latin America,
North America, the small island nations, temperate Asia, and tropical Asia.
It must be borne in mind that while current climate models are successful in simulating
present annual mean climate and the seasonal cycle on continental scales, they are less
successful at regional scales. Clouds and humidity, essential factors in defining local and
regional ("sub-grid") climate, are significantly uncertain (IPCC, 2001a). Due to modeling
uncertainties, both in reproducing regional climate and in predicting future economic activity,
the projected impacts discussed above are also uncertain.
4-207
-------
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 Likelihood"
Representative Examples of Projected Impacts'1
(all high confidence of occurrence in some areas")
Simple Extremes
Higher maximum temperatures; more hot days and heat
waves'1 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 waves'1 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, avalanche, 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
Increased tropical cyclone peak wind intensities, mean
and peak precipitation intensities (likely" over some
areas)'
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 hydropower 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)11
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, 2001a).
bThese impacts can be lessened by appropriate response measures.
'High 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 (2001b).
4-208
-------
Findings from the IPCC TAR (200la), Regional Impacts Assessment (1998), and other
regional assessments illustrate well the considerable uncertainties and difficulties in projecting
likely climate change impacts on regional or local scales. The findings predict a mixture of
positive and negative projected potential climate change impacts for U.S. regions differing
widely among regions. Projections of region-specific climate change impacts are complicated
by the need to evaluate potential effects of local- or regional-scale changes in key air pollutants.
4.5.1.2 Airborne Particle Relationships to Global Warming and Climate Change
Atmospheric particles both scatter and absorb incoming solar radiation. Visibility
reduction is caused by particle scattering in all directions, whereas climate effects are mainly
caused by particle scattering in the upward direction. Upward scattering of solar radiation
reduces the total amount of energy received by the Earth system, leading to surface cooling.
The effect on climate due to upward scattering and to absorption of radiation by aerosol can be
roughly quantified as a "radiative forcing" (Houghton et al., 1990). Global and regional climate
(at equilibrium) is defined by the balance between a large number of "positive" and "negative"
forcings induced by different components of the Earth system. The Earth system responds to
these forcings in a potentially complex way due to feedback mechanisms that are theorized but
difficult to model. In the absence of information about climate feedbacks, radiative forcing
values for many components of the climate system are estimated as a tool for approximating
their relative importance in climate change. Forcing estimates for various classes of atmospheric
particles are derived on the basis of climate modeling studies and reported by the IPCC (IPCC,
2001b).
Particulate matter appears to play a significant role in defining climate on both global and
regional scales. Significant reductions over the past 50 years in solar radiation received at the
Earth's surface on a globally averaged basis correlate with increases in atmospheric aerosol
(Stanhill and Cohen, 2001). While this correlation seems clear, quantifying the cooling and
warming effects of aerosol in relation to GHG-related warming is difficult. Aerosols complicate
the prediction of climate change due to their spatial and temporal inhomogeneities and uncertain
radiative properties. However, inclusion of modeled atmospheric sulfate concentrations
substantially improved the agreement between modeled and observed surface temperatures
(Kiehl and Briegleb, 1993). Haywood et al. (1999) also found that the inclusion of
4-209
-------
anthropogenic aerosols results in a significant improvement between calculations of reflected
sunlight at the top of the atmosphere and satellite observations in oceanic regions close to
sources of anthropogenic PM. On a regional scale, the suspected influence of aerosols upon
climate relates to regional hydrological cycles. Evidence is accumulating that pollution aerosols
reduce precipitation frequency by clouds, potentially leading to drought in some parts of the
world (Ramanathan et al., 2001).
Greenhouse Gases, Particulate Matter, and the Earth's Radiative Equilibrium
According to simple radiative transfer theory, at thermal equilibrium, the Earth's surface
should be near -15 °C. This is the temperature of a theoretical "black body" that is receiving
and then re-emitting 342.5 W nT2, the globally averaged amount of solar radiation absorbed and
then re-emitted by the Earth as infrared terrestrial radiation. In fact, satellite observations
indicate that the Earth's average planetary temperature is remarkably close to the theoretical
black body value at -18 °C, a temperature at which liquid water ordinarily does not exist.
At its surface, however, the Earth's average temperature is +15 °C. The 33 °C temperature
differential between Earth's planetary temperature and its surface temperature is due to the
existence of infrared radiation-absorbing components in the atmosphere, i.e., GHGs including
carbon dioxide, methane, several other trace gases, and some types of particles and clouds. The
phenomenon of planetary surface warming due to the atmospheric absorption and re-emission of
infrared radiation is popularly known as the "greenhouse effect" (Arrhenius, 1896; Schneider,
1992). Radiation trapped by the Earth's atmosphere is reflected back to its surface, with some
small fraction absorbed by dark atmospheric particles. The fraction of this radiation that is not
directly re-emitted as long-wave terrestrial radiation transforms into heat energy that drives the
atmospheric processes that form the basis of weather and climate. Eventually this energy is
transformed a second time, to terrestrial radiation, and is re-emitted as part of the process that
maintains Earth's radiative equilibrium.
The evidence of a physical relationship between radiatively active air pollutants (GHGs
and particles) and solar energy retained by the Earth system suggests that continuously changing
atmospheric concentrations of air pollutants, along with other alterations to the climate system
due to human activities, may be shifting the Earth's radiative equilibrium.
4-210
-------
Radiatively active gases in the atmosphere are largely responsible for the greenhouse
effect, although some light-absorbing particles and clouds contribute to atmospheric and surface
warming (IPCC, 200la). The majority of clouds and particles play a role in counteracting the
greenhouse effect by increasing the degree to which the Earth is able to reflect solar radiation,
i.e. its albedo. Successful modeling of the Earth's climate and, therefore, assessment of the
degree of human-induced climate change and development of appropriate policy depends on the
high quality information on the relative efficiencies, amounts, and spatial and temporal
distributions of the various radiatively active components of the atmosphere involved in
absorbing and/or reflecting solar and terrestrial radiation.
"Forcing" and the Earth's Radiative Balance
A measure of the relative influence of a given component of the climate system on the
Earth's radiative balance is its "radiative forcing". Radiative forcing, in W nT2, is a quantity that
was developed by the climate modeling community as a first order-only means of estimating
relative effects of anthropogenic and natural processes on the surface-troposphere system.
No more precise metric has yet been found to replace radiative forcing as a measure of impact of
upon global climate (IPCC, 200la). The convention for this quantity assigns a negative forcing
to climate system components that reflect solar radiation back into space and assigns a positive
forcing to those that enhance the greenhouse effect or otherwise act to enhance the heat-
absorbing capacity of the Earth system. Purely reflective atmospheric aerosol, snow-covered
land surfaces, and dense sea ice provide a negative forcing, while black carbon-containing
atmospheric aerosol, GHGs, and dark ocean surfaces provide a positive forcing of the climate
system.
The IPCC reports included estimated values for forcing by the individual, radiatively active
gas and particle-phase components of the atmosphere. These estimates are derived primarily
through expert judgment, incorporating the results of peer-reviewed modeling studies.
Uncertainty ranges are assigned that reflect the range of modeled values reported in these
studies. According to the available research on climate forcing by aerosol, the panel has
provided estimates for sulfate, organic, black carbon, biomass burning, and mineral dust aerosol.
The current estimate of forcing due to long-lived, well-mixed GHGs accumulated in the
atmosphere from the preindustrial era (-1750) through the year 2000 is +2.4 W nT2 (IPCC,
4-211
-------
200la). In contrast, forcing due to a sulfate aerosol-related increase in planetary albedo has been
assigned a value of -0.4 W nT2. Biomass burning and fossil-fuel-related organic aerosol also
increase the Earth's reflectivity and are estimated to contribute a -0.2 W nT2 and -0.1 W nT2
forcing, respectively. Fossil-fuel black carbon is expected to warm the atmosphere, resulting in
an estimated +0.2 W nT2 forcing. No estimate for forcing by nitrate aerosol has been proposed
due to the wide discrepancies in current global modeling results and the difficulties associated
with obtaining accurate ambient samples of nitrate concentrations and size distributions.
Likewise, no specific estimate has been offered for forcing by mineral dust aerosol introduced
into the atmosphere due to human activities beyond the assignment of a tentative range of
-0.6 to +0.4 W nT2. The estimated forcing and associated uncertainty for each aerosol type is
shown in relation to forcing estimates for the known GHGs along with an indication of the level
of confidence in each of these estimates in Figure 4-42.
The relationship between perturbations to the Earth's radiative balance and climate is
complicated by various feedbacks within the climate system. An example would be the positive
feedback associated with melting sea ice. As sea ice melts with increasing surface temperatures,
the dark ocean surface is revealed which absorbs, rather than reflects, solar radiation. Such a
feedback increases the rate of surface warming. The role of feedbacks in determining the
sensitivity of climate to changes in radiative forcing is described in detail by the IPCC in its third
assessment report (IPCC, 200la).
One possible feedback worth consideration in the discussion of the role of aerosols in
climate may result from the possible sensitivity of aerosol number and mass to atmospheric
temperature (Hemming and Seinfeld, 2001). Increasing atmospheric temperatures may result in
a reduction of aerosol as semivolatile organic and inorganic aerosol constituents evaporate,
leading to a change in aerosol forcing. As described in Chapter 2 of this document, and further
discussed below, ambient aerosols are known to contain complex chemical mixtures of both
scattering and absorbing materials. The feedback that may result from this phenomenon will
depend upon whether the aerosols become more absorbing or more reflective with the loss of
semivolatile material. Research is currently underway to evaluate both the role of temperature in
determining aerosol mass and in defining the link between air quality and climate, but no
literature yet exists to support an assessment of these effects. The following discussion,
4-212
-------
1Z^
_0>
E
1_ D>
=5 E
rT i_
C/5 •§
CD ^ 1-
Q.
to
s
> 0~
D5
C
"o CT
£ 1 -1-
(D O
"03
TS _2-
03 ^
01
_L
1
c
Halocarbons
N,O Aerosols
CH4 r Black
„_ carbon from
1
^U2 fossil
Tropospheric ^ | Mineral Aviation-induced
ozone burning Dust , ^ , Solar
i__ T
r— I— ,
±
ozone 1 carbon Biomass
Sulphate f^°"?. burning
fuel
burning
High Medium Medium Low Very Very
Low Low
Contrails Cirrus r-j-i
u as u -r- IE
III
Land-
Aerosol use
indirect (albedo)
effect only
1
Very Very Very Very Very Very
Low Low Low Low Low Low
Level of Scientific Understanding
Figure 4-42. Estimated global mean radiative forcing exerted by gas and various particle
phase species for the year 2000, relative to 1750.
Source: IPCC(2001a).
therefore, will focus solely on the relative forcing properties of aerosols, for which a body of
scientific research is available for consideration.
The physical and chemical properties of atmospheric aerosols, and their regional
distribution and temporal nature, all play a role in determining the degree to which they force
climate. These details are described, below, followed by a description of the "indirect" effect of
aerosols due to changes in cloud properties, and then by brief discussions of the sources of
uncertainty in determining aerosol-related climate forcing and of the link to human health and
the environment.
4-213
-------
The Physics of Particulate Matter and Its Climate Effects
Four wavelength-dependent physical properties of ambient aerosols are needed to calculate
their optical depth and, thus, their radiative properties: the mass light-scattering efficiency (asp),
the functional dependence of light-scattering on relative humidity f(RH), the single-scattering
albedo (co0), and the scattering asymmetry parameter (g); (Charlson et al., 1992; Penner et al.,
1994). Direct forcing by aerosols is especially sensitive to single-scattering albedo. Depending
on the color of the underlying surface (i.e. the surface albedo) small changes in co0 can change
the sign of the calculated aerosol forcing (Hansen et al., 1997).
The wavelength-dependent phase function and scattering and absorption coefficients are
calculated using Mie scattering theory. These values are then used to calculate mass light-
scattering efficiency, single-scattering albedo, and the asymmetry factor. Mie calculations
require the ratio of particle size versus wavelength and the complex refractive index of the
particle — a composition-dependent property (Salby, 1996). The relationship between light-
scattering and relative humidity likewise depends upon composition, as water absorption
depends upon the presence of hygroscopic material within the particle. Therefore, according to
the current understanding of aerosol optics, good information about composition and size
distribution is needed to successfully predict aerosol-related forcing.
The Chemistry of Particulate Matter and Its Climate Effects
Although forcing estimates are reported for specific aerosol classes (i.e., sulfate, black
carbon, dust, etc.) it is understood that shortly after emission, primary aerosols undergo chemical
transformation in the atmosphere. These transformations occur through the partitioning of gas
phase compounds, by coagulation with another aerosol, or through a combination of both
processes (Seinfeld and Pandis, 1998). While the radiative properties of nonabsorbing aerosol,
such as sulfates or nitrates, are primarily sensitive to particle size, the radiative properties of
aerosols containing absorbing constituents (i.e., black carbon and mineral aerosols) are also
sensitive to chemical composition and mixing state. Studies of refractive index changes as a
function of composition have shown that the type of mixing present both in the population of
aerosols (i.e., "internal" versus "external" mixtures) and the extent of mixing within individual
aerosols influence aerosol optical properties when absorbing material is present (Fuller et al.,
1999). Modeled estimates for radiative forcing by black carbon-containing aerosols do, in fact,
4-214
-------
range widely depending on whether the aerosol population is assumed to be internally or
externally mixed and whether the absorbing carbon is uniformly distributed in the particle or
whether it exists as a core surrounded by nonabsorbing material. The IPCC Third Assessment
Report (200la) provides a tabulation of studies and their forcing estimates for black carbon as it
exists in different mixing states. For example, Hay wood and Shine (1995) calculated that
externally mixed fossil-fuel black carbon forces climate by +0.2 W nT2. Jacobson (2000)
assumed that if black carbon exists as a solid core contained within an otherwise nonabsorbing
droplet, the global forcing will be +0.54 W nT2.
Chemically and physically detailed models at high spatial resolution are required to
accurately represent the chemical transformation and size evolution of aerosol within the
ambient atmosphere. Such models, however, are impractical for global scale climate modeling.
Parameterizations are formulated to represent processes occurring at spatial and temporal scales
that are too fine for climate models (i.e., "sub-grid" processes). Several important fundamental
chemical and physical processes, however, are not yet well-enough defined to assure reasonable
parameterization in large scale climate models. Prediction of the organic and black carbon
content of ambient aerosols and their associated radiative properties, as described above, and the
aerosol-induced changes in cloud properties, remain especially weak.
Modeling of the effect of black carbon aerosols on climate, to date, has been done on the
basis of limited and poor quality data regarding total global emissions, aerosol composition, and
the mixing state of ambient aerosols. Appendix 2B.2 of Chapter 2 in this document describes
the many problems associated with the most commonly used organic carbon/elemental carbon
measurement method (thermal-optical reflectance and transmittance). For example, elemental
carbon comprises a relatively small fraction of total carbon in biomass burning-related aerosol
which is, nevertheless, an important source of black carbon in the atmosphere due to the
extensive use of biofuels in developing countries and to the prevalence of natural and
agriculture-related open biomass burning (Ludwig et al., 2003). The concentrations of elemental
carbon in these aerosols are close to the detection limits of thermal-optical measurement
systems, adding a high degree of uncertainty to the reported values. Thermal-optical
measurements also do not provide information about the specific chemical composition of
carbonaceous aerosol that is needed to calculate refractive indices. Furthermore, the technique is
4-215
-------
based on filter collection, which only provides an average concentration of organic and
elemental carbon and no information about mixing states.
The magnitude of the errors introduced into estimates of black carbon-related radiative
forcing due to the lack of detailed information about optical properties of ambient black carbon
aerosol cannot yet be estimated. The forcing estimate of +0.2 W nT2 provided by the IPCC is
based upon a summary of the available modeling estimates, some of which are founded on
simplistic assumptions about the composition and mixing state of ambient BC aerosol.
New methods, such as photoacoustic spectroscopy, hold promise for detecting quantities of
light absorbing materials in ambient aerosol streams (Arnott et al., 1999; Moosmuller et al.,
2001). The photoacoustic method is capable of identifying the absolute fraction of a given
ambient aerosol sample that absorbs radiation at a given wavelength at high temporal resolution.
The method does not alter the composition of the aerosol sample through heating and oxidation,
thus introduces no chemical artifacts. Present photoacoustic instruments detect absorption at
wavelengths of interest for visibility studies, but the measurement principle can be adapted to
climate studies through the selection of climate-forcing relevant wavelengths. Specific chemical
speciation at the organic and inorganic compound level coupled with photoacoustic
measurements of the light absorbing properties of ambient aerosol would provide the best
available test of atmospheric chemistry models attempting to simulate the evolution in aerosol
radiative properties between emission and atmospheric removal.
The Regional and Temporal Dependence of Aerosol Forcing
In addition to information about the composition and size distributions of ambient aerosol,
details regarding atmospheric concentrations on a spatial and temporal basis are needed to
estimate climate forcing. Unlike the long-lived GHGs, aerosols have lifetimes averaging only
days to weeks or less, leading to inhomogeneity in regional and global-scale concentrations.
Industrial processes and other human activities that produce air pollution aerosols also vary on a
seasonal, monthly, weekly, or diurnal basis, introducing other complications into forcing
estimates (IPCC, 200la). Peak aerosol concentrations, along with the greatest temporal
variation, exist near emissions sources.
Deviations from global mean forcing estimates can be very large on the regional scale.
For instance, Tegen et al. (1996) found that local radiative forcing exerted by dust raised from
4-216
-------
disturbed lands ranges from -2.1 W nT2 to +5.5 W nT2 over desert areas and their adjacent seas.
The largest regional values of radiative forcing caused by anthropogenic sulfate are about
-3 W nT2 in the eastern United States, south central Europe, and eastern China (Kiehl and
Briegleb, 1993). These regional maxima in aerosol forcing are at least a factor of 10 greater than
their global mean values shown in Figure 4-41. By comparison, regional maxima in forcing by
the well-mixed GHGs are only about 50% greater than their global mean value (Kiehl and
Briegleb, 1993). Thus, the estimates of local radiative forcing by particles also are large enough
to completely cancel the effects of GHGs in many regions and to cause a number of changes in
the dynamic structure of the atmosphere that still need to be evaluated. A number of
anthropogenic pollutants whose distributions are highly variable are also effective greenhouse
absorbers. These gases include O3 and, possibly HNO3, C2H4, NH3, and SO2, all of which are
not commonly considered in radiative forcing calculations (Wang et al., 1976). High-ozone
values are found downwind of urban areas and areas where there is biomass burning. However,
Van Borland et al. (1997) found that there may not be much cancellation between the radiative
effects for ozone and for sulfate because both species have different seasonal cycles and show
significant differences in their spatial distribution.
"Indirect" Effects of Particulate Matter on Climate
Aerosols directly affect climate by scattering and absorbing solar and terrestrial radiation.
Depending on chemical composition, they can also nucleate new cloud droplets. For a given
total liquid water content (LWC), increasing cloud droplet number means smaller droplets that
scatter solar radiation more effectively, reduce the amount of precipitation from the cloud, and,
consequently, increase cloud lifetime. Observational evidence exists for this impact of
hygroscopic aerosol on both warm and ice cloud properties. Increased albedo and increased
cloud lifetime effects are treated separately by the climate modeling community. Particle-
induced increases in cloud albedo are referred to as the "first" indirect effect, and changes to the
cloud lifetime due to reduced precipitation are referred to as the "second" indirect effect. The
highly uncertain estimated forcing due to the effects of particles on cloud albedo is given as a
range from 0 to +2.0 W nT2 (IPCC, 2001a). A further effect, referred to as the "semi-indirect
effect" is the reduction in cloud reflectivity due to the inclusion of black carbon-containing
aerosol within the cloud drops themselves or as interstitial aerosol (Hansen et al., 2000).
4-217
-------
Sulfate aerosols, especially those > 50 nm in diameter, are believed to alter clouds to the
largest extent due to their efficient nucleation of cloud drops and ice crystals (Twomey, 1974).
Organic aerosols that contain highly oxidized carbon compounds may be similarly efficient in
nucleating cloud droplets (Novakov and Penner, 1993). Both satellite and in-situ aircraft
observations reinforce the hypothesis that pollutant aerosols increase cloud reflectivity and
lifetime (Ramanathan et al., 2001).
An important consequence of this property of aerosols on regional climate includes
suppression of rain over polluted regions. Satellite observations show that precipitation occurs
only outside of pollution tracks, while clouds within pollution tracks show a reduction in
effective cloud drop radius to below the precipitation threshold (Rosenfeld, 1999). Desert dust
also appears to alter the microphysical properties of clouds, suppressing precipitation from warm
clouds while nucleating ice crystals in cold clouds.
While climate models are not yet equipped for modeling the effect of aerosols on regional
and global hydrological cycles, it has been proposed that aerosols will reduce precipitation
efficiency over land where anthropogenic activities provide a major source of cloud
condensation nuclei. Several studies using coupled ocean-atmosphere general circulation
models support the possibility of a "spin down" effect upon hydrological cycles due only to a
reduction in surface radiation receipts from sulfate aerosol scattering. When indirect effects are
included, the reduction in precipitation rates from clouds is large enough to reverse the effect of
GHG-related forcing (Ramanathan et al., 2001).
Sources of Uncertainty in Aerosol Forcing Estimates
Uncertainties in calculating the direct effect of airborne particles arise from a lack of
knowledge concerning their vertical and horizontal variability, their size distribution, chemical
composition, and the distribution of components within individual particles. For instance,
gas-phase sulfur species may be oxidized to form a layer of sulfate around existing particles in
continental environments or they may be incorporated in sea-salt particles (e.g., Li-Jones and
Prospero, 1998). In either case, the radiative effects of a given mass of the sulfate will be much
lower than if pure sulfate particles were formed. It must also be stressed that the overall
radiative effect of particles at a given location is not simply determined by the sum of effects
4-218
-------
caused by individual classes of particles because of radiative interactions between particles with
different radiative characteristics and with gases.
That calculations of the indirect effects of particles on climate are subject to much larger
uncertainties than are calculations of their direct effects reflects uncertainties in a large number
of chemical and microphysical processes that determine aerosol chemistry, size distribution, and
the number of droplets within a cloud. A complete assessment of the radiative effects of PM
will require computationally intensive calculations that incorporate the spatial and temporal
behavior of particles of varying composition that have been emitted from, or formed from
precursors emitted from, different sources. Refining the values of model-input parameters (such
as improving emissions estimates) may be as important as improving the models per se in
calculations of direct radiative forcing (Pan et al., 1997) and indirect radiative forcing (Pan et al.,
1998). However, the uncertainties associated with the calculation of radiative effects of particles
will likely remain much larger than those associated with well-mixed GHGs.
Aerosol-Related Climate Effects, Human Health, and the Environment
The specific impacts on human health and the environment due to aerosol effects on the
climate system can not be calculated with confidence given the present difficulty in accurately
modeling an aerosol's physical, chemical, and temporal properties and its regionally dependent
atmospheric concentration levels, combined with the difficulties of projecting location-specific
increases or decreases in anthropogenic emissions of atmospheric particles (or their precursors).
However, substantial qualitative information, from observation and modeling, indicates that
aerosol forces climate both positively and negatively, globally and regionally, and may be
negatively impacting hydrological cycles on a regional scale. Global and other regional scale
impacts are suspected on the basis of current, though uncertain, modeling studies suggesting that
climate change in general can have positive and negative effects on human health, human
welfare, and the environment.
4-219
-------
4.5.2 Atmospheric Particulate Matter Effects on the Transmission of Solar
Ultraviolet Radiation Transmission: Impacts on Human Health and
the Environment
4.5.2.1 Potential Effects of Increased Ultraviolet Radiation Transmission
The transmission of solar UV-B radiation through the Earth's atmosphere is controlled by
ozone, clouds, and particles. The depletion of stratospheric ozone caused by the release of
anthropogenically produced chlorine (Cl)- and bromine (Br)-containing compounds has resulted
in heightened concern about potentially serious increases in the amount of solar UV-B radiation
(SUVB) reaching the Earth's surface. SUVB is also responsible for initiating the production of
hydroxyl radicals ('OH) that oxidize a wide variety of volatile organic compounds, some of
which can deplete stratospheric ozone (e.g., CH3C1, CH3Br), absorb terrestrial infrared radiation
(e.g., CH4), and contribute to photochemical smog formation (e.g., C2H4, C5H8).
Increased penetration of SUVB to the Earth's surface as the result of stratospheric ozone
depletion continues to be of much concern because of projections of consequent increased
surface-level SUVB exposure and associated potential negative impacts on human health, plant
and animal biota, and man-made materials. Several summary overviews (Kripke, 1989; Grant,
1989; Kodama and Lee, 1994; Van der Leun et al., 1995, 1998) of salient points related to
stratospheric ozone depletion and bases for concern provide a concise introduction to the subject.
Only a brief summary will be given here. Stratophospheric ozone depletion results from
(a) anthropogenic emissions of certain trace gases having long atmospheric residence times, e.g.,
chlorofluorocarbons (CFCs), carbon tetrachloride (CC14), and Halon 1211 (CF2C1 Br) and 1301
(CF3Br) — which have atmospheric residence times of 75 to 100 years, 50 years, 25 years, and
110 years, respectively; (b) their tropospheric accumulation and gradual transport, over decades,
up to the stratosphere, where (c) they photolyze to release Cl and Br that catalyze ozone
destruction; leading to (d) stratospheric ozone depletion. Such ozone depletion is most marked
over Antarctica during spring in the southern hemisphere, to a less marked but still significant
extent over the Arctic polar region during late winter and spring in the northern hemisphere, and
to a lesser extent, over midlatitude regions during any season. Given the long time frame
involved in the transport of such gases to the stratosphere and their long residence times there,
any effects already seen on stratospheric ozone are likely to have been caused by the
atmospheric loadings of trace gases from anthropogenic emissions over the past few decades.
Ozone-depleting gases already in the atmosphere will continue to affect stratospheric ozone
4-220
-------
concentrations well into the 21st century. Shorter-lived gases, such as CH3Br, also exert
significant ozone depletion effects.
The main types of deleterious effects hypothesized as likely to result from stratospheric
ozone depletion and consequent increased SUVB penetration through the Earth's atmosphere
include the following:
(1) Direct Human Health Effects, such as skin damage (sunburn), leading to more rapid
aging and increased incidence of skin cancer; ocular effects (retinal damage and
increased cataract formation, possibly leading to blindness); and suppression of
some immune system components (contributing to skin cancer induction and spread
to nonirradiated skin areas, as well as possibly increasing susceptibility to certain
infectious diseases).
(2) Agricultural/Ecological Effects, mediated largely through altered biogeochemical
cycling resulting in consequent damaging impacts on terrestrial plants (leading to
possible reduced yields of rice, other food crops, and commercially important trees, as
well as to biodiversity shifts in natural terrestrial ecosystems); and deleterious effects
on aquatic life (including reduced ocean zooplankton and phytoplankton as important
base components of marine food-chains supporting the existence of commercially
important, edible fish and other seafood, as well as to other aquatic ecosystem shifts).
(3 ) Indirect Human Health and Ecological Effects, mediated through increased
tropospheric ozone formation (and consequent exacerbation of surface-level, ozone-
related health and ecological impacts) and alterations in the concentrations of other
important trace species, most notably the hydroxyl radical and acidic aerosols.
(4) Other Types of Effects, such as faster rates of polymer weathering because of
increased UV-B radiation and other effects on man-made commercial materials
and cultural artifacts, secondary to climate change or exacerbation of air pollution
problems.
Extensive qualitative and quantitative characterizations of stratospheric ozone depletion
processes and projections of their likely potential impacts on human health and the environment
have been the subjects of periodic (1988, 1989, 1991, 1994, 1998) international assessments
carried out under WMO and UNEP auspices since the 1987 signing of the Montreal Protocol on
Substances that Deplete the Ozone Layer. For more detailed up-to-date information, the reader
is referred to recent international assessments of (a) processes contributing to stratospheric ozone
depletion and the status of progress towards ameliorating the problem (WMO, 1999) and
(b) revised qualitative and quantitative projections of potential consequent human health and
environmental effects (UNEP, 1998, 2000).
4-221
-------
Of considerable importance is the growing recognition, as reflected in these newer
assessments, of the impacts of enhanced solar radiation on biogeochemical cycles (see, for
example, Zepp et al. [1998]). As noted in that paper, the effects of UV-B radiation (both in
magnitude and direction) on trace gas (e.g., CO) emissions and mineral nutrient cycling are
species specific and can affect a variety of processes. These include, for example, changes in the
chemical composition of living plant tissue, photodegradation of dead plant matter (e.g., ground
litter), release of CO from vegetation previously charred by fire, changes in microbial
decomposer communities, and effects on nitrogen-fixing microorganisms and plants. In
addition, changes in the amount and composition of organic matter, caused by enhanced UV-B
penetration, affect the transmission of solar ultraviolet and visible radiation through the water
column. These changes in light quality broadly impact the effects of UV-B on aquatic
biogeochemical cycles. Enhanced UV-B levels cause both positive and negative effects on
microbial activities in aquatic ecosystems that can affect nutrient cycling and the uptake or
release of GHGs. Thus, there are emerging complex issues regarding interactions and feedbacks
between climate change and changes in terrestrial and marine biogeochemical cycles because of
increased UV-B penetration to the Earth's surface.
In contrast to the above types of negative impacts projected as likely to be associated with
increased UV-B penetration to Earth's surface, some research results suggest possible beneficial
effects of increased UV-B radiation. For example, a number of U.S. and international studies
have focused on the protective effects of UV-B radiation with regard to the incidence of cancers
other than skin cancer. One of the first of these studies investigated potential relationships
between sunlight, vitamin D, and colon cancer (Garland and Garland, 1980). More recent
studies continue to provide suggestive evidence that UV-B radiation may be protective against
several types of cancer and some other diseases. For example, Grant (2002) conducted a number
of ecologic-type epidemiologic studies, which suggest that UV-B radiation, acting through the
production of vitamin D, is a risk-reduction factor for mortality due to several types of cancer,
including cancer of the breast, colon, ovary, and prostate, as well as non-Hodgkin lymphoma.
Other related studies that provide evidence for protective effects of UV-B radiation include
Gorham et al. (1989, 1990); Garland et al. (1990); Hanchette and Schwartz (1992); Ainsleigh
(1993); Lefkowitz and Garland (1994); Hartge et al. (1996); and Freedman et al. (1997).
4-222
-------
As noted in the above detailed international assessments, since the signing of the Montreal
Protocol, much progress has been made in reducing emissions of ozone depleting gases, leading
to estimates that the maximum extent of stratospheric ozone depletion may have leveled off
during recent years. This is expected to be followed by a gradual lessening of the problem and
its impacts during the next half-century. However, the assessments also note that the modeled
projections are subject to considerable uncertainty (see, for example, UNEP [2000]). Varying
potential roles of atmospheric particles, discussed below, are among salient factors complicating
predictive modeling efforts.
4.5.2.2 Airborne Particle Effects on Atmospheric Transmission of Solar
Ultraviolet Radiation
A given amount of ozone in the lower troposphere has been shown to absorb more solar
radiation than an equal amount of ozone in the stratosphere, because of the increase in its
effective optical path produced by Rayleigh scattering in the lower atmosphere (Briihl and
Crutzen, 1988). The effects of particles are more complex. The impact of particles on the
SUVB flux throughout the boundary layer are highly sensitive to the altitude of the particles and
to their single-scattering albedo. Even the sign of the effect can reverse as the composition of
the particle mix changes from scattering to absorbing types (e.g., from sulfate to elemental
carbon or PAHs) (Dickerson et al., 1997). In addition, scattering by particles also may increase
the effective optical path of absorbing molecules, such as ozone, in the lower atmosphere.
The effects of particles present in the lower troposphere on the transmission of SUVB have
been examined both by field measurements and by radiative transfer model calculations. The
presence of particles in urban areas modifies the spectral distribution of solar irradiance at the
surface. Shorter wavelength (i.e., ultraviolet) radiation is attenuated more than visible radiation
(e.g., Peterson et al., 1978; Jacobson, 1999). Wenny et al. (1998) also found greater attenuation
of SUVB than SUVA (315 to 400 nm). However, this effect depends on the nature of the
specific particles involved and, therefore, is expected to depend strongly on location. Lorente
et al. (1994) observed an attenuation of SUVB ranging from 14 to 37%, for solar zenith angles
ranging from about 30° to about 60°, in the total (direct and diffuse) SUVB reaching the surface
in Barcelona during cloudless conditions on very polluted days (aerosol scattering optical depth
at 500 nm, 0.46 £ T500nm < 1.15) compared to days on which the turbidity of urban air was
similar to that for rural air (T500nm < 0.23).
4-223
-------
Particle concentrations that can account for these observations can be estimated roughly by
combining Koschmieder's relation for expressing visual range in terms of extinction coefficient
with one for expressing the mass of PM25 particles in terms of visual range (Stevens et al.,
1984). By assuming a scale height (i.e., the height at which the concentration of a substance
falls off to 1/e of its value at the surface) of 1 km for PM2 5, an upper limit of 30 |ig/m3 can be
derived for the clear case and between 60 and 150 |ig/m3 for the polluted case. Estupifian et al.
(1996) found that summertime haze under clear sky conditions attenuates SUVB between 5 and
23% for a solar zenith angle of 34° compared to a clear-sky day in autumn. Minis (1996)
measured a decrease in SUVB by about 80% downwind of major biomass burning areas in
Amazonia in 1995. This decrease in transmission corresponded to optical depths at 340 nm
ranging from 3 to 4. Justus and Murphey (1994) found that SUVB reaching the surface
decreased by about 10% because of changes in aerosol loading in Atlanta, GA from 1980 to
1984. In addition, higher particle levels in Germany (48 °N) may be responsible for the greater
attenuation of SUVB than seen in New Zealand (Seckmeyer and McKenzie, 1992).
In a study of the effects of nonurban haze on SUVB transmission, Wenny et al. (1998)
derived a very simple regression relation between the measured aerosol optical depth at 312 nm,
ln(SUVB solar noon) = -0.1422T312nm - 0.138, R2 = 0.90, (4-13)
and the transmission of SUVB to the surface at solar noon. In principle, values of T312nm could
be found from knowledge of the aerosol optical properties and visual range values. Wenny et al.
(1998) also found that absorption by particles accounted for 7 to 25% of the total (scattering +
absorption) extinction. Relations such as the one above are strongly dependent on local
conditions and should not be used in other areas without knowledge of the differences in aerosol
properties. Although all of the above studies reinforce the idea that particles play a major role in
modulating the attenuation of SUVB, none included measurements of ambient PM
concentrations, so direct relations between PM levels and SUVB transmission could not be
determined.
Vuilleumier et al. (2001) concluded that variations in aerosol scattering and absorption
were responsible for 97% of the variability in the optical depth measured at seven wavelengths
from 300 to 360 nm at Riverside, CA from 1 July to 1 November 1997. Similar measurements
4-224
-------
made at nearby Mt. Wilson, located above the main surface haze layer, showed that 80% of the
variations in optical depth were still driven by variations in aerosol scattering and absorption.
The remainder of the variability in optical depth was attributed mainly to variability in ozone
under clear-sky conditions. However, these results cannot be extrapolated to other locations,
because these effects are coupled and nonlinear and, thus, are not straightforward. They depend
on the concentrations of these species and on the physical and chemical characteristics of the
particles. Hence, any quantitative statements regarding the relative importance of particles and
ozone will be location-specific.
Liu et al. (1991) roughly estimated the overall effects on atmospheric transmission of
SUVB of increases of anthropogenic airborne particles that have occurred since the beginning of
the industrial revolution. Based on (a) estimates of the reduction in visibility from about 95 km
to about 20 km over nonurban areas in the eastern United States and in Europe, (b) calculations
of optical properties of airborne particles found in rural areas to extrapolate the increase in
extinction at 550 to 310 nm, and (c) radiative transfer model calculations, Liu et al. (1991)
concluded that the amount of SUVB reaching Earth's surface likely has decreased from 5 to 18%
since the beginning of the industrial revolution. This was attributed mainly to scattering of
SUVB back to space by sulfate-containing particles. Radiative transfer model calculations have
not been done for urban particles.
Although aerosols are expected to decrease the flux of SUVB reaching the surface,
scattering by particles is expected to result in an increase in the actinic flux within and above the
aerosol layer. However, when the particles significantly absorb SUVB, a decrease in the actinic
flux is expected. Actinic flux is the radiant energy integrated over all directions at a given
wavelength incident on a point in the atmosphere and is the quantity needed to calculate rates of
photolytic reactions in the atmosphere. Blackburn et al. (1992) measured attenuation of the
photolysis rate of ozone and found that aerosol optical depths near unity at 500 nm reduced
ozone photolysis rate by as much as a factor of two. Dickerson et al. (1997) showed that the
photolysis rate for NO2, a key parameter for calculating the overall intensity of photochemical
activity, could be increased within and above a scattering aerosol layer extending from the
surface, although it would be decreased at the surface. This effect is qualitatively similar to what
is seen in clouds, where photolysis rates are increased in the upper layers of a cloud and above
the cloud (Madronich, 1987). For a simulation of an ozone episode that occurred during July
4-225
-------
1995 in the Mid-Atlantic region, Dickerson et al. (1997) calculated ozone increases of up to
20 ppb compared to cases that did not include the radiative effects of particles in urban airshed
model (UAM-IV) simulations. In contrast, Jacobson (1998) found that particles may have
caused a 5 to 8% decrease in O3 levels during the Southern California Air Quality Study in 1987.
Absorption by organic compounds and nitrated inorganic compounds was hypothesized to
account for the reductions in UV radiation intensity.
The photolysis of ozone in the Hartley bands also leads to production of electronically
excited oxygen atoms, O(JD), that then react with water vapor to form hydroxyl radicals. Thus,
enhanced photochemical production of ozone is accompanied by the scavenging of species
involved in greenhouse warming and stratospheric depletion. However, these effects may be
neutralized, or even reversed, by the presence of absorbing material in the particles. Any
evaluation of the effects of particles on photochemical activity, therefore, will depend on the
composition of the particles and will also be location-specific.
Further complicating any straightforward evaluation of UV-B penetration to specific areas
of the Earth's surface are the influences of clouds, as discussed by Erlick et al. (1998), Frederick
et al. (1998), and Soulen and Fredrick (1999). The transmission of solar UV and visible
radiation is highly sensitive to cloud type, cloud amount, and the extent of their external or
internal mixing with cloud droplets. Even in situations of very low atmospheric PM (e.g., over
Antarctica), interannual variations in cloudiness over specific areas can be as important as ozone
levels in determining ultraviolet surface irradiation, with net impacts varying monthly to
seasonally (Soulen and Fredrick, 1999). Evaluations of the effects of changes in the
transmission of solar UV-B radiation to the surface have been performed usually for cloud-free
or constant cloudiness conditions.
Given the above considerations, the quantification of projected effects of variations in
atmospheric PM on human health or on the environment related to the effects of particles on the
transmission of solar UV-B radiation would require location-specific evaluations that take into
account (a) composition, concentration, and internal structure of the particles; (b) temporal
variations in atmospheric mixing height and depths of layers containing the particles; and
(c) the abundance of ozone and other absorbers within the planetary boundary layer and the free
troposphere. The outcome of such modeling effects would likely vary from location to location
in terms of increased or decreased surface level UV-B exposures because of location-specific
4-226
-------
changes in atmospheric PM concentrations or composition. For example, to the extent that any
location-specific scattering by airborne PM were to affect the directional characteristics of
ultraviolet radiation at ground level, and thereby enhance radiation incident from low angles
(Dickerson et al., 1997), the biological effectiveness (whether deleterious or beneficial) of
resulting ground-level UV-B exposures could be enhanced. Airborne PM also can reduce the
ground-level ratio of photorepairing radiation (UV-A and short-wavelength visible) to damaging
UV-B radiation. Lastly, PM deposition is a major source of PAHs in certain freshwater lakes
and coastal areas; and the adverse effects of solar UV are enhanced by the uptake of PAHs by
aquatic organisms. Thus, although airborne PM may, in general, tend to reduce ground-level
UV-B, its net effect in some locations may be to increase UV damage to certain aquatic and
terrestrial organisms, as discussed by Cullen andNeale (1997).
4.6 SUMMARY AND KEY CONCLUSIONS
4.6.1 Particulate Matter Effects on Vegetation and Ecosystems
The first section of this chapter assessed and characterized the overall ecological integrity
and indicated the status of ecosystems within the United States affected by the deposition of the
anthropogenic stressors associated with PM. There are six essential environmental attributes
(EEAs) — landscape condition, biotic condition, and chemical/physical characteristics, and
ecological processes, hydrology/geomorphology, and natural disturbance regimes that can be
used to provide a hierarchical framework for characterizing ecosystem status. The first three can
be separated into "patterns" and the last three into "processes." The ecological processes create
and maintain the ecosystem elements in patterns. The patterns, in turn, affect how the ecosystem
processes are expressed. Patterns at the higher level of biological organization emerge from the
interactions and selection processes at localized levels. Changes in patterns or processes result
in changes in the status and functioning of an ecosystem. The relationships among the EEAs are
complex because all are interrelated (i.e., changes in one EEA may affect, directly or indirectly,
every other EEA).
Human existence on Earth depends on the life-support services provided by the interaction
of the different EEAs. Ecosystem processes and patterns provide the functions that maintain
clean water, clean air, a vegetated earth, and a balance of organisms that enable humans to
4-227
-------
survive. The benefits they impart include the absorption and breakdown of pollutants, cycling of
nutrients, binding of soil, degradation of organic waste, maintenance of a balance of gases in the
air, regulation of radiation balance, climate, and fixation of solar energy. Concern has arisen in
recent years regarding biodiversity and the integrity of ecosystems. Because the activities of few
ecosystems on Earth today are not influenced by humans, understanding the changes in
biodiversity and nutrient cycling resulting from anthropogenic PM deposition are of great
importance.
The PM whose effects on vegetation and ecosystems is discussed in this section is not a
single pollutant but represents a heterogeneous mixture of particles differing in origin, size, and
chemical constituents. The effects of exposure to a given mass concentration of PM of particular
size may, depending on the particular mix of deposited particles, lead to widely differing
phytotoxic responses.
The deposition of PM onto vegetation and soil, depending on its chemical composition
(acid/base, trace metal, or nutrients, e.g., nitrates or sulfates), can produce direct or indirect
responses within an ecosystem. The ecosystem response to pollutant deposition is a direct
function of the level of sensitivity of the ecosystem and its ability to ameliorate resulting change.
Changes in ecosystem structural patterns and the functioning of ecological processes, must be
scaled in both time and space and propagated to the more complex levels of community
interaction to produce observable ecosystem changes.
The stressors of greatest environmental significance are particulate nitrates and sulfates
whose indirect effects occur primarily via their deposition onto the soil. Upon entering the soil
environment, they can alter the ecological processes of energy flow and nutrient cycling, inhibit
nutrient uptake, change ecosystem structure, and affect ecosystem biodiversity. The soil
environment is one of the most dynamic sites of biological interaction in nature. It is inhabited
by microbial communities of bacteria, fungi, and actinomycetes. Bacteria are essential
participants in the nitrogen and sulfur cycles that make these elements available for plant uptake.
Fungi in association with plant roots form mycorrhizae, a symbiotic relationship that is integral
in mediating plant uptake of mineral nutrients. Changes in the soil environment that influence
the role of the bacteria in nutrient cycling and fungi in making minerals available for plant
utilization, determine plant and ultimately ecosystem response.
4-228
-------
The effects on the growth of plants resulting from the deposition of nitrates and sulfates
and the acidifying effect of the associated H+ ion in wet and dry deposition are the most
important environmentally. Nitrogen is of overriding importance in plant metabolism and, to a
large extent, governs the utilization of phosphorus, potassium, and other nutrients. Typically,
the availability of nitrogen via the nitrogen cycle controls net primary productivity, and possibly,
the decomposition rate of plant litter. Plants usually obtain nitrogen directly from the soil by
absorbing NH4+ or NO3 through their roots, or it is formed by symbiotic organisms in their
roots. Plants vary in their ability to absorb ammonium and nitrate from the soil.
Although nitrogen as molecular nitrogen (N2) is the most abundant element in the
atmosphere, it only becomes available to support the growth of plants after it is converted into
reactive forms. In nature, nitrogen may be divided into two groups: reactive (Nr) and
nonreactive (N2). Reactive Nr includes the inorganic reduced forms of nitrogen (e.g., ammonia
[NH3] and ammonium [NH4+]), inorganic oxidized forms (e.g., nitrogen oxides [NOX], nitric acid
[HNO3], nitrous oxide [N2O], and nitrate [NO3~]), and organic compounds (e.g., urea, amine,
proteins, and nucleic acids).
Reactive nitrogen is now accumulating in the environment on all spatial scales — local,
regional, and global. The three main causes of the increase in global Nr is the result of the
(1) widespread cultivation of legumes, rice, and other crops that promote conversion of N2 to
organic nitrogen through biological nitrogen fixation; (2) combustion of fossil fuels, which
converts both atmospheric N2 and fossil nitrogen to reactive NOX; and (3) the Haber-Bosch
process, which converts nonreactive NH3 to sustain food production and some industrial
activities. The major changes in the nitrogen cycle due to the cited causes can be both beneficial
and detrimental to the health and welfare of humans and ecosystems.
Reactive nitrogen can be widely dispersed and accumulate in the environment when the
rates of its formation exceed the rates of removal via denitrification. Reactive nitrogen creation
and accumulation is projected to increase as per capita use of resources by human populations
increases. The cascade of environmental effects of increases in Nr include the following:
(1) production of tropospheric ozone and aerosols that induce human health problems;
(2) increases in the productivity in forests and grasslands followed by decreases wherever
deposition increases significantly and exceeds critical loads; Nr additions probably also decrease
biodiversity in many natural habitats; (3) acidification and loss of biodiversity in lakes and
4-229
-------
streams in many regions of the world in association with sulfur; (4) eutrophication (now
considered the biggest pollution problem in coastal waters), hypoxia, loss of biodiversity, and
habitat degradation in coastal ecosystems; and (5) contributes to global climate change and
stratospheric ozone depletion, which can in turn affect ecosystems and human health (see
Table 4-21).
TABLE 4-21. EFFECTS OF REACTIVE NITROGEN
Direct effects ofNr on ecosystems include:
- Increased productivity of Nr-limited natural ecosystems.
- Ozone-induced injury to crop, forest, and natural ecosystems and predisposition to attack by pathogens
and insects.
- Acidification and eutrophication effects on forests, soils, and freshwater aquatic ecosystems.
- Eutrophication and hypoxia in coastal ecosystems.
- N-saturation of soils in forests and other natural ecosystems.
- Biodiversity losses in terrestrial and aquatic ecosystems and invasions by N-loving weeds.
- Changes in abundance of beneficial soil organisms that alter ecosystem functions.
Indirect effects ofNr on other societal values include:
- Increased wealth and well being of human populations in many parts of the world.
- Significant changes in patterns of land use.
- Regional hazes that decrease visibility at scenic vistas and airports.
- Depletion of stratospheric ozone by N2O emissions.
- Global climate change induced by emissions of N2O and formation of tropospheric ozone.
- Damage to useful materials and cultural artifacts by ozone, other oxidants, and acid deposition.
- Long-distance transport of Nr, which causes harmful effects in countries distant from emission sources
and/or increased background concentrations of zone and fine PM.
In addition to these effects, it is important to recognize that:
- The magnitude of Nr flux often determines whether effects are beneficial or detrimental.
- All of these effects are linked by biogeochemical circulation pathways of Nr.
- Nr is easily transformed among reduced and oxidized forms in many systems. Nr is easily distributed by
hydrologic and atmospheric transport processes.
Changes in nitrogen supply can have a considerable effect on an ecosystem's nutrient
balance. Large chronic additions of nitrogen influence normal nutrient cycling and alter many
plant and soil processes involved in nitrogen cycling. "Nitrogen saturation" results when Nr
concentrations exceed the capacity of a system to use it. Saturation implies that some resource
other than nitrogen is limiting biotic function. Water and phosphorus for plants and carbon for
microorganisms are most likely to be secondary limiting factors. The appearance of nitrogen in
4-230
-------
soil solution is an early symptom of excess nitrogen. In the final stage, disruption of ecosystem
structure becomes apparent.
Possible ecosystem responses to nitrogen saturation include (1) permanently increased
foliar nitrogen and reduced foliar phosphorus and lignin caused by the lower availability of
carbon, phosphorus, and water; (2) reduced productivity in conifer stands because of disruptions
of physiological function; (3) decreased rootbiomass and increased nitrification and nitrate
leaching; (4) reduced soil fertility, resulting from increased cation leaching, increased nitrate
and aluminum concentrations in streams; (5) decreased water quality; and (6) changes in soil
microbial communities from predominantly fungal (mycorrhizal) communities to those
dominated by bacteria during saturation.
Although soils of most North American forest ecosystems are nitrogen-limited,
some exhibit severe symptoms of nitrogen saturation (See Table 4-14). They include the
high-elevation, nonaggrading spruce-fir ecosystems in the Appalachian Mountains; the eastern
hardwood watersheds at the Fernow Experimental Forest near Parsons, WV; forests in southern
California, the southwestern Sierra Nevada in Central California; and the Front Range in
northern Colorado. The mixed conifer forest and chaparral watershed with high smog exposure
in the Los Angeles Air Basin exhibit the highest stream water NO3 concentrations for wildlands
in North America.
Increases in soil nitrogen play a selective role in ecosystems. Plants adapted to living in an
environment of low nitrogen availability will be replaced by nitrophilic plants capable of using
increased nitrogen because they have a competitive advantage when nitrogen becomes more
readily available. Plant succession patterns and biodiversity are affected significantly by chronic
nitrogen additions in some North American ecosystems. Long-term nitrogen fertilization studies
in both New England and Europe suggest that some forests receiving chronic inputs of nitrogen
may decline in productivity and experience greater mortality. Declining coniferous forest
stands with slow nitrogen cycling may be replaced by deciduous fast-growing forests that
cycle nitrogen.
Linked to the nitrogen cascade is the deposition of Nr and sulfates and the associated
hydrogen ion in acidic precipitation, a critical environmental stress that affects forest landscapes
and aquatic ecosystems in North America, Europe, and Asia. Composed of ions, gases, and
particles derived from gaseous emissions of sulfur dioxide (SO2), nitrogen oxides (NOX),
4-231
-------
ammonia (NH3) and particulate emissions of acidifying and neutralizing compounds, acidic
precipitation varies highly across time and space. Its deposition and the resulting soil acidity can
lead to nutrient deficiencies and to high aluminum-to-nutrient ratios that limit uptake of calcium
and magnesium and create a nutrient deficiency. Aluminum accumulation in root tissue can
reduce calcium uptake and causes Ca2+deficiencies. Tree species can be adversely affected if
altered Ca:Al ratios impair the uptake of calcium or magnesium. Calcium is essential in the
formation of wood and the maintenance of the primary plant tissues necessary for tree growth.
Studies suggest that the decline of red spruce stands in Vermont may be related to the Ca:Al
ratio (Johnson and Lindberg, 1992b).
The evidence of the effects of wet and dry particulate deposition SO4 and Nr species on
nutrient cycling in forest ecosystems is provided by the IPS. The deposition data from the study
illustrates several important aspects of the atmospheric exposure characteristics across a wide
elevational gradient and over a wide spatial scale. Atmospheric deposition plays a significant
role in the biogeochemical cycles at all IPS sites, but is most important in the east at high-
elevation sites. The flux of the sulfate ion, Nr compounds, and H+ ions in throughfall at all sites
is dominated by atmospheric deposition. Atmospheric deposition may have significantly
affected the nutrient status of some IPS sites through the mobilization of aluminum by impeding
cation uptake. Nitrates and sulfate are the dominant anions in the Smokies. Pulses of nitrates
are the major causes of aluminum pulses in soil solutions. However, the connection between
aluminum mobilization and forest decline is not clear, hence, aluminum mobilization presents a
situation worthy of further study.
Notable impacts of excess nitrogen deposition also have been observed with regard to
aquatic systems. For example, atmospheric nitrogen deposition into soils in watershed areas
feeding into estuarine sound complexes (e.g., the Pamlico Sound of North Carolina) appear to
contribute to excess nitrogen flows in runoff (especially during and after heavy rainfall events
such as hurricanes) from agricultural practices or other uses (e.g., fertilization of lawns or
gardens). Massive influxes of such nitrogen into watersheds and sounds can lead to dramatic
decreases in water oxygen and increases in algae blooms that can cause extensive fish kills and
damage to commercial fish and sea food harvesting.
The critical loads concept is useful for estimating the amounts of pollutants (e.g., Nr and
acidic precipitation) that sensitive ecosystems can absorb on a sustained basis without
4-232
-------
experiencing measurable degradation. The estimation of ecosystem critical loads requires an
understanding of how an ecosystem will respond to different loading rates in the long term and
can be of special value for ecosystems receiving chronic deposition of Nr and sulfur
independently and as acidic deposition when in combination.
Changes in the soil can result from the deposition of heavy metals. Exposures to heavy
metals are highly variable, depending whether deposition is by wet or dry processes. Few (e.g.,
Cu, Ni, Zn) have been documented to have direct phytoxicity under field conditions. Exposure
to coarse particles of natural origin and elements such as Fe and Mg are more likely to occur via
dry deposition, while fine particles of atmospheric origin are more likely to contain elements
such as Ca, Cr, Pb, N, and V. Ecosystems immediately downwind of major emissions sources
such as power generating, industrial, or urban complexes can receive locally heavy inputs.
By negatively affecting litter decomposition, heavy-metal accumulation presents the greatest
potential for influencing nutrient cycling. Microbial populations decreased and logarithmic rates
of microbial increase were prolonged as a result of cadmium toxicity. Additionally, the presence
of Ca, Cu, and Ni were observed to impair the symbiotic activity of fungi, bacteria, and
actinomycetes.
Phytochelatins produced by plants as a response to sublethal concentrations of heavy
metals are indicators of metal stress and can be used to indicate that heavy metals are involved in
forest decline. The increasing concentrations of phytochelatins across regions and at greater
altitudes associated with greater levels of forest injury implicate them in forest decline.
The ambient concentration of particles, the parameter for which there is most data
(Chapter 3), is, at best, an indicator of exposure. The amount of PM entering the immediate
plant environment and deposited onto the plant surfaces or soil in the vicinity of the roots,
determines the biological effect. Three major routes are involved during the wet and dry
deposition processes: (1) precipitation scavenging in which particles are deposited in rain and
snow; (2) occult (fog, cloud water, and mist interception); and (3) dry deposition, a much slower,
yet more continuous removal to plant surfaces.
Deposition of PM on the surfaces of above-ground plant parts can have either a physical
and/or a chemical effect. Particles transferred from the atmosphere to plant surfaces may cause
direct effects if they (1) reside on the leaf, twig, or bark surface for an extended period; (2) are
taken up through the leaf surface; or (3) are removed from the plant via resuspension to the
4-233
-------
atmosphere, washing by rainfall, or litter-fall with subsequent transfer to the soil. Ecosystem
effects have been observed in the neighborhoods of limestone quarries.
Secondary organics formed in the atmosphere have been variously subsumed under the
following terms: toxic substances, pesticides, HAPS, air toxics, SVOCs, and POPs. The
substances alluded to under the above headings are controlled under CAA Sect. 112, Hazardous
Air Pollutants, not as criteria pollutants controlled by NAAQS under CAA Sections 108 and 109
(U.S. Code, 1991). Their possible effects on humans and ecosystems are discussed in many
other government documents and publications. They are noted in this chapter because, in the
atmosphere, many of the chemical compounds are partitioned between gas and particle phases
and are deposited as PM. As particles, they become airborne and can be distributed over a wide
area and affect remote ecosystems. Some of the chemical compounds are of concern to humans
because they may reach toxic levels in the food chains of both animals and humans, whereas
others tend to decrease or maintain the same toxicity as they move through food chains.
An important characteristic of fine particles is their ability to affect the flux of solar
radiation passing through the atmosphere directly, by scattering and absorbing solar radiation,
and, indirectly, by acting as cloud condensation nuclei that, in turn, influence the optical
properties of clouds. Regional haze has been estimated to diminish surface solar visible
radiation by approximately 8%. Crop yields have been reported as being sensitive to the amount
of sunlight received, and crop losses have been attributed to increased airborne particle levels in
some areas of the world.
Key conclusions and findings about PM-related effects on vegetation and ecosystems
include the following:
• A number of ecosystem-level conditions (e.g., nitrogen saturation, terrestrial and aquatic
acidification, coastal eutrophication) that can lead to negative impacts on human health
and welfare have been associated with chronic, long-term exposure of ecosystems to
elevated inputs of compounds containing Nr, sulfur and/or associated hydrogen ions.
• Some percentage of total ecosystem inputs of these chemicals is contributed by deposition
of atmospheric particles, although the percentage greatly varies temporally and
geographically and has not generally been well quantified.
• Unfortunately, our ability to relate ambient concentrations of PM to ecosystem response is
hampered by the following significant data gaps and uncertainties:
4-234
-------
- The lack of a long-term, historic database of annual speciated PM deposition rates
in sensitive ecosystems precludes establishing relationships between PM deposition
(exposure) and ecosystem response at this time. Except for data from the IMPROVE
network and from some CASTNet sites, much of the PM monitoring effort has focused
on urban or near urban areas, rather than on sensitive ecosystems.
- Modeled deposition rates, used in the absence of monitored data, can be highly
uncertain, since there are a multitude of factors that influence the amounts of PM
that get deposited from the air onto sensitive receptors. These factors include the
mode of deposition (wet, dry, occult), windspeed, surface roughness/stickiness,
elevation, particle characteristics (e.g., size, shape, chemical composition, etc.),
and relative humidity.
- Each ecosystem is unique from all others, since each has developed within a context
framed by the topography, underlying bedrock, soils, climate, meteorology, hydrologic
regime, natural and land use history, species associations that co-occur at that location
(i.e., soil organisms, plants, etc.), and successional stage. Because of this variety, and
insufficient baseline data on each of these features for most ecosystems, it is currently
impossible to extrapolate with much confidence any effect from one ecosystem to
another, or to predict an appropriate "critical load." Thus, a given PM deposition rate
or load of nitrates in one ecosystem may produce entirely different responses than the
same deposition rate at another location.
- There remain large uncertainties associated with the length of residence time of Nr in
a particular ecosystem component or reservoir, and thus, its impact on the ecosystem
as it moves through the various levels of the N cascade.
As additional PM speciated air quality and deposition monitoring data become available,
there is much room for fruitful research into the areas of uncertainty identified above.
4.6.2 Particulate Matter-Related Effects on Visibility
Visibility is defined as the degree to which the atmosphere is transparent to visible light
and the clarity and color fidelity of the atmosphere. Visual range is the farthest distance a black
object can be distinguished against the horizontal sky. Visibility impairment is any humanly
perceptible change in visibility. For regulatory purposes, visibility impairment, characterized by
light extinction, visual range, contrast, and coloration, is classified into two principal forms:
(1) "reasonably attributable" impairment, attributable to a single source or small group of
sources, and (2) regional haze, any perceivable change in visibility caused by a combination of
many sources over a wide geographical area.
4-235
-------
Visibility is measured by human observation, light scattering by particles, the light
extinction-coefficient, and parameters related to the light-extinction coefficient (visual range and
deciview scale), and fine PM mass concentrations.
The air quality within a sight path will affect the illumination of the sight path by scattering
or absorbing solar radiation before it reaches the Earth's surface. The rate of energy loss with
distance from a beam of light is the light extinction coefficient. The light extinction coefficient
is the sum of the coefficients for light absorption by gases (oag), light scattering by gases (osg),
light absorption by particles (oap), and light scattering by particles (osp). Corresponding
coefficients for light scattering and absorption by fine and coarse particles are osfpand oafp, and
oscp and oacp, respectively. Visibility within a sight path longer than approximately 100 km (60
mi) is affected by the change in the optical properties of the atmosphere over the length of the
sight path.
Visual range was developed for, and continues to be used as, an aid in military operations
and to a lesser degree in transportation safety. Visual range is commonly taken to be the greatest
distance a dark object can be seen against the background sky. The deciview is an index of
haziness. A change of 1 or 2 deciviews is seen as a noticeable change in the appearance of a
scene.
Visibility impairment is associated with airborne particle properties, including size
distributions (i.e., fine particles in the 0.1- to 1.0-|im size range) and aerosol chemical
composition, and with relative humidity. With increasing relative humidity, the amount of
moisture available for absorption by particles increases, thus causing the particles to increase in
both size and volume. As the particles increase in size and volume, the light scattering potential
of the particles also generally increases. Visibility impairment is greatest in the eastern United
States and Southern California. In the eastern United States, visibility impairment is caused
primarily by light scattering by sulfate aerosols and, to a lesser extent, by nitrate particles and
organic aerosols, carbon soot, and crustal dust. Up to 86% of the haziness in the eastern United
States is caused by atmospheric sulfate. Farther West, scattering contributions to visibility
impairment decrease from 25 to 50%. Light scattering by nitrate aerosols is the major cause of
visibility impairment in southern California. Nitrates contribute about 45% to the total light
extinction in the West and up to 17% of the total extinction in the East. Organic particles are the
second largest contributors to light extinction in most U.S. areas. Organic carbon is the greatest
4-236
-------
cause of light extinction in the West, accounting for up to 40% of the total extinction and up to
18% of the visibility impairment in the East. Coarse mass and soil, primarily considered
"natural extinction," is responsible for some of the visibility impairment in the West, accounting
for up to 25% of the light extinction.
Key conclusions and findings about PM-related effects on visibility include the following:
• Airborne particles degrade visibility by scattering and absorbing light. These optical
properties can be well characterized in terms of a light extinction coefficient, which is the
fractional attentuation of light per unit distance. The extinction coefficient produced by a
given distribution of particle sizes and compositions is strictly proportional to the particle
mass concentration.
• The efficiency with which different particles attenuate light depends upon particle size,
with fine particles in the accumulation mode being much more important in causing
visibility impairment that coarse-mode particles. Thus, it is fine-particle mass
concentrations that tend to drive extinction coefficients in polluted air.
• The spatial and temporal variability in the observed extinction coefficient per mass of fine
particles is mainly due to the effects of particle-bound water, which varies with relative
humidity and is removed by drying when ambient PM2 5 mass concentrations are measured
using the Federal Reference Method.
• Improved relationships between ambient fine particle levels and visibility impairment
have been developed as a result of refinements in algorithms that relate particle size and
composition and relative humidity to light extinction, and thus, to visual range, as well as
the availability of much expanded databases of PM25 concentrations and related
compositional information and higher resolution visibility data.
• Various local initiatives to address visibility impairment have demonstrated the usefulness
of approaches now being used to evaluate public perceptions and attitudes about visibility
impairment and public judgments as to the importance of standards to improve visibility
relative to current conditions.
• Various such initiatives, conducted in areas with notable scenic vistas (e.g., Denver, CO),
have resulted in local standards that reflect what might be referred to as "adverse
thresholds" associated with a minimum visual range of approximately 40 to 60 km.
• These various local standards take into account that visibility impairment is an
instantaneous effect of ambient PM25 levels and that the public primarily values enhanced
visibility during daylight hours. These considerations are reflected in local standards that
are based on sub-daily averaging times (e.g., 4 to 6 hours), typically averaged across mid-
day hours.
4-237
-------
• This general convergence of visual range values and averaging times that have been
determined to be acceptable to the public in a number of such locations suggests that these
values and averaging times are relevant for consideration in assessing the need for a
national secondary standard to protect visibility in such areas.
4.6.3 Particulate Matter-Related Effects on Materials
Building materials (metals, stones, cements, and paints) undergo natural weathering
processes from exposure to environmental elements (wind, moisture, temperature fluctuations,
sunlight, etc.). Metals form a protective film of oxidized metal (e.g., rust) that slows
environmentally induced corrosion. On the other hand, the natural process of metal corrosion
from exposure to natural environmental elements is enhanced by exposure to anthropogenic
pollutants, particularly SO2, that render the protective film less effective.
Dry deposition of SO2 enhances the effects of environmental elements on calcareous stones
(limestone, marble, and cement) by converting calcium carbonate (calcite) to calcium sulfate
dihydrate (gypsum). The rate of deterioration is determined by the SO2 concentration, the
stone's permeability and moisture content, and the deposition rate; however, the extent of the
damage to stones produced by the pollutant species apart from the natural weathering processes
is uncertain. Sulfur dioxide also has been found to limit the life expectancy of paints by causing
discoloration, loss of gloss, and thinning of the paint film layer.
A significant detrimental effect of particle pollution is the soiling of painted surfaces and
other building materials. Soiling changes the reflectance of opaque materials and reduces the
transmission of light through transparent materials. Soiling is a degradation process that requires
remediation by cleaning or washing, and, depending on the soiled surface, repainting. Available
data on pollution exposure indicates that particles can result in increased cleaning frequency of
the exposed surface and may reduce the usefulness of the soiled material. Attempts have been
made to quantify the pollutants exposure levels at which materials damage and soiling have been
perceived. However, to date, insufficient data are available to advance our knowledge regarding
perception thresholds with respect to pollutant concentration, particle size, and chemical
composition.
4-238
-------
4.6.4 Effects of Atmospheric Particulate Matter on Global Warming
Processes and Transmission of Solar Ultraviolet Radiation
The physical processes (i.e., scattering and absorption) responsible for airborne particle
effects on transmission of solar visible and ultraviolet radiation are the same as those responsible
for visibility degradation. Scattering of solar radiation back to space and absorption of solar
radiation determine the effects of an aerosol layer on solar radiation.
Atmospheric particles greatly complicate projections of future trends in global warming
processes because of emissions of GHGs; consequent increases in global mean temperature;
resulting changes in regional and local weather patterns; and mainly deleterious (but sometimes
beneficial) location-specific human health and environmental effects. The body of available
evidence, ranging from satellite to in situ measurements of aerosol effects on radiation receipts
and cloud properties, is strongly indicative of an important role in climate for aerosols. This
role, however, is poorly quantified. No significant advances have been made in reducing the
uncertainties assigned to forcing estimates provided by the IPCC for aerosol-related forcing,
especially for black carbon-containing aerosol. The IPCC characterizes the scientific
understanding of GHG-related forcing as "high" in contrast to that for aerosol, which it describes
as "low" to "very low."
Quantification of the effect of anthropogenic aerosols on hydrological cycles requires more
information than is presently available regarding ecosystems responses to reduced solar radiation
and other changes occurring in the climate system. However, several global-scale studies
indicate that aerosol cooling alone can slow down the hydrological cycle, while cooling plus the
nucleation of additional cloud droplets can dramatically reduce precipitation rates.
In addition to direct climate effects through the scattering and absorption of solar radiation,
particles also exert indirect effects on climate by serving as cloud condensation nuclei, thus
affecting the abundance and vertical distribution of clouds. The direct and indirect effects of
particles appear to have significantly offset global warming effects caused by the buildup of
GHGs on a globally-averaged basis. However, because the lifetime of particles is much shorter
than the time required for complete mixing within the northern hemisphere, the climate effects of
particles generally are felt much less homogeneously than are the effects of long-lived GHGs.
4-239
-------
Any effort to model the impacts of local alterations in particle concentrations on projected
global climate change or consequent local and regional weather patterns would be subject to
considerable uncertainty.
Atmospheric particles also complicate estimation of potential future impacts on human
health and the environment projected as possible to occur because of increased transmission of
SUVB through the Earth's atmosphere, secondary to stratospheric ozone depletion due to
anthropogenic emissions of chlorofluorcarbons (CFCs), halons, and certain other gases. The
transmission of SUVB radiation is affected strongly by atmospheric particles. Measured
attenuations of UV-B under hazy conditions range up to 37% of the incoming solar radiation.
Measurements relating variations in PM mass directly to SUVB transmission are lacking.
Particles also can affect the rates of photochemical reactions occurring in the atmosphere, e.g.,
those involved in catalyzing tropospheric ozone formation. Depending on the amount of
absorbing substances in the particles, photolysis rates either can be increased or decreased.
Thus, the atmospheric particle effects on SUVB radiation vary depending on the size and
composition of particles and can differ substantially over different geographic areas, and from
season to season over the same area. Any projection of the effects of location-specific airborne
PM alterations on increased atmospheric transmission of solar UV radiation (and associated
potential human health or environmental effects) due to stratospheric ozone-depletion would,
therefore, also be subject to considerable uncertainty.
4-240
-------
REFERENCES
Abdullah, U. M; Iqbal, M. Z. (1991) Response of automobile, stone and cement paniculate matters on stomatal
clogging of plants. Geobios (Jodhpur, India) 18: 196-202.
Aber, J. D.; Nadelhoffer, K. J.; Steudler, P.; Melillo, J. M. (1989) Nitrogen saturation in northern forest ecosystems:
excess nitrogen from fossil fuel combustion may stress the biosphere. BioScience 39: 378-386.
Aber, J. D.; Magill, A.; McNulty, S. G.; Boone, R. D.; Nadelhoffer, K. J.; Downs, M.; Hallett, R. (1995) Forest
biogeochemistry and primary production altered by nitrogen saturation. Water Air Soil Pollut. 85: 1665-1670.
Aber, J.; McDowell, W.; Nadelhoffer, K.; Magill, A.; Berntson, G.; Kamakea, M.; McNulty, S.; Currie, W.;
Rustad, L.; Fernandez, I. (1998) Nitrogen saturation in temperate forest ecosystems. BioScience 48: 921-934.
Ainsleigh, H. G. (1993) Beneficial effects of sun exposure on cancer mortality. Prev. Med. 22: 132-140.
Alexander, M. (1977) Introduction to soil microbiology. 2nd ed. New York, NY: John Wiley & Sons; pp. 225-230;
239-246; 272-286; 355-356.
Allen, M. F. (1991) The ecology of mycorrhizae. Cambridge, MA: Cambridge University Press.
Allen, A. G.; Harrison, R. M.; Nicholson, K. W. (1991) Dry deposition of fine aerosol to a short grass surface.
Atmos. Environ. Part A 25: 2671-2676.
Allen, E. R.; Cabrera, N.; Kim, J.-C. (1994) Atmospheric deposition studies in rural forested areas of the
southeastern United States. Presented at: 87th annual meeting & exhibition of the Air & Waste Management
Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no.
94-WP88.02.
Allen, E. B.; Padgett, P. E.; Bytnerowicz, A.; Minich, R. (1998) Nitrogen deposition effects on coastal sage
vegetation of southern California. USDA Forest Service Gen. Tech. Rep. PSW-GTR-166, pp. 131-139.
Altshuller, A. P.; Linthurst, R. A., eds. (1984) Acidic deposition phenomenon and its effects: critical assessment
review papers; volume I, atmospheric sciences; volume II, effects sciences. Washington, DC: U.S.
Environmental Protection Agency, Office of Research and Development; EPA report nos.
EPA/600/8-83/016aF-bF. Available from: NTIS, Springfield, VA; PB85-100030 andPB85-100048.
Andersen, C. P.; Rygiewicz, P. T. (1991) Stress interactions and mycorrhizal plant response: understanding carbon
allocation priorities. Environ. Pollut. 73: 217-244.
Anderson, J. B.; Baumgardner, R. E.; Mohnen, V. A.; Bowser, J. J. (1999) Cloud chemistry in the eastern United
States, as sampled from three high-elevation sites along the Appalachian Mountains. Atmos. Environ.
33: 5105-5114.
Aneja, V. P.; Murthy, A. B. (1994) Monitoring deposition of nitrogen-containing compounds in a high-elevation
forest canopy. Air Waste 44: 1109-1115.
Antonovics, J.; Bradshaw, A. D.; Turner, R. G. (1971) Heavy metal tolerance in plants. In: Cragg, J. B., ed.
Advances in ecological research. Volume 7. London, United Kingdom: Academic Press; pp. 1-85.
April, R.; Newton, R. (1992) Mineralogy and mineral weathering. In: Johnson, D. W.; Lindberg, S. E., eds.
Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
Springer-Verlag; pp. 378-425. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
Ecological studies: analysis and synthesis: v. 91).
Arizona Department of Environmental Quality. (2002) Air quality report. A.R.S. § 49-424.10. Phoenix, AZ: Arizona
Department of Environmental Quality; Air Quality Assessment Section. Available:
http://www.azdeq.gov/function/forms/download/2002/aq.pdf [29 September, 2004].
Arnott, W. P.; Moosmuller, H.; Rogers, C. F.; Jin, T.; Bruch, R. (1999) Photoacoustic spectrometer for measuring
light absorption by aerosol; instrument description. Atmos. Environ. 33: 2845-2852.
Arrhenius, S. (1896) On the influence of carbonic acid in the air upon the temperature of the ground. Philos. Mag.
41:237-276.
Askey, A.; Lyon, S. B.; Thompson, G. E.; Johnson, J. B.; Wood, G. C.; Sage, P. W.; Cooke, M. J. (1993) The effect
of fly-ash particulates on the atmospheric corrosion of zinc and mild steel. Corros. Sci. 34: 1055-1081.
Ausset, P.; Bannery, F.; Del Monte, M.; Lefevre, R. A. (1998) Recording of pre-industrial atmospheric environment
by ancient crusts on stone monuments. Atmos. Environ. 32: 2859-2863.
Ayensu, E.; Van R. Claasen, D.; Collins, M.; Dearing, A.; Fresco, L.; Gadgil, M.; Gitay, H.; Glaser, G.; Juma, C.;
Krebs, J.; lenton, R.; Lubchenco, J.; McNeely, J. A.; Mooney, H. A.; Pinstrup-Andersen, P.; Ramos, M.;
Raven, P.; Reid, W. V; Samper, C.; Sarukhan, J.; Schei, P.; Tundisi, J. G.; Watson, R. T.; Guanhua, X.;
Zakri, A. H. (1999) International ecosystem assessment. Science (Washington, DC) 286: 685-686.
Babich, H.; Stotzky, G. (1978) Effects of cadmium on the biota: influence of environmental factors. In: Perlman, D.,
ed. Advances in applied microbiology, volume 23. New York, NY: Academic Press; pp. 55-117.
4-241
-------
Bache, D. H. (1979a) Particle transport within plant canopies -1. A framework for analysis. Atmos. Environ.
13: 1257-1262.
Bache, D. H. (1979b) Paniculate transport within plant canopies - II. Prediction of deposition velocities. Atmos.
Environ. 13: 1681-1687.
Baedecker, P. A.; Edney, E. O.; Moran, P. I; Simpson, T. C.; Williams, R. S. (1991) Effects of acidic deposition on
materials. In: Irving, P. M, ed. Acidic deposition: state of science and technology, volume III: terrestrial,
materials, health and visibility effects. Washington, DC: The U.S. National Acid Precipitation Assessment
Program. (State of science and technology report no. 19).
Baedecker, P. A.; Reddy, M. M.; Reimann, K. J.; Sciammarella, C. A. (1992) Effects of acidic deposition on the
erosion of carbonate stone-experimental results from the U.S. National Acid Precipitation Assessment
Program (NAPAP). Atmos. Environ. Part B 26: 147-158.
Baldocchi, D. (1988) A multi-layer model for estimating sulfur dioxide deposition to a deciduous oak forest canopy.
Atmos. Environ. 22: 869-884.
Bange, M. P.; Hammer, G. L.; Rickert, K. G. (1997) Effect of radiation environment on radiation use efficiency and
growth of sunflower. Crop Sci. 37: 1208-1214.
Baron, J. S.; Rueth, H. M.; Wolfe, A. M.; Nydick, K. R.; Allstott, E. J.; Minear, J. T.; Moraska, B. (2000) Ecosystem
responses to nitrogen deposition in the Colorado front range. Ecosystems 3: 352-368.
Barrie, L. A.; Anlauf, K.; Wiebe, H. A.; Fellin, P. (1984) Acidic pollutants in air and precipitation at selected rural
locations in Canada. In: Hicks, B. B., ed. Deposition both wet and dry: [papers presented at the American
Chemical Society acid rain symposium]; March 1982; Las Vegas, NV. Boston, MA: Butterworth Publishers;
pp. 15-35. (Teasley, J. L, ed. Acid precipitation series: v. 4).
Barton, K. (1958) The influence of dust on atmospheric corrosion of metals. Werkst. Korros. 8/9: 547-549.
BBC Research & Consulting. (2003) Phoenix area visibility survey. Phoenix, AZ: Arizona Department of
Environmental Quality. Available: http://www.azdeq.gov/environ/air/download/vis_021903f.pdf [17
September, 2004].
Beier, C. (1991) Separation of gaseous and paniculate dry deposition of sulfur at a forest edge in Denmark.
J. Environ. Qual. 20: 460-466.
Beier, C.; Gundersen, P.; Rasmussen, L. (1992) A new method for estimation of dry deposition of particles based on
throughfall measurements in a forest edge. Atmos. Environ. Part A 26: 1553-1559.
Bellan, L. M.; Salmon, L. G.; Cass, G. R. (2000) A study on the human ability to detect soot deposition onto works
of art. Environ. Sci. Technol. 34: 1946-1952.
Bennett, J. H.; Hill, A. C. (1973) Absorption of gaseous air pollutants by a standardized plant canopy. J. AirPollut.
Control Assoc. 23: 203-206.
Bennett, J. H.; Hill, A. C. (1975) Interactions of air pollutants with canopies of vegetation. In: Mudd, J. B.;
Kozlowski, T. T., eds. Responses of plants to air pollution. New York, NY: Academic Press; pp. 273-306.
Bewley, R. J. F.; Parkinson, D. (1984) Effects of sulphur dioxide pollution on forest soil microorganisms. Can. J.
Microbiol. 30: 179-185.
Biefer, G. J. (1981) Atmospheric corrosion of steel in the Canadian Arctic. Mater. Perform. 20: 16-19.
Billings, W. D. (1978) Plants and the ecosystem. 3rded. Belmont, CA: Wadsworth Publishing Company, Inc.;
pp. 1-62, 83-108.
Binkley, D.; Valentine, D.; Wells, C.; Valentine, U. (1989) An empirical analysis of the factors contributing to
20-year decrease in soil pH in an old-field plantation of loblolly pine. Biogeochemistry 8: 39-54.
Blackburn, T. E.; Bairai, S. T.; Stedman, D. H. (1992) Solar photolysis of ozone to singlet D oxygen atoms.
J. Geophys. Res. [Atmos.] 97: 10,109-10,117.
Blaschke, H. (1990) Mycorrhizal populations and fine root development on Norway spruce exposed to controlled
doses of gaseous pollutants and simulated acidic rain treatments. Environ. Pollut. 68: 409-418.
Bobbink, R. (1998) Impacts of tropospheric ozone and airborne nitrogenous pollutants on natural and semi-natural
ecosystems: a commentary. NewPhytol. 139: 161-168.
Bobbink, R.; Hornung, M.; Roelofs, J. G. M. (1998) The effects of air-borne nitrogen pollutants on species diversity
in natural and semi-natural European vegetation. J. Ecol. 86: 717-738.
Bondietti, E. A.; McLaughlin, S. B. (1992) Evidence of historical influences of acidic deposition on wood and soil
chemistry. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient cycling: a
synthesis of the integrated forest study. New York, NY: Springer-Verlag: pp. 358-377. (Billings, W. D.;
Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological studies analysis and synthesis: v. 91).
Bormann, F. H.; Likens, G. E. (1979) Catastrophic disturbance and the steady state in northern hardwood forests.
Am. Sci. 67: 660-669.
4-242
-------
Bowden, R. D.; Geballe, G. T.; Bowden, W. B. (1989) Foliar uptake of 15N from simulated cloud water by red
spruce (Picea rubens) seedlings. Can. J. For. Res. 19: 382-386.
Bowman, W. D. (2000) Biotic controls over ecosystem response to environmental change in alpine tundra of the
Rocky Mountains. Ambio 29: 396-400.
Bowman, W. D.; Steltzer, H. (1998) Positive feedbacks to anthropogenic nitrogen deposition in Rocky Mountain
alpine tundra. Ambio 27: 514-517.
Boyce, S. G. (1954) The salt spray community. Ecol. Monogr. 24: 29-67.
Bradley, R.; Burt, A. J.; Read, D. J. (1981) Mycorrhizal infection and resistance to heavy metal toxicity in Calluna
vulgaris. Nature (London) 292: 335-337.
Brandt, C. J.; Rhoades, R. W. (1972) Effects of limestone dust accumulation on composition of a forest community.
Environ. Pollut. 3: 217-225.
Brandt, C. J.; Rhoades, R. W. (1973) Effects of limestone dust accumulation on lateral growth of forest trees.
Environ. Pollut. 4: 207-213.
Brook, J. R.; Zhang, L.; Di-Giovanni, F.; Padro, J. (1999) Description and evaluation of a model of deposition
velocities for routine estimates of air pollutant dry deposition over North America. Part I: model development.
Atmos. Environ. 33: 5037-5051.
Brookshire, D. S.; d'Arge, R. C.; Schulze, W. D.; Thayer, M. D. (1979) Methods development for assessing air
pollution control benefits: volume II. alternative benefit measures of air pollution control in the south coast air
basin of southern California. Washington, DC: U.S. Environmental Protection Agency, Office of Research
and Development; report no. EPA-600/6-79-001b.
Briihl, C.; Crutzen, P. J. (1988) On the disproportionate role of tropospheric ozone as a filter against solar UV-B
radiation. Geophys. Res. Lett. 16: 703-706.
Brumme, R.; Leimcke, U.; Matzner, E. (1992) Interception and uptake of NH4 and NO3 from wet deposition by
above-ground parts of young beech (Fagus silvatica L.) trees. Plant Soil 142: 273-279.
Brunei, J.; Diekmann, M.; Falkengren-Grerup, U. (1998) Effects of nitrogen deposition on field layer vegetation in
south Swedish oak forests. Environ. Pollut. (Oxford, U.K.) 102(suppl. 1): 35-40.
Bugini, R.; Tabasso, M. L.; Realini, M. (2000) Rate of formation of black crusts on marble. A case study. J. Cult.
Heritage 1: 111-116.
Businger, J. A. (1986) Evaluation of the accuracy with which dry deposition can be measured with current
micrometeorological techniques. J. Clim. Appl. Meteorol. 25: 1100-1124.
Butlin, R. N.; Coote, A. T.; Devenish, M.; Hughes, I. S. C.; Hutchens, C. M.; Irwin, J. G.; Lloyd, G. O.; Massey,
S. W.; Webb, A. H.; Yates, T. J. S. (1992a) Preliminary results from the analysis of metal samples from the
National Materials Exposure Programme (NMEP). Atmos. Environ. Part B 26: 199-206.
Butlin, R. N.; Coote, A. T.; Devenish, M.; Hughes, I. S. C.; Hutchens, C. M.; Irwin, J. G.; Lloyd, G. O.; Massey,
S. W.; Webb, A. H.; Yates, T. J. S. (1992b) Preliminary results from the analysis of stone tablets from the
National Materials Exposure Programme (NMEP). Atmos. Environ. Part B 26: 189-198.
Bytnerowicz, A.; Fenn, M. E. (1996) Nitrogen deposition in California forests: a review. Environ. Pollut.
92: 127-146.
Bytnerowicz, A.; Miller, P. R.; Olszyk, D. M. (1987a) Dry deposition of nitrate, ammonium and sulfate to a
Ceanothus crassifolius canopy and surrogate surfaces. Atmos. Environ. 21: 1749-1757.
Bytnerowicz, A.; Miller, P. R.; Olszyk, D. M.; Dawson, P. J.; Fox, C. A. (1987b) Gaseous and paniculate air
pollution in the San Gabriel Mountains of southern California. Atmos. Environ. 21: 1805-1814.
Cahill, T. A.; Eldred, R. A.; Wakabayashi, P. H. (1996) Trends in fine particle concentrations at Great Smoky
Mountains National Park. For presentation at: Annual meeting of the Air & Waste Management Association;
June; Nashville, TN. Pittsburgh, PA: Air & Waste Management Association; paper no. 96-MP1A.05.
Campbell, G. G.; Schurr, G. G.; Slawikowski, D. E.; Spence, J. W. (1974) Assessing air pollution damage to
coatings. J. Paint Technol. 46: 59-71.
Camuffo, D. (1995) Physical weathering of stones. In: Saiz-Jimenez, C., ed. The deterioration of monuments:
proceedings of the 2nd international symposium on biodeterioration and biodegradation; January 1994;
Sevilla, Spain. Sci. Total Environ. 167: 1-14.
Cape, J. N. (1993) Direct damage to vegetation caused by acid rain and polluted cloud: definition of critical levels
for forest trees. Environ. Pollut. 82: 167-180.
Cape, J. N.; Sheppard, L. J.; Fowler, D.; Harrison, A. F.; Parkinson, J. A.; Dao, P.; Paterson, I. S. (1992)
Contribution of canopy leaching to sulphate deposition in a Scots pine forest. Environ. Pollut. 75: 229-236.
Carey, W. F. (1959) Atmospheric deposits in Britain: a study of dinginess. Int. J. Air Pollut. 2: 1-26.
4-243
-------
Ceccotti, S. P.; Messick, D. L. (1997) A global review of crop requirements, supply, and environmental impact on
nutrient sulphur balance. In: Cram, W. I; De Kok, L. I; Stulen, L; Brunold, C.; Rennenberg, H., eds. Sulphur
metabolism in higher plants. Leiden, The Netherlands: Backhuys Publishers; pp. 155-163.
Chamberlain, A. C. (1975) The movement of particles in plant communities. In: Monteith, J. L., ed. Vegetation and
the atmosphere: v. 1, principles. London, United Kingdom: Academic Press; pp. 155-203.
Chamberlain, A. C.; Little, P. (1981) Transport and capture of particles by vegetation. In: Grace, J.; Ford, E. D.;
Jarvis, P. G., eds. Plants and their atmospheric environment: the 21st symposium of the British Ecological
Society; 1979; Edinburgh, United Kingdom. Oxford, United Kingdom: Blackwell Scientific Publications;
pp. 147-173.
Chameides, W. L.; Yu, H.; Liu, S. C.; Bergin, M; Zhou, X.; Mearns, L.; Wang, G.; Kiang, C. S.; Saylor, R. D.;
Luo, C.; Huang, Y.; Steiner, A.; Giorgi, F. (1999) Case study of the effects of atmospheric aerosols and
regional haze on agriculture: an opportunity to enhance crop yields in China through emission controls?
Proc. Natl. Acad. Sci. U. S. A. 96: 13626-13633.
Chapin, F. S., III. (1991) Integrated responses of plants to stress: a centralized system of physiological responses.
BioScience 41: 29-36.
Chapin, F. S., Ill; Ruess, R. W. (2001) The roots of the matter. Nature (London) 411: 749-752.
Chapin, F. S., Ill; Bloom, A. J.; Field, C. B.; Waring, R. H. (1987) Plant responses to multiple environmental
factors. BioScience 37: 49-57.
Chapin, F. S., Ill; Sala, O. E.; Burke, I. C.; Grime, J. P.; Hooper, D. U.; Lauenroth, W. K.; Lombard, A.; Mooney,
H. A.; Mosier, A. R.; Naeem, S.; Pacala, S. W.; Roy, J.; Steffen, W. L.; Tilman, D. (1998) Ecosystem
consequences of changing biodiversity. BioScience 48: 45-52.
Charlson, R. J.; Ahlquist, N. C.; Horvath, H. (1968) On the generality of correlation of atmospheric aerosol mass
concentration and light scatter. Atmos. Environ. 2: 455-464.
Charlson, R. J.; Schwartz, S. E.; Hales, J. M.; Cess, R. D.; Coakley, J. A., Jr.; Hansen, J. E.; Hofmann, D. J. (1992)
Climate forcing by anthropogenic aerosols. Science (Washington, DC) 255: 423-430.
Chestnut, L. (1997) Methodology for estimating values for changes in visibility at national parks [draft
memorandum]. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards; contract no. 68-D-98-001; EPA air docket A-96-56, document no. VI-B-09-(ooo).
Chestnut, L. G.; Dennis, R. L. (1997) Economic benefits of improvements invisibility: acid rain. Provisions of the
1990 clean air act amendments. J. Air Waste Manage. Assoc. 47: 395-402.
Chestnut, L. G.; Rowe, R. D. (1990) Preservation values for visibility protection at the national parks (draft final
report). Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, Economic Analysis Branch.
Chevone, B. I.; Herzfeld, D. E.; Krupa, S. V.; Chappelka, A. H. (1986) Direct effects of atmospheric sulfate
deposition on vegetation. J. AirPollut. Control Assoc. 36: 813-815.
Choudhury, B. J. (1987) Relationships between vegetation indices, radiation absorption, and net photosynthesis
evaluated by a sensitivity analysis. Remote Sens. Environ. 22: 209-233.
Chow, J. C.; Watson, J. G.; Pritchett, L.; Lowenthal, D.; Frazier, C.; Neuroth, G.; Evans, K.; Moon, D. (1990)
Wintertime visibility in Phoenix, Arizona. Presented at: the 83rd annual meeting & exhibition; June;
Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; paper no. 90-66.6.
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Richards, L. W. (2002a) Comparability between PM25 and particle
light scattering measurements. Environ. Monit. Assess. 79: 29-45.
Chow, J. C.; Bachmann, J. D.; Wierman, S. S. G.; Mathai, C. V.; Malm, W. C.; White, W. H.; Mueller, P. K.;
Kumar, N.; Watson, J. G. (2002b) Visibility: science and regulation. J. Air Waste Manage. Assoc.
52: 973-999.
Chughtai, A. R.; Williams, G. R.; Atteya, M. M. O.; Miller, N. J.; Smith, D. M. (1999) Carbonaceous particle
hydration. Atmos. Environ. 33: 2679-2687.
Cione, N. K.; Padgett, P. E.; Allen, E. B. (2002) Restoration of a native shrubland impacted by exotic grasses,
frequent fire and nitrogen deposition in southern California. Restor. Ecol. 10: 376-384.
Clarke, J. F.; Edgerton, E. S.; Martin, B. E. (1997) Dry deposition calculations for the clean air status and trends
network. Atmos. Environ. 31: 3667-3678.
Clarkson, D. T.; Sanderson, J. (1971) Inhibition of the uptake and long-distance transport of calcium by aluminum
and other polyvalent cations. J. Exp. Bot. 22: 837-851.
Clevenger, T. E.; Saiwan, C.; Koirtyohann, S. R. (1991) Lead speciation of particles on air filters collected in the
vicinity of a lead smelter. Environ. Sci. Technol. 25: 1128-1133.
Clough, W. S. (1975) The deposition of particles on moss and grass surfaces. Atmos. Environ. 9: 1113-1119.
4-244
-------
Cocker, D. R., Ill; Whitlock, N. E.; Flagen, R. C.; Seinfield, J. H. (2001) Hygroscopic properties of Pasadena,
California aerosol. Aerosol Sci. Technol. 35: 637-647.
Cole, D. W.; Johnson, D. W. (1977) Atmospheric sulfate additions and cation leaching in a Douglas fir ecosystem.
Water Resour. Res. 13: 313-317.
Collett, J. L., Jr.; Hoag, K. J.; Sherman, D. E.; Bator, A.; Richards, L. W. (1999) Spatial and temporal variations in
San Joaquin Valley fog chemistry. Atmos. Environ. 33: 129-140.
Cooke, R. U.; Gibbs, G. B. (1994) Crumbling heritage? Studies of stone weathering in polluted atmospheres. Atmos.
Environ. 28: 1355-1356.
Costanza, R.; d'Arge, R.; De Groot, R.; Farber, S.; Grasso, M; Harmon, B.; Limburg, K.; Naeem, S.; O'Neill, R. V.;
Paruelo, J.; Raskin, R. G.; Sutton, P.; Van Den Belt, M. (1997) The value of the world's ecosystem services
and natural capital. Nature (London) 387: 253-259.
Cotrufo, M. F.; De Santo, A. V; Alfani, A.; Bartoli, G.; De Cristofaro, A. (1995) Effects of urban heavy metal
pollution on organic matter decomposition in Quercus ilex L. woods. Environ. Pollut. 89: 81-87.
Cowling, J. E.; Roberts, M. E. (1954) Paints, varnishes, enamels, and lacquers. In: Greathouse, G. A.; Wessel, C. J.,
eds. Deterioration of materials: causes and preventive techniques. New York, NY: Reinhold Publishing
Corporation; pp. 596-645.
Cowling, E. B.; Erisman, J. W.; Smeulders, S. M.; Holman, S. C.; Nicholson, B. M. (1998) Optimizing air quality
management in Europe and North America: justification for integrated management of both oxidized and
reduced forms of nitrogen. Environ. Pollut. 102(suppl. 1): 599-608.
Cramer, S. D.; McDonald, L. G.; Bhagia, G.; Flinn, D. R.; Linstrom, P. J.; Carter, J. P. (1989) Effects of acid
deposition on the atmospheric corrosion of structural metals. Research Triangle Park, NC: U.S.
Environmental Protection Agency; internal report. Incorporated in: Irving, P. M., ed. Acidic Deposition:
State of Science and Technology. Volume III: Terrestrial, Materials, Health and Visibility Effects; report 19,
section 2. Washington, DC: National Acid Precipitation Assessment Program.
Creighton, P. J.; Lioy, P. J.; Haynie, F. H.; Lemmons, T. J.; Miller, J. L.; Gerhart, J. (1990) Soiling by atmospheric
aerosols in an urban industrial area. J. Air Waste Manage. Assoc. 40: 1285-1289.
Cronan, C. S.; Grigal, D. F. (1995) Use of calcium/aluminum ratios as indicators of stress in forest ecosystems.
J. Environ. Qual. 24: 209-226.
Cullen, J. J.; Neale, P. J. (1997) Biological weighting functions for describing the effects of ultraviolet radiation on
aquatic ecosystems. In: Hader, D.-P., ed. The effects of ozone depletion on aquatic ecosystems. Austin, TX:
Academic Press, pp. 97-118.
Dahlgren, R. A. (1994) Soil acidification and nitrogen saturation from weathering of ammonium-bearing rock.
Nature (London) 368: 838-841.
Daily, G. C. (1997) Introduction: what are ecosystem services? In: Daily, G. C., ed. Nature's services: societal
dependence on natural ecosystems. Washington, DC: Island Press; pp. 1-10.
Daily, G. C.; Ehrlich, P. R. (1999) Managing Earth's ecosystems: an interdisciplinary challenge. Ecosystems
2: 277-280.
Darley, E. F. (1966) Studies on the effect of cement-kiln dust on vegetation. J. Air Pollut. Control Assoc.
16: 145-150.
Dasch, J. M. (1987) Measurement of dry deposition to surfaces in deciduous and pine canopies. Environ. Pollut.
44: 261-277.
Dasch, J. M.; Cadle, S. H. (1985) Wet and dry deposition monitoring in southeastern Michigan. Atmos. Environ.
19: 789-796.
Dasch, J. M.; Cadle, S. H. (1986) Dry deposition to snow in an urban area. Water Air Soil Pollut. 29: 297-308.
Dassler, H.-G.; Ranft, H.; Rehn, K.-H. (1972) Zur Widerstandsfahigkeit von Geholzen gegentiber
Fluorverbindungen und Schwefeldioxid [The susceptibility of woody plants exposed to fluorine compounds
and SO2]. Flora (Jena) 161: 289-302.
Davidson, C. I.; Friedlander, S. K. (1978) A filtration model for aerosol dry deposition: application to trace metal
deposition from the atmosphere. J. Geophys. Res. 83: 2343-2352.
Davidson, C. I.; Wu, Y.-L. (1989) Dry deposition of trace elements. In: Pacyna, J.M.; Ottar, B., eds. Control and fate
of atmospheric trace metals: proceedings of the NATO advanced research workshop on fate and control of
toxic metals in the atmosphere; September 1988; Oslo, Norway. Boston, MA: Kluwer Academic Publishers;
pp. 147-202. (NATO ASI series C, mathematical and physical sciences: v. 268).
Davidson, C. I.; Wu, Y.-L. (1990) Dry deposition of particles and vapors. In: Lindberg, S. E.; Page, A. L.; Norton,
S. A., eds. Acidic precipitation: v. 3, sources, deposition, and canopy interactions. New York, NY:
Springer-Verlag; pp. 103-216.
4-245
-------
Davidson, C. I.; Miller, J. M; Pleskow, M. A. (1982) The influence of surface structure on predicted particle dry
deposition to natural grass canopies. Water Air Soil Pollut. 18: 25-43.
Davidson, K.; Deck, L.; McCubbin, D.; Post, E. (2000a) Out of sight: the science and economics of visibility
impairment. Cambridge, MA: Abt Associates, Inc.
Davidson, C. L; Tang, W.; Finger, S.; Etyemezian, V.; Striegel, M. F.; Sherwood, S. I. (2000b) Soiling patterns on a
tall limestone building: changes over 60 years. Environ. Sci. Technol. 34: 560-565.
Davis, G. D.; Shaw, B. A.; Arah, C. O.; Fritz, T. L.; Moshier, W. C.; Simpson, T. C.; Moran, P. J.; Zankel, K. L.
(1990) Effects of SO2 deposition on painted steel surfaces. Surf. Interface Anal. 15: 107-112.
Day, D. E.; Malm, W. C.; Kreidenweis, S. M.; Tree, R. M. (1997) Aerosol light scattering measurements: a
comparison of differently configured optec nephelometers. In: Visual air quality: aerosols and global radiation
balance: proceedings of a specialty conference sponsored by the Air & Waste Management Association and
the American Geophysical Union; September; Bartlett, NH. Pittsburgh, PA: Air & Waste Management
Association; pp. 952-962. (A&WMA publication VIP-76).
DeHayes, D. H.; Schaberg, P. G.; Hawley, G. J.; Strimbeck, G. R. (1999) Acid rain impacts on calcium nutrition and
forest health. BioScience 49: 789-800.
Del Monte, M.; Vittori, O. (1985) Air pollution and stone decay: the case of Venice. Endeavour 9: 117-122.
Diakumaku, E.; Gorbushina, A. A.; Krumbein, W. E.; Panina, L.; Soukharjevski, S. (1995) Black fungi in marble
and limestones~an aesthetical, chemical and physical problem for the conservation of monuments.
In: Saiz-Jimenez, C., ed. The deterioration of monuments: proceedings of the 2nd international symposium on
biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci. Total Environ. 167: 295-304.
Dickerson, R. R.; Kondragunta, S.; Stenchikov, G.; Civerolo, K. L.; Doddridge, B. G.; Holben, B. N. (1997)
The impact of aerosols on solar ultraviolet radiation and photochemical smog. Science (Washington, DC)
278: 827-830.
Dobbins, R. A.; Mulholland, G. W.; Bryner, N. P. (1994) Comparison of a fractal smoke optics model with light
extinction measurements. Atmos. Environ. 28: 889-897.
Dolske, D. A. (1988) Dry deposition of airborne sulfate and nitrate to soybeans. Environ. Pollut. 53: 1-12.
Dolske, D. A. (1995) Deposition of atmospheric pollutants to monuments, statues, and buildings. Sci. Total Environ.
167: 15-31.
Dolske, D. A.; Gatz, D. F. (1984) Field intercomparison of sulfate dry deposition monitoring and measurement
methods: preliminary results. In: Hicks, B. B., ed. Deposition both wet and dry: [papers presented at the
American Chemical Society acid rain symposium]; March 1982; Las Vegas, NV. Boston, MA: Butterworth
Publishers; pp. 121-131. (Teasley, J. L, ed. Acid precipitation series: v. 4).
Draaijers, G. P. J.; Erisman, J. W. (1993) Atmospheric sulphur deposition to forest stands: throughfall estimates
compared to estimates from inference. Atmos. Environ. Part A 27: 43-55.
Draaijers, G. P. J.; Ivens, W. P. M. F.; Bleuten, W. (1988) Atmospheric deposition in forest edges measured by
monitoring canopy throughfall. Water Air Soil Pollut. 42: 129-136.
Draaijers, G. P. J.; van Ek, R.; Meijers, R. (1992) Research on the impact of forest stand structure on atmospheric
deposition. Environ. Pollut. 75: 243-249.
Dreisinger, B. R.; McGovern, P. C. (1970) Monitoring atmospheric sulphur dioxide and correlating its effects on
crops and forests in the Sudbury area. In: Linzon, S. N., ed. Impact of air pollution on vegetation conference;
April; Toronto, ON, Canada. Pittsburgh, PA: Air Pollution Control Association.
Driscoll, C. T.; Van Dreason, R. (1993) Seasonal and long-term temporal patterns in the chemistry of Adirondack
lakes. Water Air Soil Pollut. 67: 319-344.
Driscoll, C. T.; Wyskowski, B. J.; DeStaffan, P.; Newton, R. M. (1989) Chemistry and transfer of aluminum in a
forested watershed in the Adirondack region of New York, USA. In: Lewis, T. E., ed. Environmental
chemistry and toxicology of aluminum. Chelsea, MI: Lewis Publishers, Inc.; pp. 83-105.
Driscoll, C. T.; Newton, R. M.; Gubala, C. P.; Baker, J. P.; Christensen, S. W. (1991) Adirondack mountains.
In: Charles, D. F., ed. Acidic deposition and aquatic ecosystems: regional case studies. New York, NY:
Springer-Verlag; pp. 133-202.
Driscoll, C. T.; Lawrence, G. B.; Bulger, A. J.; Butler, T. J.; Cronan, C. S.; Eagar, C.; Lambert, K. F.; Likens, G. E.;
Stoddard, J. L.; Weathers, K. C. (2001) Acidic deposition in the northeastern United States: sources and
inputs, ecosystem effects, and management strategies. BioScience 51: 180-198.
Driscoll, C. T.; Whitall, D.; Aber, J.; Boyer, E.; Castro, M.; Cronan, C.; Goodale, C. L.; Groffman, P.; Hopkinson,
C.; lambert, K.; Lawrence, G.; Ollinger, S. (2003) Nitrogen pollution in the northeastern United States:
sources, effects, and management options. BioScience 53: 357-374.
Duan, B.; Fairall, C. W.; Thomson, D. W. (1988) Eddy correlation measurements of the dry deposition of particles
in wintertime. J. Appl. Meteorol. 27: 642-652.
4-246
-------
Duyzer, J. H.; Bouman, A. M. M; Van Aalst, R. M; Diederen, H. S. M. A. (1987) Assessment of dry deposition
fluxes of NH3 and NH4+ over natural terrains. In: Asman, W. A. H.; Diederen, H. S. M. A., eds. Ammonia and
acidification: proceedings of a symposium of the European Association for the Science of Air Pollution
(EURASAP); April; Bilthoven, The Netherlands. European Association for the Science of Air Pollution;
pp. 97-106.
Eaton, J. S.; Likens, G. E.; Bormann, F. H. (1973) Throughfall and stemflow chemistry in a northern hardwood
forest. J. Ecol. 61: 495-508.
Edgerton, E. S.; Lavery, T. F.; Boksleitner, R. P. (1992) Preliminary data from the USEPA dry deposition network:
1989. Environ. Pollut. 75: 145-156.
Edgerton-Warburton, L. M.; Allen, E. B. (2000) Shifts in arbuscular mycorrhizal communities along an
anthropogenic gradient nitrogen deposition gradient. Ecol. Appl. 10: 484-496.
Edney, E. O. (1989) Paint coatings: controlled field and chamber experiments. Research Triangle Park, NC:
U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment Laboratory; report
no. EPA/600/3-89/032. Available from: NTIS, Springfield, VA; PB89-189849.
Edney, E. O.; Stiles, D. C; Corse, E. W.; Wheeler, M. L.; Spence, J. W.; Haynie, F. H.; Wilson, W. E., Jr. (1988)
Field study to determine the impact of air pollutants on the corrosion of galvanized steel. Mater. Perform.
27: 47-50.
Edney, E. O.; Cheek, S. F.; Corse, E. W.; Spence, J. W.; Haynie, F. H. (1989) Atmospheric weathering caused by
dry deposition of acidic compounds. J. Environ. Sci. Health Part A 24: 439-457.
Edwards, R. S.; Claxton, S. M. (1964) The distribution of air-borne salt of marine origin in the Aberystwyth area.
J. Appl. Ecol. 1:253-263.
El-Shobokshy, M. S. (1985) The dependence of airborne paniculate deposition on atmospheric stability and surface
conditions. Atmos. Environ. 19: 1191-1197.
Eldering, A.; Cass, G. R.; Moon, K. C. (1994) An air monitoring network using continuous particle size distribution
monitors: connecting pollutant properties to visibility via Mie scattering calculations. Atmos. Environ.
28: 2733-2749.
Eldering, A.; Hall, J. R.; Hussey, K. J.; Cass, G. R. (1996) Visibility model based on satellite-generated landscape
data. Environ. Sci. Technol. 30: 361-370.
Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the
United States of America. Atmos. Environ. 28: 1009-1019.
Eldred, R. A.; Cahill, T. A.; Feeney, P. J. (1993) Comparison of independent measurements of sulfur and sulfate in
the IMPROVE network. Presented at: 86th annual meeting and exhibition of the Air & Waste Management
Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management Association; paper no.
93-RA-l 10.02.
Eldred, R. A.; Cahill, T. A.; Flocchini, R. G. (1997) Composition of PM25 andPM10 aerosols in the IMPROVE
network. In: Special issue on aerosols and atmospheric optics-part I, [published papers from the international
specialty conference; September 1994; Snowbird, UT]. J. Air Waste Manage. Assoc. 47: 194-203.
Eliason, S. A.; Allen, E. B. (1997) Exotic grass competition in suppressing native shrubland re-establishment.
Restor. Ecol. 5:245-255.
Ellenberg, H. (1987) Floristic changes due to eutrophication. In: Asman, W. A. H.; Diederen, S. M. A., eds.
Ammonia and acidification: proceedings of a symposium of the European Association for the Science of Air
Pollution (EURASAP); April; Bilthoven, The Netherlands. European Association for the Science of Air
Pollution; pp. 301-308.
Eller, B. M. (1977) Road dust induced increase of leaf temperature. Environ. Pollut. 13: 99-107.
Ely, D. W.; Leary, J. T.; Stewart, T. R.; Ross, D. M. (1991) The establishment of the Denver Visibility Standard.
Presented at: 84th annual meeting & exhibition of the Air & Waste Management Association; June;
Vancouver, British Columbia. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-48.4.
Eriksson, P.; Johansson, L.-G.; Strandberg, H. (1993) Initial stages of copper corrosion in humid air containing SO2
andNO2. J. Electrochem. Soc. 140: 53-59.
Erisman, J. W. (1993) Acid deposition onto nature areas in the Netherlands; part II. Throughfall measurements
compared to deposition estimates. Water Air Soil Pollut. 71: 81-99.
Erisman, J.-W.; Vermetten, A. W. M.; Asman, W. A. H.; Waijers-IJpelaan, A.; Slanina, J. (1988) Vertical
distribution of gases and aerosols: the behaviour of ammonia and related components in the lower
atmosphere. Atmos. Environ. 22: 1153-1160.
Erisman, J. W.; Van Pul, A.; Wyers, P. (1994) Parametrization of surface resistance for the quantification of
atmospheric deposition of acidifying pollutants and ozone. Atmos. Environ. 28: 2595-2607.
4-247
-------
Erisman, J. W.; Draaijers, G.; Duyzer, I; Hofschreuder, P.; VanLeeuwen, N.; Romer, F.; Ruijgrok, W.; Wyers, P.;
Gallagher, M. (1997) Particle deposition to forests-summary of results and application. Atmos. Environ.
o 1 . o ^ 1 n ^
31: 321-332.
Erlick, C.; Frederick, J. E.; Saxena, V. K.; Wenny, B. N. (1998) Atmospheric transmission in the ultraviolet and
visible: aerosols in cloudy atmospheres. J. Geophys. Res. 103: 31,541-31,556.
Eshleman, K. N.; Miller-Marshall, L. M.; Webb, J. R. (1995) Long-term changes in episodic acidification of streams
in Shenandoah National Park, Virginia (U.S.A.). Water Air Soil Pollut. 85: 517-522.
Estupinan, J. B.; Raman, S.; Crescenti, G. H.; Streicher, J. J.; Barnard, W. F. (1996) Effects of clouds and haze on
UV-B radiation. J. Geophys. Res. 101: 16,807-16,816.
Evans, E. G.; Pitchford, M. (1991) Federal visibility monitoring in the eastern United States. Presented at:
84th annual meeting and exhibition of the Air & Waste Management Association; June; Vancouver, BC,
Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-49.6.
Fahey, T. J.; Yavitt, J. B.; Joyce, G. (1988) Precipitation and throughfall chemistry mPinus contorta ssp. latifolia
ecosystems, southeastern Wyoming. Can. J. For. Res. 18: 337-345.
Falke, S. (2001) Web-based visibility information system. Available:
http://capita.wustl.edu/datawarehouse/datasets/ASOS/WebVIS.htm (28 May 2003).
Falkengren-Grerup, U. (1987) Long-term changes in pH of forest soils in southern Sweden. Environ. Pollut.
43:79-90.
Falkengren-Grerup, U. (1998) Nitrogen response of herbs and graminoids in experiments with simulated acid soil
solution. Environ. Pollut. 102(suppl. 1): 93-99.
Farber, R. J.; Welsing, P. R.; Rozzi, C. (1994) PM10 and ozone control strategy to improve visibility in the
Los Angeles Basin. Atmos. Environ. 28: 3277-3283.
Farmer, A. M. (1993) The effects of dust on vegetation~a review. Environ. Pollut. 79: 63-75.
Fassina, V.; Favaro, M.; Crivellari, F.; Naccari, A. (2001) The stone decay of monuments in relation to atmospheric
environment. Ann. Chim. 91: 767-774.
Federal Register. (1999) Regional haze regulations; final rule. F. R. (July 1) 64: 35,713-35,774.
Fenn, M. E.; Keifer, J. W. (1999) Throughfall deposition of nitrogen and sulfur in a Jeffrey pine forest in the
San Gabriel Mountains, Southern California. Environ. Pollut. 104: 179-187.
Fenn, M. E.; Poth, M. A. (1998) Indicators of nitrogen status in California forests. U.S.D.A. Forest Service Gen.
Tech. Rep. PSW-GTR-166.
Fenn, M. E.; Poth, M. A. (1999) Temporal and spatial trends in streamwater nitrate concentrations in the
San Bernardino Mountains, southern California. J. Environ. Qual. 28: 822-836.
Fenn, M. E.; Poth, M. A.; Johnson, D. W. (1996) Evidence for nitrogen saturation in the San Bernardino Mountains
in southern California. For. Ecol. Manage. 82: 211-230
Fenn, M. E.; Poth, M. A.; Aber, J. D.; Baron, J. S.; Bormann, B. T.; Johnson, D. W.; Lemly, A. D.; McNulty, S. G.;
Ryan, D. F.; Stottlemyer, R. (1998) Nitrogen excess in North American ecosystems: predisposing factors,
ecosystem responses, and management strategies. Ecol. Appl. 8: 706-733.
Fenn, M. E.; Haeuber, R.; Tonnesen, G. S.; Baron, J. S.; Grossman-Clarke, S.; Hope, D.; Jaffe, D. A.; Copeland, S.;
Geiser, L.; Rueth, H. M.; Sickman, J. O. (2003a) Nitrogen emissions, deposition, and monitoring in the
western United States. BioScience 53: 391-403.
Fenn, M. E.; Baron, J. S.; Allen, E. B.; Rueth, H. M.; Nydick, K. R.; Geiser, L.; Bowman, W. D.; Sickman, J. O.;
Meixner, T.; Johnson, D. W.; Neitlich, P. (2003b) Ecological effects of nitrogen deposition in the western
United States. BioScience 53: 404-420.
Fink, F. W.; Buttner, F. H.; Boyd, W. K. (1971) Final report on technical-economic evaluation of air-pollution
corrosion costs on metals in the U.S. Columbus, OH: Battelle Memorial Institute, Columbus Laboratories;
EPA contract no. CPA 70-86; EPA report no. APTD-0654. Available from: NTIS, Springfield, VA;
PB-198-453.
Flum, T.; Nodvin, S. C. (1995) Factors affecting streamwater chemistry in the Great Smoky Mountains, USA.
Water Air Soil Pollut. 85: 1707-1712.
Fochtman, E. G.; Langer, G. (1957) Automobile paint damaged by airborne iron particles. J. Air Pollut. Control
Assoc. 6: 243-247.
Foltescu, V. L.; Isakson, J.; Selin, E.; Stikans, M. (1994) Measured fluxes of sulphur, chlorine and some
anthropogenic metals to the Swedish west coast. Atmos. Environ. 28: 2639-2649.
Foster, J. R. (1991) Effects of organic chemicals in the atmosphere on terrestrial plants. In: Moser, T. J.; Barker,
J. R.; Tingey, D. T., eds. Ecological exposure and effects of airborne toxic chemicals: an overview. Corvallis,
OR: U.S. Environmental Protection Agency, Environmental Research Laboratory; pp. 60-89; EPA report no.
EPA-600/3-91-001. Available from: NTIS, Springfield, VA; PB91-148460.
4-248
-------
Foster, N. W.; Nicolson, J. A.; Hazlett, P. W. (1989) Temporal variation in nitrate and nutrient cations in drainage
waters from a deciduous forest. J. Environ. Qual. 18: 238-244.
Fowler, D.; Cape, J. N.; Unsworth, M. H. (1989) Deposition of atmospheric pollutants on forests. Philos. Trans. R.
Soc. London B 324: 247-265.
Fowler, D.; Duyzer, J. H.; Baldocchi, D. D. (1991) Inputs of trace gases, particles and cloud droplets to terrestrial
surfaces. Proc. R. Soc. Edinburgh Sect. B: Biol. Sci. 97: 35-59.
Fox, D. G.; Malm, W. C.; Mitchell, B.; Fisher, R. W. (1999) Where there's fire, there's smoke: fine paniculate and
regional haze. EM (November): 15-24.
Frederick, J. E.; Qu, Z.; Booth, C. R. (1998) Ultraviolet radiation at sites on the Antarctic coast. J. Photochem.
Photobiol. B 68: 183-190.
Freedman, B.; Hutchinson, T. C. (1981) Sources of metal and elemental contamination of terrestrial environments.
In: Lepp, N. W., ed. Effect of heavy metal pollution on plants. Volume 2: metals in the environment. London,
United Kingdom: Applied Science Publishers; pp. 35-94.
Freeman, A. M., III. (1993) The measurement of environmental and resource values: theory and methods.
Washington, DC: Resources for the Future; pp. 1-27.
Freedman, D. M.; Zahm, S. H.; Dosemeci, M. (1997) Residential and occupational exposure to sunlight and
mortality from non-Hodgkin's lymphoma: composite (threefold) case-control study. Br. Med. J.
314: 1451-1455.
Freudenburg, W. R.; Alario, M. (1999) What ecologists can learn from nuclear scientists. Ecosystems 2: 286-291.
Fuller, K. A.; Malm, W. C.; Kreidenweis, S. M. (1999) Effects of mixing on extinction by carbonaceous particles.
J. Geophys. Res. [Atmos.] 104: 15,941-15,954.
Ftirst, P.; Krause, G. H. M.; Hein, D.; Delschen, T.; Winners, K. (1993) PCDD/PCDF in cow's milk in relation to
their levels in grass and soil. Chemosphere 27: 1349-1357.
Gaggi, C.; Bacci, E.; Calamari, D.; Fanelli, R. (1985) Chlorinated hydrocarbons in plant foliage: an indication of the
tropospheric contamination level. Chemosphere 14: 1673-1686.
Gallagher, M. W.; Choularton, T. W.; Morse, A. P.; Fowler, D. (1988) Measurements of the size dependence of
cloud droplet deposition at a hill site. Q. J. R. Meteorol. Soc. 114: 1291-1303.
Galloway, J. N. (1998) The global nitrogen cycle: changes and consequences. Environ. Pollut. 102(suppl. 1): 15-24.
Galloway, J. N.; Cowling, E. B. (2002) Reactive nitrogen and the world: 200 years of change. Ambio 31: 64-71.
Galloway, J. N.; Aber, J. D.; Erisman, J. W.; Seitzinger, S. P.; Howarth, R. W.; Cowling, E. B.; Cosby, B. J. (2003)
The nitrogen cascade. BioScience 53: 341-356.
Garcia-Valles, M.; Vendrell-Saz, M.; Molera, J.; Blazquez, F. (1998) Interaction of rock and atmosphere: patinas on
Mediterranean monuments. Environ. Geol. (Berlin, Ger.) 36: 137-149.
Garland, J. A. (1978) Dry and wet removal of sulphur from the atmosphere. Atmos. Environ. 12: 349-362.
Garland, C.; Garland, F. (1980) Do sunlight and vitamin D reduce the likelihood of colon cancer? Int. J. Epidemiol.
9:227-231.
Garland, F. C.; Garland, C. F.; Gorham, E. D.; Young, J. F. (1990) Geographic variation in breast cancer mortality in
the United States: a hypothesis involving exposure to solar radiation. Prev. Med. 19: 614-622.
Garner, J. H. B. (1991) Ozone exposure and nitrogen loading in the forests of eastern North America. In: Berglund,
R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
Association; pp. 289-310. (A&WMA transactions series no. TR-19).
Garner, J. H. B. (1994) Nitrogen oxides, plant metabolism, and forest ecosystem response. In: Alscher, R. G.;
Wellburn, A. R., eds. Plant responses to the gaseous environment: molecular, metabolic and physiological
aspects, [3rd international symposium on air pollutants and plant metabolism]; June 1992; Blacksburg, VA.
London, United Kingdom: Chapman & Hall; pp. 301-314.
Garner, J. H. B.; Pagano, T.; Cowling, E. B. (1989) An evaluation of the role of ozone, acid deposition, and other
airborne pollutants in the forests of eastern North America. Asheville, NC: U.S. Department of Agriculture,
Forest Service, Southeastern Forest Experiment Station; general technical report SE-59.
Garten, C. T., Jr. (1988) Fate and distribution of sulfur-35 in yellow poplar and red maple trees. Oecologia
76: 43-50.
Garten, C. T., Jr.; Hanson, P. J. (1990) Foliar retention of 15N-nitrate and 15N-ammonium by red maple
(Acer rubrum) and white oak (Quercus alba) leaves from simulated rain. Environ. Exp. Bot. 30: 333-342.
Garten, C. T., Jr.; Bondietti, E. A.; Lomax, R. D. (1988) Contribution of foliar leaching and dry deposition to sulfate
in net throughfall below deciduous trees. Atmos. Environ. 22: 1425-1432.
Gawel, J. E.; Ahner, B. A.; Friedland, A. J.; Morel, F. M. M. (1996) Role for heavy metals in forest decline
indicated by phytochelatin measurements. Nature (London) 381: 64-65.
4-249
-------
Gazzi, M; Georgiadis, T.; Vicentini, V. (2001) Distant contrast measurements through fog and thick haze. Atmos.
Environ. 35: 5143-5149.
Geron, C.; Rasmussen, R.; Arnts, R. R.; Guenther, A. (2000) A review and synthesis of monoterpene speciation
from forests in the United States. Atmos. Environ. 34: 1761-1781.
Ghedini, N.; Gobbi, G.; Sabbioni, C.; Zappia, G. (2000) Determination of elemental and organic carbon on damaged
stone monuments. Atmos. Environ. 34: 4383-4391.
Gildon, A.; Tinker, P. B. (1983) Interactions of vesicular-arbuscular mycorrhizal infection and heavy metals in
plants: I. the effects of heavy metals on the development of vesicular-arbuscular mycorrhizas. New Phytol.
95: 247-261.
Gilliam, F. S.; Adams, M. B.; Yurish, B. M. (1996) Ecosystem nutrient responses to chronic nitrogen inputs at
Fernow Experimental Forest, West Virginia. Can. J. For. Res. 26: 196-205.
Gobbi, G.; Zappia, G.; Sabbioni, C. (1998) Sulphite quantification on damaged stones and mortars. Atmos. Environ.
32: 783-789.
Goodwin, J. E.; Sage, W.; Tilly, G. P. (1969) Study of erosion by solid particles. Proc. Inst. Mech. Eng.
184: 279-292.
Gorham, E. D.; Garland, C. F.; Garland, F. C. (1989) Acid haze air pollution and breast and colon cancer mortality
in 20 Canadian cities. Can. J. Public Health 80: 96-100.
Gorham, E. D.; Garland, F. C.; Garland, C. F. (1990) Sunlight and breast cancer incidence in the USSR. Int. J.
Epidemiol. 19: 820-824.
Goulder, L. H.; Kennedy, D. (1997) Valuing ecosystem services: philosophical bases and empirical methods. In:
Daily, G. C. Nature's Services: Societal Dependence on Natural Ecosystems. Washington, DC: Island Press;
pp. 23-47.
Graedel, T. E.; Nassau, K.; Franey, J. P. (1987) Copper patinas formed in the atmosphere—I. Introduction.
Corros. Sci. 27: 639-657.
Grand Canyon Visibility Transport Commission. (1996) Report of the Grand Canyon Visibility Transport
Commission to the United States Environmental Protection Agency: recommendations for improving western
vistas. Denver, CO: Grand Canyon Visibility Transport Commission.
Grant, L. D. (1989) Health effects issues associated with regional and global air pollution problems. In: Proceedings
of the changing atmosphere: implications for global security; June, 1988; Toronto, Canada. Geneva,
Switzerland: World Meteorological Organization; report no. WMO-No. 710.
Grant, W. B. (2002) An estimate of premature cancer mortality in the U.S. due to inadequate doses of solar
ultraviolet-B radiation. Cancer 6: 1867-1875.
Grattan, S. R.; Maas, E. V.; Ogata, G. (1981) Foliar uptake and injury from saline aerosol. J. Environ. Qual.
10: 406-409.
Gravenhorst, G.; Hoefken, K. D.; Georgii, H. W. (1983) Acidic input to a beech and spruce forest. In: Beilke, S.;
Elshout, A. J., eds. Acid deposition: proceedings of the CEC workshop organized as part of the concerted
action "Physico-chemical behaviour of atmospheric pollutants;" September 1982; Berlin, Federal Republic of
Germany. Dordrecht, The Netherlands: D. Reidel Publishing Company; pp. 155-171.
Greenberg, B. M.; Wilson, M. L; Huang, X.-D.; Duxbury, C. L.; Gerhardt, K. E.; Gensemer, R. W. (1997)
The effects of ultraviolet-B radiation on higher plants. In: Wang, W.; Gorsuch, J. W.; Hughes, J. S., eds.
Plants for Environmental Studies. New York, NY: Lewis Publishers; pp. 1-35.
Grennfelt, P. (1987) Deposition processes for acidifying compounds. Environ. Technol. Lett. 8: 515-527.
Grimm, N. B.; Grove, J. M.; Pickett, S. T. A.; Redman, C. L. (2000) Integrated approaches to long-term studies of
urban ecological systems. BioScience 50: 571-584.
Gu, L. H.; Baldocchi, D.; Verma, S. B.; Black, T. A.; Vesala, T.; Falge, E. M.; Dowty, P. R. (2002) Advantages of
diffuse radiation for terrestrial ecosystem productivity. J. Geophys. Res. [Atmos.] 107(D6):
10.1029/2001JDOO1242.
Guderian, R. (1977) Accumulation of pollutants in plant organs. In: Air pollution: phytotoxicity of acidic gases and
its significance in air pollution control. Berlin, Germany: Springer-Verlag; pp. 66-74. (Ecological studies:
v. 22).
Guderian, R., ed. (1985) Air pollution by photochemical oxidants: formation, transport, control, and effects on
plants. New York, NY: Springer-Verlag. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
Remmert, H., eds. Ecological studies: analysis and synthesis: v. 52).
Guderian, R. (1986) Terrestrial ecosystems: paniculate deposition. In: Legge, A. H.; Krupa, S. V., eds.
Air pollutants and their effects on the terrestrial ecosystem. New York, NY: John Wiley & Sons; pp. 339-363.
(Advances in environmental science and technology: v. 18).
4-250
-------
Guderian, R.; Tingey, D. T.; Rabe, R. (1985) Effects of photochemical oxidants on plants. In: Guderian, R., ed. Air
pollution by photochemical oxidants: formation, transport, control, and effects on plants. Berlin, Federal
Republic of Germany: Springer-Verlag; pp. 127-333. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
Remmert, H., eds. Ecological studies: analysis and synthesis, v. 52).
Gundersen, P.; Callesen, L; De Vries, W. (1998a) Nitrate leaching in forest ecosystems is related to forest floor C/N
ratios. Environ. Pollut. 102(suppl. 1): 403-407.
Gundersen, P.; Emmett, B. A.; Kjonaas, O. J.; Koopmans, C. J.; Tietema, A. (1998b) Impact of nitrogen deposition
on nitrogen cycling in forests: a synthesis of NITREX data. For. Ecol. Manage. 101: 37-55.
Hamilton, R. S.; Revitt, D. M; Vincent, K. J.; Butlin, R. N. (1995) Sulphur and nitrogen paniculate pollutant
deposition on to building surfaces. In: Saiz-Jimenez, C., ed. The deterioration of monuments: proceedings of
the 2nd international symposium on biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci.
Total Environ. 167: 57-66.
Hanchette, C. L.; Schwartz, G. G. (1992) Geographic patterns of prostate cancer mortality. Cancer 70: 2861-2869.
Haneef, S. J.; Johnson, J. B.; Thompson, G. E.; Wood, G. C. (1993) Simulation of the degradation of limestones and
dolomitic sandstone under dry deposition conditions. In: Corrosion control for low-cost reliability:
preceedings [of the] 12th international corrosion congress. Volume 2: process industries plant operation.
Houston, TX: National Association of Corrosion Engineers; pp. 700-710.
Hanley, N.; Spash, C. L. (1993) Cost-benefit analysis and the environment. Hants, United Kingdom: Edward Elgar.
Hansen, K.; Draaijers, G. P. J.; Ivens, W. P. M. F.; Gundersen, P.; vanLeeuwen, N. F. M. (1994) Concentration
variations in rain and canopy throughfall collected sequentially during individual rain events. Atmos. Environ.
28: 3195-3205.
Hansen, J. E.; Sato, M.; Ruedy, R. (1997) Radiative forcing and climate response. J. Geophys. Res. 102: 6831-6864.
Hansen, J.; Sato, M.; Ruedy, R.; Lacis, A.; Oinas, V. (2000) Global warming in the twenty-first century: an
alternative scenario. Proc. Natl. Acad. Sci. U. S. A. 97: 9875-9880.
Hanson, P. J.; Lindberg, S. E. (1991) Dry deposition of reactive nitrogen compounds: a review of leaf, canopy and
non-foliar measurements. Atmos. Environ. Part A 25: 1615-1634.
Hartge, P.; Devessa, S. S.; Graumen, D.; Fears, T. R.; Fraumeni, J. F. (1996) Non-Hodgkin's lymphoma and
sunlight. J. Natl. Cancer Inst. 88: 298-300.
Harwell, M. A.; Myers, V; Young, T.; Bartuska, A.; Gassman, N.; Gentile, J. H.; Harwell, C. C.; Appelbaum, S.;
Barko, J.; Causey, B.; Johnson, C.; McLean, A.; Smola, R.; Templet, P.; Tosini, S. (1999) A framework for
an ecosystem integrity report card. BioScience 49: 543-556.
Hauhs, M. (1989) Lange Bramke: an ecosystem study of a forested catchment. In: Adriano, D. C.; Havas, M., eds.
Acidic precipitation: volume 1, case studies. New York, NY: Springer-Verlag; pp. 275-305. (Adriano, D. C.;
Salomons, W., eds. Advances in environmental science).
Haynie, F. H. (1976) Air pollution effects on stress induced intergranular corrosion of 7005-T53 aluminum alloy. In:
Stress corrosion-new approaches. Philadelphia, PA: American Society for Testing and Materials; pp. 32-43.
(ASTM special technical publication 610).
Haynie, F. H. (1980) Theoretical air pollution and climate effects on materials confirmed by zinc corrosion data.
In: Sereda, P. J.; Litvan, G. G., eds. Durability of building materials and components: proceedings of the
1st international conference; August 1978; Ottawa, ON, Canada. Philadelphia, PA: American Society for
Testing and Materials; pp. 157-175. (ASTM special technical publication 691).
Haynie, F. H. (1986) Environmental factors affecting corrosion of weathering steel. In: Baboian, R., ed. Materials
degradation caused by acid rain: developed from the 20th state-of-the-art symposium of the American
Chemical Society; June 1985; Arlington, VA. Washington, DC: American Chemical Society; pp. 163-171.
(ACS symposium series 318).
Haynie, F. H.; Lemmons, T. J. (1990) Paniculate matter soiling of exterior paints at a rural site. Aerosol Sci.
Technol. 13: 356-367.
4-251
-------
Haynie, F. H.; Upham, J. B. (1974) Correlation between corrosion behavior of steel and atmospheric pollution data.
In: Coburn, S. K., ed. Corrosion in natural environments: presented at the 76th annual meeting American
Society for Testing and Materials; June 1973; Philadelphia, PA. Philadelphia, PA: American Society for
Testing and Materials; pp. 33-51. (ASTM special technical publication 558).
Haywood, J. M.; Shine, K. P. (1995) The effect of anthropogenic sulfate and soot aerosol on the clear sky planetary
radiation budget. Geophys. Res. Lett. 22: 603-606.
Haywood, J. M.; Ramaswamy, V.; Soden, B. J. (1999) Tropospheric aerosol climate forcing in clear-sky satellite
observations over the oceans. Science (Washington, DC) 283: 1299-1303.
Heal, G. (2000) Valuing ecosystem services. Ecosystems 3: 24-30.
Healey, K. D.; Rickert, K. G.; Hammer, G. L.; Bange, M. P. (1998) Radiation use efficiency increases when the
diffuse component of incident radiation is enhanced under shade. Aust. J. Agric. Res. 49: 665-672.
Heath, J. A.; Huebert, B. J. (1999) Cloudwater deposition as a source of fixed nitrogen in a Hawaiian montane
forest. Biogeochemistry 44: 119-134.
Hedin, L. O.; Granat, L.; Likens, G. E.; Buishand, T. A.; Galloway, J. N.; Butler, T. J.; Rodhe, H. (1994)
Steep declines in atmospheric base cations in regions of Europe and North America. Nature (London)
367: 351-354.
Heichel, G. H.; Hankin, L. (1972) Particles containing lead, chlorine, and bromine detected on trees with an electron
microprobe. Environ. Sci. Technol. 6: 1121-1122.
Heichel, G. H.; Hankin, L. (1976) Roadside coniferous windbreaks as sinks for vehicular lead emissions. J. Air
Pollut. Control Assoc. 26: 767-770.
Heinsdorf, D. (1993) The role of nitrogen in declining Scots pine forests (Pinus sylvestris) in the lowland of
East Germany. Water Air Soil Pollut. 69: 21-35.
Hemming, B. L.; Seinfeld, J. H. (2001) On the hygroscopic behavior of atmospheric organic aerosols. Ind. Eng.
Chem. Res. 40: 4162-4171.
Hermosin, B.; Saiz-Jimenez, C. (2000) Characterization of the organic compounds deposited on monument surfaces.
Fresenius' Environ. Bull. 9: 322-327.
Hess, A.; Iyer, H.; Malm, W. (2001) Linear trend analysis: a comparison of methods. Atmos. Environ.
35: 5211-5222.
Hicks, B. B. (1986) Differences in wet and dry particle deposition parameters between North America and Europe.
In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk assessment and control
strategies; proceedings of the second U.S.-Dutch international symposium; May 1985; Williamsburg, VA.
Chelsea, MI: Lewis Publishers; pp. 973-982.
Hicks, B. B.; Matt, D. R.; McMillen, R. T.; Womack, J. D.; Shelter, R. E. (1984) Eddy fluxes of nitrogen oxides to a
deciduous forest in complex terrain. In: Samson, P. J., ed. Transactions: the meteorology of acid deposition,
an APCA specialty conference; October 1983; Hartford, CT. Pittsburgh, PA: Air Pollution Control
Association; pp. 189-201.
Hicks, B. B.; Baldocchi, D. D.; Meyers, T. P.; Hosker, R. P., Jr.; Matt, D. R. (1987) A preliminary multiple
resistance routine for deriving dry deposition velocities from measured quantities. Water Air Soil Pollut.
36:311-330.
Hicks, B. B.; Matt, D. R.; McMillen, R. T.; Womack, J. D.; Wesely, M. L.; Hart, R. L.; Cook, D. R.; Lindberg,
S. E.; de Pena, R. G.; Thomson, D. W. (1989) A field investigation of sulfate fluxes to a deciduous forest.
J. Geophys. Res. [Atmos.] 94: 13,003-13,011.
Hidy, G. M.; Hales, J. M.; Roth, P. M.; Scheffe, R. (2000) Fine particles and oxidant pollution: developing an
agenda for cooperative research. J. Air Waste Manage. Assoc. 50: 613-632.
Hinckley, T.; Ford, D.; Segura, G.; Sprugel, D. (1992) Key processes from tree to stand level. In: Wall, G., ed.
Implications of climate change for Pacific Northwest forest management. Waterloo, ON, Canada: University
of Waterloo, Department of Geography; pp. 33-43. (Department of Geography publication series: occasional
paper no. 15).
Hofken, K. D.; Gravenhorst, G. (1982) Deposition of atmospheric aerosol particles to beech- and spruce forest. In:
Georgii, H.-W.; Pankrath, J., eds. Deposition of atmospheric pollutants: proceedings of a colloquium;
November 1981; Oberursel/Taunus, Federal Republic of Germany. Dordrecht, The Netherlands: D. Reidel
Publishing Company; pp. 191-194.
Hofken, K. D.; Meixner, F. X.; Ehhalt, D. H. (1983) Deposition of atmospheric trace constituents onto different
natural surfaces. In: Pruppacher, H. R.; Semonin, R. G.; Slinn, W. G. N., eds. Precipitation scavenging, dry
deposition, and resuspension: v. 2, dry deposition and resuspension: proceedings of the fourth international
conference; November-December 1982; Santa Monica, CA. New York, NY: Elsevier; pp. 825-835.
4-252
-------
Hofstra, G.; Hall, R. (1971) Injury on roadside trees: leaf injury on pine and white cedar in relation to foliar levels of
sodium and chloride. Can. J. Bot. 49: 613-622.
Hogan, G. D.; Rennenberg, H.; Fink, S. (1998) Role and effect of sulfur in tree biology. In: Maynard, D. G., ed.
Sulfur in the environment. New York, NY: Marcel Dekker, Inc.; pp. 173-217. (Books in soils, plants, and the
environment: v. 67).
Holbrow, G. L. (1962) Atmospheric pollution: its measurement and some effects on paint. J. Oil Colour Chem.
Assoc. 45: 701-718.
Holland, D. M; Principe, P. P.; Sickles, J. E., II. (1999) Trends in atmospheric sulfur and nitrogen species in the
eastern United States for 1989-1995. Atmos. Environ. 33: 37-49.
Hollander, M.; Wolfe, D. A. (1973) Nonparametric statistical methods. New York: John Wiley & Sons.
Holling, C. S. (1986) The resilience of terrestrial ecosystems: local surprise and global change. In: Clark, W. C.;
Munn, R. E., eds. Sustainable development of the biosphere. Cambridge, UK: Cambridge University;
pp. 292-317.
Hooper, D. U.; Vitousek, P. M. (1997) The effects of plant composition and diversity on ecosystem processes.
Science 277: 1302-1305.
Hornung, M.; Langan, S. J. (1999) Nitrogen deposition, sources, impacts and responses in natural and semi-natural
ecosystems. In: Bangan, S. J., ed. Impact of nitrogen deposition on natural ecosystems and semi-natural
ecosystems. Dordrect, Netherlands: Kluwer Academic Publishers; pp. 1-14. [Environmental Pollution, no.3].
Horvath, H. (1993) Atmospheric light absorption-a review. Atmos. Environ. Part A 27: 293-317.
Hosker, R. P., Jr.; Lindberg, S. E. (1982) Review: atmospheric deposition and plant assimilation of gases and
particles. Atmos. Environ. 16: 889-910.
Houghton, J. T.; Jenkins, G. J.; Ephraums, J. J., eds. (1990) Climate change: the IPCC scientific assessment.
Cambridge, MA: Cambridge University Press; p. 55.
Howarth, R. W.; Boyer, E. W.; Pabich, W. J.; Galloway, J. N. (2002) Nitrogen use in the United States from
1961-2000 and potential future trends. Ambio 31: 88-96.
Hoyt, D. V. (1978) A model for the calculation of solar global insolation. Sol. Energy 21: 27-35.
Huebert, B. J.; Luke, W. T.; Delany, A. C.; Brost, R. A. (1988) Measurements of concentrations and dry surface
fluxes of atmospheric nitrates in the presence of ammonia. J. Geophys. Res. [Atmos.] 93: 7127-7136.
Hughes, M. K. (1981) Cycling of trace metals in ecosystems. In: Lepp, N. W., ed. Effect of heavy metal pollution on
plants. Volume 2: metals in the environment. London, United Kingdom: Applied Science Publishers;
pp. 95-118.
Humphreys, F. R.; Lambert, M. J.; Kelly, J. (1975) The occurrence of sulphur deficiency in forests. In: McLachlan,
K. D., ed. Sulphur in Australasian agriculture. Sydney, Australia: Sydney University Press; pp. 154-162.
Hunsaker, C.; Graham, R.; Turner, R. S.; Ringold, P. L.; Holdren, G. R., Jr.; Strickland, T. C. (1993) A national
critical loads framework for atmospheric deposition effects assessment: II. Defining assessment end points,
indicators, and functional subregions. Environ. Manage. (N. Y.) 17: 335-341.
Huntington, T.G. (2000) The potential for calcium depletion in forest ecosystems of southeastern United States:
review and analysis. Global Biogeochem. Cycles 14: 623-638.
Husar, R. B.; Wilson, W. E. (1993) Haze and sulfur emission trends in the eastern United States. Environ. Sci.
Technol. 27: 12-16.
Husar, R. B.; Holloway, J. M.; Patterson, D. E.; Wilson, W. E. (1981) Spatial and temporal pattern of eastern U.S.
haziness: a summary. Atmos. Environ. 15: 1919-1928.
Husar, R. B.; Elkins, J. B.; Wilson, W. E. (1994) U.S. visibility trends, 1960-1992, regional and national. Presented
at: 87th annual meeting of the Air & Waste Management Association; June; Cincinnati, OH. Pittsburgh, PA:
Air & Waste Management Association; paper no. 94-MP3.05.
Hutchinson, A. J.; Johnson, J. B.; Thompson, G. E.; Wood, G. C.; Sage, P. W.; Cooke, M. J. (1992) The role of
fly-ash paniculate material and oxide catalysts in stone degradation. Atmos. Environ. Part A 26: 2795-2803.
Huttunen, S. (1994) Effects of air pollutants on epicuticular wax structure. In: Percy, K. E.; Cape, J. N.; Jagels, R.;
Simpson, C. J., eds. Air pollutants and the leaf cuticle. New York, NY: Springer-Verlag; pp. 81-96.
(NATO ASI series).
Hyer, K. E.; Webb, J. R.; Eshleman, K. N. (1995) Episodic acidification of three streams in Shenandoah National
Park, Virginia, USA. Water Air Soil Pollut. 85: 523-528.
Ibrahim, M.; Barrie, L. A.; Fanaki, F. (1983) An experimental and theoretical investigation of the dry deposition of
particles to snow, pine trees and artificial collectors. Atmos. Environ. 17: 781-788.
4-253
-------
IMPROVE: interagency monitoring of protected visual environments [database]. (1998) [Data on gross chemical
composition of PM2.5 particles in rural areas of the eastern and western United States as of 1998]. Fort
Collins, CO: National Park Service (NPS); Cooperative Institute for Research in the Atmosphere (CIRA).
Available at: http://vista.cira.colostate.edu/improve/ [2001, January 26].
Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
of vulnerability. Cambridge, United Kingdom: Cambridge University Press.
Intergovernmental Panel on Climate Change (IPCC). (2001a) Climate change 2001: the scientific basis. Contribution
of working group I to the third assessment report of the Intergovernmental Panel on Climate Change.
Cambridge, United Kingdom: Cambridge University Press.
Intergovernmental Panel on Climate Change (IPCC). (200 Ib) Climate change 2001: impacts, adaptation, and
vulnerability. Contribution of working group II to the third assessment report of the Intergovernmental Panel
on Climate Change. Cambridge, United Kingdom: Cambridge University Press.
Iyer, H.; Patterson, P.; Malm, W. C. (2000) Trends in the extremes of sulfur concentration distributions. J. Air Waste
Manage. Assoc. 50: 802-808.
Jacobson, M. Z. (1998) Studying the effects of aerosols on vertical photolysis rate coefficient and temperature
profiles over an urban airshed. J. Geophys. Res. [Atmos.] 103: 10,593-10,604.
Jacobson, M. Z. (1999) Isolating nitrated and aromatic aerosols and nitrated aromatic gases as sources of ultraviolet
light absorption. J. Geophys. Res. 104 (D3): 3527-3542.
Jacobson, M. Z. (2000) A physically-based treatment of elemental carbon optics: implications for global direct
forcing of aerosols. Geophys. Res. Lett. 27: 217-220.
Jaworski, N. A.; Howarth, R. W.; Hetling, L. J. (1997) Atmospheric deposition of nitrogen oxides onto the landscape
contributes to coastal eutrophication in the northeast United States. Environ. Sci. Technol. 31: 1995-2004.
Jensen, V. (1974) Decomposition of angiosperm tree leaf litter. In: Dickinson, C. H.; Pugh, G. J. F., eds. Biology of
plant litter decomposition: v. I. London, United Kingdom: Academic Press; pp. 69-104.
John, W.; Wall, S. M.; Ondo, J. L. (1985) Dry acid deposition on materials and vegetation: concentrations in
ambient air. Berkeley, CA: California Department of Health Services, Air and Industrial Hygiene Laboratory;
report no. CA/DOH/AIHL/SP-34. Available from: NTIS, Springfield, VA; PB85-241206.
Johnson, D. W. (1984) Sulfur cycling in forests. Biogeochemistry 1: 29-43.
Johnson, D. W. (1992) Base cation distribution and cycling. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric
deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
Springer-Verlag, Inc.; pp. 275-333. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
Ecological studies, analysis and synthesis: v. 91).
Johnson, F. R.; Desvousges, W. H. (1997) Estimating stated preferences with rated-pair data: environmental, health,
and employment effects of energy programs. J. Environ. Econ. Manage. 34: 79-99.
Johnson, D. W.; Lindberg, S. E., eds. (1992a) Atmospheric deposition and forest nutrient cycling: a synthesis of the
integrated forest study. New York, NY: Springer-Verlag, Inc. (Billings, W. D.; Golley, F.; Lange, O. L.;
Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
Johnson, D. W.; Lindberg, S. E., eds. (1992b) Nitrogen chemistry, deposition, and cycling in forests. In: Johnson,
D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated
forest study. New York, NY: Springer-Verlag, Inc.; pp. 150-213. (Billings, W. D.; Golley, F.; Lange, O. L.;
Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
Johnson, D. W.; Mitchell, M. J. (1998) Responses of forest ecosystems to changing sulfur inputs. In: Maynard,
D. G., ed. Sulfur in the environment. New York, NY: Marcel Dekker, Inc.; pp. 219-262. [Books in Soils,
Plants, and the Environment, v. 67].
Johnson, D. W.; Taylor, G. E. (1989) Role of air pollution in forest decline in eastern North America. Water Air Soil
Pollut. 48: 21-43.
Johnson, D. W.; Todd, D. E., Jr. (1998) Harvesting effects on long-term changes in nutrient pools of mixed oak
forest. Soil Sci. Soc. Amer. J. 62: 1725-1735.
Johnson, D. W.; Henderson, G. S.; Huff, D. D.; Lindberg, S. E.; Richter, D. D.; Shriner, D. S.; Todd, D. E.;
Turner, J. (1982a) Cycling of organic and inorganic sulphur in a chestnut oak forest. Oecologia 54: 141-148.
Johnson, D. W.; Cole, D. W.; Bledsoe, C. S.; Cromack, K.; Edmonds, R. L.; Gessel, S. P.; Grier, C. C.; Richards,
B. N.; Vogt, K. A. (1982b) Nutrient cycling in forests of the Pacific Northwest. In: Edmonds, R. L., ed.
Analysis of coniferous forest ecosystems in the Western United States. Stroudsburg, PA: Hutchinson Ross
Publishing Company; pp. 186-232. (US/IBP synthesis series: v. 14).
Johnson, D. W.; VanMiegroet, H.; Lindberg, S. E.; Todd, D. E.; Harrison, R. B. (1991a) Nutrient cycling in red
spruce forests of the Great Smoky Mountains. Can. J. For. Res. 21: 769-787.
4-254
-------
Johnson, D. W.; Cresser, M. S.; Nilsson, S. I.; Turner, I; Ulrich, B.; Binkley, D.; Cole, D. W. (1991b) Soil changes
in forest ecosystems: evidence for and probable causes. Proc. R. Soc. Edinburgh Sect. B: Biol. Sci.
97B: 81-116.
Johnson, D. W.; Swank, W. T.; Vose, J. M. (1993) Simulated effects of atmospheric sulfur deposition on nutrient
cycling in a mixed deciduous forest. Biogeochemistry 23: 169-196.
Johnson, A. H.; Friedland, A. J.; Miller, E. K.; Siccama, T. G. (1994) Acid rain and soils of the Adirondacks. III.
Rates of soil acidification in a montane spruce-fir forest at Whiteface Mountain, New York. Can. J. For. Res.
24: 663-669.
Johnson, D.W.; Susfalk, R.B.; Brewer, P.,F.; Swank, W.T. (1999) Simulated effects of reduced sulfur, nitrogen, and
base cation deposition on soils and solutions in southern Appalachian forests. J. Environ. Qua! 28: 1336-1346.
Joslin, J. D.; Wolfe, M. H. (1992) Red spruce soil chemistry and root distribution across a cloud water deposition
gradient. Can. J. For. Res. 22: 893-904.
Joslin, J. D.; Wolfe, M. H. (1994) Foliar deficiencies of mature Southern Appalachian red spruce determined from
fertilizer trials. Soil Sci. Soc. Am. J. 58: 1572-1579.
Joslin, J. D.; Kelly, J. M.; VanMiegroet, H. (1992) Soil chemistry and nutrition of North American spruce-fir
stands: evidence for recent change. J. Environ. Qual. 21: 12-30.
Justus, C. G.; Murphey, B. B. (1994) Temporal trends in surface irradiance at ultraviolet wavelengths. J. Geophys.
Res. 99 (Dl): 1389-1394.
Kaminski, M. D.; Landsberger, S. (2000a) Heavy metals in urban soils of East St. Louis, IL, part I: total
concentration of heavy metals in soils. J. Air Waste Manage. Assoc. 50: 1667-1679.
Kaminski, M. D.; Landsberger, S. (2000b) Heavy metals in urban soils of East St. Louis, IL, part II: leaching
characteristics and modeling. J. Air Waste Manage. Assoc. 50: 1680-1687.
Kandeler, E.; Kampichler, C.; Horak, O. (1996) Influence of heavy metals on the functional diversity of soil
microbial communities. Biol. Fertil. Soils 23: 299-306.
Karmoker, J. L.; Clarkson, D. T.; Saker, L. R.; Rooney, J. M.; Purves, J. V. (1991) Sulphate deprivation depresses
the transport of nitrogen to the xylem and the hydraulic conductivity of barley (Hordeum vulgare L) roots.
Planta 185: 269-278.
Kaupp, H.; McLachlan, M. S. (1999) Atmospheric particle size distributions of polychlorinated dibenzo-p-dioxins
and dibenzofurans (PCDD/Fs) and polycyclic aromatic hydrocarbons (PAHs) and their implications for wet
and dry deposition. Atmos. Environ. 33: 85-95.
Kelly, J.; Lambert, M. J. (1972) The relationship between sulphur and nitrogen in the foliage ofPinus radiata.
Plant Soil 37: 395-407.
Kelly, J. M.; Meagher, J. F. (1986) Nitrogen input/output relationships for three forested watersheds in eastern
Tennessee. In: Correll, D. L., ed. Watershed research perspectives. Washington, DC: Smithsonian Institution
Press; pp. 360-391.
Kenk, G.; Fischer, H. (1988) Evidence from nitrogen fertilisation in the forests of Germany. Environ. Pollut.
54: 199-218.
Kiehl, J. T.; Briegleb, B. P. (1993) The relative roles of sulfate aerosols and greenhouse gases in climate forcing.
Science (Washington, DC) 260: 311-314.
Kim, E.; Kalman, D.; Larson, T. (2000) Dry deposition of large airborne particles on a surrogate surface. Atmos.
Environ. 34: 2387-2397.
Kleeman, M. J.; Eldering, A.; Hall, J. R.; Cass, G. R. (2001) Effect of emissions control programs on visibility in
southern California. Environ. Sci. Technol. 35: 4668-4674.
Klepper, B.; Craig, D. K. (1975) Deposition of airborne particulates onto plant leaves. J. Environ. Qual. 4: 495-499.
Kodama, Y.; Lee, S. D., eds. (1994) The 13th UOEH international symposium and the 2nd pan Pacific cooperative
symposium on impact of increased UV-B exposure on human health and ecosystem; October 1993;
Kitakyushu, Japan. Kitakyushu, Japan: University of Occupational and Environmental Health.
Koslow, E. E.; Smith, W. H.; Staskawicz, B. J. (1977) Lead-containing particles on urban leaf surfaces.
Environ. Sci. Technol. 11: 1019-1021.
Krajickova, A.; Mejstfik, V. (1984) The effect of fly-ash particles on the plugging of stomata. Environ. Pollut.
36: 83-93.
Kripke, M. L. (1989) Health effects of stratospheric ozone depletion: an overview. In: Schneider, T.; Lee, S. D.;
Wolters, G. J. R.; Grant, L. D. eds. Atmospheric ozone research and its policy implications: proceedings of
the Third US-Dutch International Symposium; May, 1988; Nijmegen, The Netherlands. Amsterdam, The
Netherlands: Elsevier, pp. 795-802. [Studies in Environmental Science 35].
Krupa, S. V.; Legge, A. H. (1986) Single or joint effects of coarse and fine paniculate sulfur aerosols and ozone on
vegetation. In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk assessment
4-255
-------
and control strategies, proceedings of the second U.S.-Dutch international symposium; May 1985;
Williamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 879-887.
Krupa, S. V.; Legge, A. H. (1998) Sulphur dioxide, paniculate sulphur and its impacts on a boreal forest ecosystem.
In: Ambasht, R. S., ed. Modern trends in ecology and environment. Leiden, The Netherlands: Backhuys
Publishers; pp. 285-306.
Kylin, H.; Grimvall, E.; Ostman, C. (1994) Environmental monitoring of polychlorinated biphenyls using pine
needles as passive samplers. Environ. Sci. Technol. 28: 1320-1324.
Lamaud, E.; Chapuis, A.; Fontan, J.; Serie, E. (1994) Measurements and parameterization of aerosol dry deposition
in a semi-arid area. Atmos. Environ. 28: 2461-2471.
Lamersdorf, N. P.; Meyer, M. (1993) Nutrient cycling and acidification of a northwest German forest site with high
atmospheric nitrogen deposition. For. Ecol. Manage. 62: 323-354.
Laurenroth, W. K.; Milchunas, D. G. (1985) SO2 effects on plant community function. In: Winner, W. E.; Mooney,
H. A.; Goldstein, R. A., eds. Sulfur dioxide and vegetation: physiology, ecology, and policy issues. Stanford,
CA: Stanford University Press; pp. 454-477.
Lawrence, G. B.; Huntington, T. G. (1999) Soil-Calcium depletion linked to acid rain and forest growth in the
eastern United States. Reston, VA: U.S. Geological Survey. Available:
http://bqs.usgs.gov/acidrain/WRIR984267.pdf [29 September, 2004].
Lefkowitz, E. S.; Garland, C. F. (1994) Sunlight, vitamin D, and ovarian cancer mortality rates in U.S. women. Int.
J. Epidemiol. 23: 1133-1136.
Legge, A. H. (1980) Primary productivity, sulfur dioxide, and the forest ecosystem: an overview of a case study.
In: Miller, P. R., ed. Proceedings of symposium on effects of air pollutants on Mediterranean and temperate
forest ecosystems; June; Riverside, CA. Berkeley, CA: U.S. Department of Agriculture, Pacific Southwest
Forest and Range Experiment Station; pp. 51-62; Forest Service general technical report no. PSW-43.
Available from: NTIS, Springfield, VA; PB81-133720.
Lerman, S. L.; Darley, E. F. (1975) Particulates. In: Mudd, J. B.; Kozlowski, T. T., eds. Responses of plants to air
pollution. New York, NY: Academic Press; pp. 141-158.
Leslie, M. (2001) Having abad air day. Science (Washington, DC) 294: 2255.
Levin, S. A. (1998) Ecosystems and the biosphere as complex adaptive systems. Ecosystems 1: 431-436.
Li-Jones, X.; Prospero, J. M. (1998) Variations in the size distribution of non-sea-salt sulfate aerosol in the marine
boundary layer at Barbados: impact of African dust. J. Geophys. Res. [Atmos.] 103: 16,073-16,084.
Likens, G. E.; Driscoll, C. T.; Buso, D. C. (1996) Long-term effects of acid rain: response and recovery of a forest
ecosystem. Science (Washington, DC) 272: 244-246.
Likens, G. E.; Driscoll, C. T.; Buso, D. C.; Siccama, T. G.; Johnson, C. E.; Lovett, G. M.; Fahey, T. J.; Reiners,
W. A.; Ryan, D. F.; Martin, C. W.; Bailey, S. W. (1998) The biogeochemistry of calcium at Hubbard Brook.
Biogeochemistry 41: 89-173.
Likens, G. E.; Butler, T. J.; Buso, D. C. (2001) Long- and short-term changes in sulfate deposition: effects of the
1990 Clean Air Act Amendments. Biogeochemistry 52: 1-11.
Lin, J.-M.; Fang, G.-C.; Holsen, T. M.; Noll, K. E. (1993) A comparison of dry deposition modeled from size
distribution data and measured with a smooth surface for total particle mass, lead and calcium in Chicago.
Atmos. Environ. Part A 27: 1131-1138.
Lindberg, S. E. (1992) Atmospheric deposition and canopy interactions of sulfur. In: Johnson, D. W.; Lindberg,
S. E., eds. Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated forest study.
New York, NY: Springer-Verlag, Inc.; pp. 74-90. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
Lindberg, S. E.; Garten, C. T., Jr. (1988) Sources of sulphur in forest canopy throughfall. Nature (London)
336: 148-151.
Lindberg, S. E.; Harriss, R. C. (1981) The role of atmospheric deposition in an eastern U.S. deciduous forest.
Water Air Soil Pollut. 16: 13-31.
Lindberg, S. E.; Lovett, G. M. (1985) Field measurements of particle dry deposition rates to foliage and inert
surfaces in a forest canopy. Environ. Sci. Technol. 19: 238-244.
Lindberg, S. E.; Lovett, G. M. (1992) Deposition and forest canopy interactions of airborne sulfur: results from the
integrated forest study. Atmos. Environ. Part A 26: 1477-1492.
4-256
-------
Lindberg, S. E.; McLaughlin, S. B. (1986) Air pollutant interactions with vegetation: research needs in data
acquisition and interpretation. In: Legge, A. H.; Krupa, S. V., eds. Air pollutants and their effects on the
terrestrial ecosystem. New York, NY: John Wiley & Sons, Inc.; pp. 449-503. (Nriagu, J. O., ed. Advances in
environmental science and technology: v. 18).
Lindberg, S. E.; Owens, J. G. (1993) Throughfall studies of deposition to forest edges and gaps in montane
ecosystems. Biogeochemistry 19: 173-194.
Lindberg, S. E.; Harriss, R. C.; Turner, R. R. (1982) Atmospheric deposition of metals to forest vegetation. Science
(Washington, DC) 215: 1609-1611.
Lindberg, S. E.; Lovett, G. M; Richter, D. D.; Johnson, D. W. (1986) Atmospheric deposition and canopy
interactions of major ions in a forest. Science (Washington, DC) 231: 141-145.
Lindberg, S. E.; Lovett, G. M.; Meiwes, K.-J. (1987) Deposition and forest canopy interactions of airborne nitrate.
In: Hutchinson, T. C.; Meema, K. M., eds. Effects of atmospheric pollutants on forests, wetlands and
agricultural ecosystems. Berlin, Federal Republic of Germany: Springer-Verlag; pp. 117-130. (NATO as
1 series: v. G16).
Lindberg, S. E.; Lovett, G. M.; Schaefer, D. A.; Bredemeier, M. (1988) Dry deposition velocities and
surface-to-canopy scaling factors for aerosol calcium from forest canopy throughfall. J. Aerosol Sci.
19: 1187-1190.
Lindberg, S. E.; Bredemeier, M.; Schaefer, D. A.; Qi, L. (1990) Atmospheric concentrations and deposition of
nitrogen and major ions in conifer forests in the United States and Federal Republic of Germany. Atmos.
Environ. Part A 24: 2207-2220.
Liousse, C.; Cachier, H.; Jennings, S. G. (1993) Optical and thermal measurements of black carbon aerosol content
in different environments: variation of the specific attenuation cross-section, sigma (a). Atmos. Environ.
Part A 27: 1203-1211.
Little, P. (1977) Deposition of 2.75, 5.0 and 8.5 |_im particles to plant and soil surfaces. Environ. Pollut. 12: 293-305.
Little, P.; Wiffen, R. D. (1977) Emission and deposition of petrol engine exhaust Pb~I. deposition of exhaust Pb to
plant and soil surfaces. Atmos. Environ. 11: 437-447.
Liu, S. C.; McKeen, S. A.; Madronich, S. (1991) Effect of anthropogenic aerosols on biologically active ultraviolet
radiation. Geophys. Res. Lett. 18: 2265-2268.
Lloyd, J. (2002) After years of smog, a shroud lifts. The Christian Science Monitor. June 4: l(s. 1).
Loehman, E.; Boldt, D.; Chaikin,K. (1985) Measuring the benefits of air quality improvements in the SanFrancisco
Bay area: property value and contingent valuation studies, volume IV of methods development for
environmental control benefits assessment. Washington, DC: U.S. Environmental Protection Agency; Office
of Policy, Planning and Evaluation; report no. EPA-230-12-85-022. Available:
http://yosemite.epa.gov/EE/epa/eerm.nsf/vwRepNumLookup/EE-0278D7OpenDocument (28 May 2003).
Lekke, H.; Bak, J.; Falkengren-Grerup, U.; Finlay, R. D.; Ilvesniemi, H.; Nygaard, P. H.; Starr, M. (1996)
Critical loads of acidic deposition for forest soils: is the current approach adequate. Ambio 25: 510-516.
Lorente, J.; Redano, A.; de Cabo, X. (1994) Influence of urban aerosol on spectral solar irradiance. J. Appl.
Meteorol. 33:406-415.
Lorusso, S.; Marabelli, M.; Troili, M. (1997) Air pollution and the deterioration of historic monuments. J. Environ.
Pathol. Toxicol. Oncol. 16: 171-173.
Lovett, G. M. (1984) Atmospheric deposition to forests. In: Breece, L.; Hasbrouck, S., eds. Forest responses to
acidic deposition: proceedings of the U. S.-Canadian conference; August 1983; Orono, ME. Orono, ME:
University of Maine, Orono, Land and Water Resources Center; pp. 7-18.
Lovett, G. M. (1992) Atmospheric deposition and canopy interactions of nitrogen. In: Johnson, D. W.; Lindberg,
S. E., eds. Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated forest study.
New York, NY: Springer-Verlag, Inc.; pp. 152-166. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
Lovett, G. M. (1994) Atmospheric deposition of nutrients and pollutants in North America: an ecological
perspective. Ecol. Appl. 4: 629-650.
Lovett, G. M.; Lindberg, S. E. (1984) Dry deposition and canopy exchange in a mixed oak forest as determined by
analysis of throughfall. J. Appl. Ecol. 21: 1013-1027.
Lovett, G. M.; Lindberg, S. E. (1986) Dry deposition of nitrate to a deciduous forest. Biogeochemistry 2: 137-148.
Lovett, G. M.; Lindberg, S. E. (1992) Concentration and deposition of particles and vapors in a vertical profile
through a forest canopy. Atmos. Environ. Part A 26: 1469-1476.
Lovett, G. M.; Lindberg, S. E. (1993) Atmospheric deposition and canopy interactions of nitrogen in forests. Can. J.
For. Res. 23: 1603-1616.
4-257
-------
Lovett, G. M; Traynor, M. M; Pouyat, R. V.; Carreiro, M. M; Zhu, W.-X.; Baxter, J. W. (2000) Atmospheric
deposition to oak forest along an urban-rural gradient. Environ. Sci. Technol. 34: 4294-4300.
Lowenthal, D. H.; Rogers, C. F.; Saxena, P.; Watson, J. G.; Chow, J. C. (1995) Sensitivity of estimated light
extinction coefficients to model assumptions and measurement errors. Atmos. Environ. 29: 751-766.
Ludwig, J.; Marufu, L. T.; Huber, B.; Andreae, M. O.; Helas, G. (2003) Domestic combustion of biomass fuels in
developing countries: a major source of atmospheric pollutants. J. Atmos. Chem. 44: 23-37.
Madronich, S. (1987) Photodissociation in the atmosphere. 1. Actinic flux and the effects of ground reflections and
clouds. J. Geophys. Res. [Atmos.] 92: 9740-9752.
Magill, A. H.; Aber, J. D.; Berntson, G. M.; McDowell, W. H.; Nadelhoffer, K. J.; Melillo, J. M.; Steudler, P. (2000)
Long-term nitrogen additions and nitrogen saturation in two temperate forests. Ecosystems 3: 238-253.
Malm, W. C. (1999) Introduction to visibility. Fort Collins, CO: Colorado State University; Cooperative Institute for
Research in the Atmosphere (CIRA); NPS Visibility Program. Available:
http://www.epa.gov/air/visibility/introvis.pdf [2 April 2002].
Malm, W. C. (2000) Spatial and seasonal patterns and temporal variability of haze and its constituents in the
United States. Report III. Fort Collins, CO: Cooperative Institute for Research in the Atmosphere, Colorado
State University. Available: http://vista.cira.colostate.edu/improve/Publications/Reports/2000/2000.htm
[22 March, 2002].
Malm, W. C.; Kreidenweis, S. M. (1997) The effects of models of aerosol hygroscopicity on the apportionment of
extinction. Atmos. Environ. 31: 1965-1976.
Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
Malm, W. C.; Molenar, J. V.; Eldred, R. A.; Sisler, J. F. (1996) Examining the relationship among atmospheric
aerosols and light scattering and extinction in the Grand Canyon area. J. Geophys. Res. [Atmos.]
101: 19,251-19,265.
Malm, W. C.; Day, D.; Kreidenweis, S. M. (1997) Comparison of measured and reconstructed scattering during an
intensive field study at Great Smoky Mountains National Park. Presented at: 90th annual meeting and
exhibition of the Air & Waste Management Association; June; Toronto, ON, CA. Pittsburgh, PA: Air &
Waste Management Association; 97-WA70.02.
Malm, W. C.; Day, D. E.; Kreidenweis, S. M. (2000a) Light scattering characteristics of aerosols as a function of
relative humidity: part I-a comparison of measured scattering and aerosol concentrations using the theoretical
models. J. Air Waste Manage. Assoc. 50: 686-700.
Malm, W. C.; Day, D. E.; Kreidenweis, S. M. (2000b) Light scattering characteristics of aerosols as a function of
relative humidity: part Il-a comparison of measured scattering and aerosol concentrations using the statistical
models. J. Air Waste Manage. Assoc. 50: 701-709.
Mansfeld, F. (1980) Regional air pollution study: effects of airborne sulfur pollutants on materials. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory;
EPA report no. EPA-600/4-80-007. Available from: NTIS, Springfield, VA; PB81-126351.
Marchwinska, E.; Kucharski, R. (1987) The combined influence of SO2 and heavy metal-containing particulates on
beans, carrots and parsley. Environ. Monit. Assess. 8: 11-25.
Markewitz, D.; Richter, D.D.; Allen, L. H.; Urrego, J. B. (1998) Three decades of observed soil acidification in the
Calhoun Experimental Forest: Has acid rain made a difference? Soil Sci. Soc. Amer. J. 62: 1428-1439.
Marschner, H. (1995) Sulfur. In: Mineral nutrition of higher plants. 2nd ed. London, United Kingdom: Academic
Press; pp. 255-265.
Martin, M. H.; Coughtrey, P. J. (1981) Impact of metals on ecosystem function and productivity. In: Lepp, N. W.,
ed. Effect of heavy metal pollution on plants: volume 2, metals in the environment. Barking, United
Kingdom: Applied Science Publishers; pp. 119-158. (Mellanby, K., ed. Pollution monitoring series).
Martin, C. E.; Gravatt, D. A.; Loeschen, V. S. (1992) Photosynthetic responses of three species to acute exposures of
nitrate- and sulphate-containing aerosols. Atmos. Environ. Part A 26: 381-391.
Mathai, C. V. (1995) The Grand Canyon visibility. EM (December): 20-31.
Matson, P. A.; Parton, W. J.; Power, A. G.; Swift, M. J. (1997) Agricultural intensification and ecosystem properties.
Science (Washington, DC) 277: 504-509.
May, P. F.; Till, A. R.; Gumming, M. J. (1972) Systems analysis of "sulphur kinetics in pastures grazed by sheep.
J. Appl. Ecol. 9: 25-49.
McCartney, H. A.; Aylor, D. E. (1987) Relative contributions of sedimentation and impactionto deposition of
particles in a crop canopy. Agric. For. Meteorol. 40: 343-358.
4-258
-------
McClelland, G.; Schulze, W.; Waldman, D.; Schenk, D.; Irwin, J.; Stewart, T.; Deck, L.; Thayer, M. (1993) Valuing
eastern visibility: a field test of the contingent valuation method. Washington, DC: U.S. Environmental
Protection Agency, Office of Policy, Planning and Evaluation; cooperative agreement no. CR-815183.
McCune, D. C.; Silberman, D. H.; Mandl, R. H.; Weinstein, L. H.; Freudenthal, P. C.; Giardina, P. A. (1977) Studies
on the effects of saline aerosols of cooling tower origin on plants. J. Air Pollut. Control Assoc. 27: 319-324.
McDonald, R. L.; Unni, C. K.; Duce, R. A. (1982) Estimation of atmospheric sea salt dry deposition: wind speed
and particle size dependence. J. Geophys. Res. 87: 1246-1250.
McDonnell, M. J.; Pickett, S. T. A.; Groffman, P.; Bohlen, P.; Pouyat, R. V.; Zipperer, W. C.; Parmelee, R. W.;
Carreiro, M. M.; Medley, K. (1997) Ecosystem processes along an urban-to-rural gradient. Urban Ecosyst.
1:21-36.
McDow, S. R.; Vartiainen, M.; Sun, Q.; Hong, Y.; Yao, Y.; Kamens, R. M. (1995) Combustion aerosol water
content and its effect on polycyclic aromatic hydrocarbon reactivity. Atmos. Environ. 29: 791-797.
McGee, E. S.; Mossotti, V. G. (1992) Gypsum accumulation on carbonate stone. Atmos. Environ. Part B
26: 249-253.
McGowan, T. F.; Lipinski, G. E.; Santoleri, J. J. (1993) New rules affect the handling of waste fuels. Chem. Eng.
(March): 122-128.
Mclntyre, N. E.; Knowles-Yanez, K.; Hope, D. (2000) Urban ecology as an indisciplinary field: differences in the
use of "urban" between the social and natural sciences. Urban Ecosyst. 4: 5-24.
McKay, W. A.; Garland, J. A.; Livesley, D.; Halliwell, C. M.; Walker, M. I. (1994) The characteristics of the
shore-line sea spray aerosol and the landward transfer of radionuclides discharged to coastal sea water.
Atmos. Environ. 28: 3299-3309.
McLachlan, M. S. (1996a) Bioaccumulation of hydrophobic chemicals in agricultural food chains. Environ. Sci.
Technol. 30: 252-259.
McLachlan, M. S. (1996b) Biological uptake and transfer of polychlorinated dibenzo-p-dioxins and dibenzofurans.
In: Hester, R. E.; Harrison, R. M., eds. Chlorinated organic micropollutants. Cambridge, United Kingdom:
Royal Society of Chemistry; pp. 31-52. (Issues in environmental science and technology).
McLachlan, M. S. (1999) Framework for the interpretation of measurements of SOCs in plants. Environ. Sci.
Technol. 33: 1799-1804.
McMahon, T. A.; Denison, P. J. (1979) Empirical atmospheric depositionparameters-a survey. Atmos. Environ.
13: 571-585.
McMillen, R. T. (1988) An eddy correlation technique with extended applicability to non-simple terrain. Boundary
Layer Meteorol. 43: 231-245.
McMurry, P. H. (2000) A review of atmospheric aerosol measurements. Atmos. Environ. 34: 1959-1999.
McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
aerosols. Atmos. Environ. 23: 497-507.
McMurry, P. H.; Zhang, X.; Lee, C.-T. (1996) Issues in aerosol measurement for optics assessments. J. Geophys.
Res. [Atmos.] 101: 19,189-19,197.
Meiwes, K. J.; Khanna, P. K. (1981) Distribution and cycling of sulphur in the vegetation of two forest ecosystems
in an acid rain environment. Plant Soil 60: 369-375.
Middleton, W. E. K. (1952) Vision through the atmosphere. Toronto, ON, Canada: University of Toronto Press.
Midwest Regional Planning Organization. (2001) Regional haze and visibility in the upper Midwest. Des Plaines,
IL: Midwest Regional Planning Organization.
Milford, J. B.; Davidson, C. I. (1987) The sizes of paniculate sulfate and nitrate in the atmosphere-a review. JAPCA
37: 125-134.
Millar, C. S. (1974) Decomposition of coniferous leaf litter. In: Dickinson, C. H.; Pugh, G. J. F., eds. Biology of
plant litter decomposition: volume I. New York, NY: Academic Press; pp. 105-128.
Miller, P. R.; McBride, J. R., eds. (1999) Oxidant air pollution impacts in the Montane forests of southern
California: a case study of the San Bernardino Mountains. New York, NY: Springer-Verlag.
(Ecological Studies: v. 134.)
Miller, W. C.; Fornes, R. E.; Gilbert, R. D.; Speer, A.; Spence, J. (1992) Removal of CaCO3 extender in residential
coatings by atmospheric acidic deposition. In: Measurement of toxic and related air pollutants: proceedings of
the 1992 U.S. EPA/A&WMA international symposium. Pittsburgh, PA: Air & Waste Management
Association; pp. 129-134. (A&WMA publication VTP-25).
Miller, E. K.; Panek, J. A.; Friedland, A. J.; Kadlecek, J.; Mohnen, V. A. (1993) Atmospheric deposition to a
high-elevation forest at Whiteface Mountain, New York, USA. Tellus Ser. B 45: 209-227.
Mims, F. M., III. (1996) Significant reduction of UVB caused by smoke from biomass burning in Brazil.
Photochem. Photobiol. 64: 814-816.
4-259
-------
Mitchell, M. I; Harrison, R. B.; Fitzgerald, J. W.; Johnson, D. W.; Lindberg, S. E.; Zhang, Y.; Autry, A. (1992a)
Sulfur distribution and cycling in forest ecosystems. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric
deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
Springer-Verlag; pp. 90-129. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
Ecological studies: analysis and synthesis: v. 91).
Mitchell, M. J.; Foster, N. W.; Shepard, J. P.; Morrison, I. K. (1992b) Nutrient cycling in Huntington Forest and
Turkey Lakes deciduous stands: nitrogen and sulfur. Can. J. For. Res. 22: 457-464.
Molenar, J. V.; Malm, W. C.; Johnson, C. E. (1994) Visual air quality simulation techniques. Atmos. Environ.
28: 1055-1063.
Molenar, J. V. (2000) Visibility science and trends in the Lake Tahoe Basin: 1989-1998. Stateline, NV: Tahoe
Regional Planning Agency; prepared by Air Resource Specialists, Inc.; February 15.
Molenar, J. V.; Malm, W. C.; Johnson, C. E. (1994) Visual air quality simulation techniques. Atmos. Environ.
28: 1055-1063.
Monteith, J. L.; Unsworth, M. H. (1990) Principles of environmental physics. 2nd ed. London, United Kingdom:
Edward Arnold; pp. 165-175.
Moomaw, W. R. (2002) Energy, industry and nitrogen: strategies for decreasing reactive nitrogen emissions. Ambio
31: 184-189.
Moosmtiller, H.; Arnott, W. P.; Rogers, C. F. (1998) Photoacoustic and filter measurements related to aerosol light
absorption during the Northern Front Range Air Quality Study (Colorado 1996/1997). J. Geophys. Res.
[Atmos.] 103: 28,149-28,157.
Moosmtiller, H.; Arnott, W. P.; Rogers, C. F.; Bowen, J. L.; Gillies, J. A.; Pierson, W. R.; Collins, J. F.; Durbin,
T. D.; Norbeck, J. M. (2001) Time-resolved characterization of diesel paniculate emissions. 2. Instruments
for elemental and organic carbon measurements. Environ. Sci. Technol. 35: 1935-1942.
Moropoulou, A.; Bisbikou, K.; Torts, K.; Van Grieken, R.; Zezza, F.; Macri, F. (1998) Origin and growth of
weathering crusts on ancient marbles in industrial atmosphere. Atmos. Environ. 32: 967-982.
Mueller, S. F. (1988) Chemical deposition to high-elevation spruce-fir forests in the eastern United States. Muscle
Shoals, AL: Tennessee Valley Authority, Air Quality Branch.
Naeem, S.; Thompson, L. J.; Lawler, S. P.; Lawton, J. H.; Woodfin, R. M. (1994) Declining biodiversity can alter
the performance of ecosystems. Nature (London) 368: 734-737.
Nash, T. H., III. (1975) Influence of effluents from a zinc factory on lichens. Ecol. Monogr. 45: 183-198.
National Park Service. (1998) IMPROVE program goals and objectives. Washington, DC: National Park Service.
Available: www2.nature.nps.gov/ard/impr/index.htm [24 November 1999].
National Research Council (NRC). (1993) Protecting visibility in national parks and wilderness areas. Washington,
DC: National Academy Press. 3v.
National Science and Technology Council. (1998) National acid precipitation assessment program biennial report to
Congress: an integrated assessment; executive summary. Silver Spring, MD: U.S. Department of Commerce,
National Oceanic and Atmospheric Administration. Available:
www.nnic.noaa.gov/CENR/NAPAP/NAPAP_96.htm [24 November 1999].
National Weather Service. (1998) Visibility. Washington, DC: National Weather Service, National Oceanic and
Atmospheric Administration. Available: www.nws.noaa.gov/asos/vsby.htm [24 November 1999].
Nazaroff, W. W.; Cass, G. R. (1991) Protecting museum collections from soiling due to the deposition of airborne
particles. Atmos. Environ. Part A 25: 841-852.
Nelis, P. M.; Branford, D.; Unsworth, M. H. (1994) A model of the transfer of radioactivity from sea to land in sea
spray. Atmos. Environ. 28: 3213-3223.
Neuberger, H.; Hosier, C. L.; Kocmond, W. C. (1967) Vegetation as aerosol filter In: Tromp, S. W.; Weihe, W. H.,
eds. Biometeorology, v. 2. pp. 693-702.
Neuroth, G.; Heisler, S. (2000) The nature and sources of visibility impairment in Phoenix. Presented at: Governor's
Brown Cloud Summit; August; Phoenix, AZ. Phoenix, AZ: Arizona Department of Environmental Quality.
Available: http://www.adeq.state.az.us/environ/air/browncloud/download/gary.pdf (28 May 2003).
Nicholson, K. W. (1988) The dry deposition of small particles: a review of experimental measurements. Atmos.
Environ. 22: 2653-2666.
Nicholson, K. W.; Davies, T. D. (1987) Field measurements of the dry deposition of paniculate sulphate. Atmos.
Environ. 21: 1561-1571.
Nielsen, K. E.; Ladekarl, U. L.; Nornberg, P. (1999) Dynamic soil processes on heathland due to changes in
vegetation to oak and Sitka spruce. For. Ecol. Manage. 114: 107-116.
Nihlgard, B. (1985) The ammonium hypothesis-an additional explanation to the forest dieback in Europe. Ambio
14: 2-8.
4-260
-------
Nikliriska, M; Laskowski, R.; Maryariski, M. (1998) Effect of heavy metals and storage time on two types of forest
litter: basal respiration rate and exchangeable metals. Ecotoxicol. Environ. Safety 41: 8-18.
Noble, I. R.; Dirzo, R. (1997) Forests as human-dominated ecosystems. Science (Washington, DC) 277: 522-525.
Nodvin, S. C.; VanMiegroet, H.; Lindberg, S. E.; Nicholas, N. S.; Johnson, D. W. (1995) Acidic deposition,
ecosystem processes, and nitrogen saturation in a high elevation southern Appalachian watershed. Water Air
SoilPollut. 85: 1647-1652.
Noll, K. E.; Fang, K. Y. P. (1989) Development of a dry deposition model for atmospheric coarse particles. Atmos.
Environ. 23: 585-594.
Nord, A. G.; Ericsson, T. (1993) Chemical analysis of thin black layers on building stone. Stud. Conserv. 38: 25-35.
Northeast States for Coordinated Air Use Management (NESCAUM). (2002) CAMNET: realtime air pollution and
visibility monitoring. Boston, MA: NESCAUM. Available at: http://www.hazecam.net. (23 January 2002).
Norton, R. A.; Childers, N. F. (1954) Experiments with urea sprays on the peach. Proc. Am. Soc. Hortic. Sci.
63:23-31.
Novakov, T.; Penner, J. E. (1993) Large contribution of organic aerosols to cloud-condensation-nuclei
concentrations. Nature (London) 365: 823-826.
Ockenden, W. A.; Stinnes, E.; Parker, C.; Jones, K. C. (1998) Observations on persistent organic pollutants in
plants: implications for their use as passive air samplers and for POP cycling. Environ. Sci. Technol.
32: 2721-2726.
Odum, E. P. (1969) The strategy of ecosystem development: an understanding of ecological succession provides a
basis for resolving man's conflict with nature. Science (Washington, DC) 164: 262-270.
Odum, E. P. (1985) Trends expected in stressed ecosystems. BioScience 35: 419-422.
Odum, E. P. (1989) Ecology and our endangered life-support systems. Sunderland, MA: Sinauer Associates, Inc.;
pp. 58-66, 108-118.
Odum, E. P. (1993) The ecosystem. In: Ecology and our endangered life-support systems. 2nd ed. Sunderland, MA:
Sinauer Associates, Inc.; pp. 38-67.
Oehme, M.; Biseth, A.; Schlabach, M.; Wiig, 0. (1995) Concentrations of polychlorinated dibenzo-p-dioxins,
dibenzofurans and non-ortho substituted biphenyls in polar bear milk from Svalbard (Norway). Environ.
Pollut. 90: 401-407.
Olszyk, D. M.; Bytnerowicz, A.; Takemoto, B. K. (1989) Photochemical oxidant pollution and vegetation: effects of
gases, fog, and particles. Environ. Pollut. 61: 11-29.
Oosting, H. J.; Billings, W. D. (1942) Factors effecting vegetational zonation on coastal dunes. Ecology
23: 131-142.
Ormrod, D. P. (1984) Impact of trace element pollution on plants. In: Treshow, M., ed. Air pollution and plant life.
Chichester, United Kingdom: John Wiley & Sons; pp. 291-319.
Ouimette, J. R.; Flagan, R. C. (1982) The extinction coefficient of multicomponent aerosols. Atmos. Environ.
16: 2405-2420.
Padgett, P.; Allen, E. B. (1999) Differential responses to nitrogen fertilization in native shrubs and exotic annuals
common to Mediterranean coastal sage scrub of California. Plant Ecol. 144: 93-101.
Padgett, P. E.; Allen, E. B.; Bytnerowicz, A.; Minich, R. A. (1999) Changes in soil inorganic nitrogen as related to
atmospheric nitrogenous pollutants in southern California. Atmos. Environ. 33: 769-781.
Paerl, H. W.; Bales, J. D.; Ausley, L. W.; Buzzelli, C. P.; Crowder, L. B.; Eby, L. A.; Go, M.; Peierls, B. L.;
Richardson, T. L.; Ramus, J. S. (2001) Ecosystem impacts of three sequential hurricanes (Dennis, Floyd, and
Irene) on the United States' largest lagoonal estuary, Pamlico Sound, NC. Proc. Natl. Acad. Sci. U. S. A.
98: 5655-5611.
Pan, W.; Tatang, M. A.; McRae, G. J.; Prinn, R. G. (1997) Uncertainty analysis of direct radiative forcing by
anthropogenic sulfate aerosols. J. Geophys. Res. [Atmos.] 102: 21,915-21,924.
Pan, W.; Tatang, M. A.; McRae, G. J.; Prinn, R. G. (1998) Uncertainty analysis of indirect radiative forcing by
anthropogenic sulfate aerosols. J. Geophys. Res. [Atmos.] 103: 3815-3823.
Pandis, S. N.; Seinfeld, J. H. (1989) Mathematical modeling of acid deposition due to radiation fog. J. Geophys. Res.
[Atmos.] 94: 12,911-12,923.
Pankow, J. F. (1987) Review and comparative analysis of the theories on partitioning between the gas and aerosol
paniculate phases in the atmosphere. Atmos. Environ. 21: 2275-2283.
Parish, S. B. (1910) The effect of cement dust on citrus trees. Plant World 13: 288-291.
Parker, A. (1955) The destructive effects of air pollution on materials. Colliery Guardian 191: 447-450.
Parker, G. G. (1983) Throughfall and stemflow in the forest nutrient cycle. In: Macfadyen, A.; Ford, E. D., eds.
Advances in ecological research, v. 13; pp. 57-120.
4-261
-------
Patz, J. A.; Engelberg, D.; Last, J. (2000a) The effects of changing weather on public health. Ann. Rev. Public
Health 21: 271-307.
Patz, J. A.; McGeehin, M. A.; Bernard, S. M; Ebi, K. L.; Epstein, P. R.; Grambsch, A.; Gubler, D. J.; Reiter, P.;
Romieu, L; Rose, J. B.; Samet, J. M.; Trtanj, J. (2000b) The potential health impacts of climate variability and
change for the United States: executive summary of the report of the health sector of the U.S. national
assessment. Environ. Health Perspect. 108: 367-376.
Pearcy, R. W.; Bjorkman, O.; Caldwell, M. M.; Keeley, J. E.; Monson, R. K.; Strain, B. R. (1987) Carbon gain by
plants in natural environments. BioScience 37: 21-29.
Penndorf, R. B. (1958) An approximation method to the Mie theory for colloidal spheres. J. Phys. Chem.
62: 1537-1542.
Penner, J. E.; Charlson, R. J.; Hales, J. M.; Laulainen, N. S.; Leifer, R.; Novakov, T.; Ogren, J.; Radke, L. F.;
Schwartz, S. E.; Travis, L. (1994) Quantifying and minimizing uncertainty of climate forcing by
anthropogenic aerosols. Bull. Am. Meteorol. Soc. 75: 375-400.
Percy, K. E. (1991) Effects of acid rain on forest vegetation: morphological and non-mensurational growth effects.
In: Rennie, P. J.; Robitaille, G., eds. Effects of Acid Rain on Forest Resources: Proceedings of the Conference
; June, 1983; Sainte-Foy, Quebec. Ottawa, Canada: Canadian Forestry Service; pp. 97-111.
Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Dutcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
Peterjohn, W. T.; Adams, M. B.; Gilliam, F. S. (1996) Symptoms of nitrogen saturation in two central Appalachian
hardwood forest ecosystems. Biogeochemistry 35: 507-522.
Peters, K.; Eiden, R. (1992) Modelling the dry deposition velocity of aerosol particles to a spruce forest. Atmos.
Environ. Part A 26: 2555-2564.
Peterson, J. T.; Flowers, E. C.; Rudisill, J. H. (1978) Urban-rural solar radiation and atmospheric turbidity
measurements in the Los Angeles basin. J. Appl. Meteorol. 17: 1595-1609.
Peterson, G.; Allen, C. R.; Holling, C. S. (1998) Ecological resilience, biodiversity, and scale. Ecosystems 1: 6-18.
Petuskey, W. T.; Richardson, D. A.; Dolske, D. A. (1995) Aspects of the deterioration of sandstone masonry in
Anasazi dwelling ruins at Mesa Verde National Park, Colorado, USA. Sci. Total Environ. 167: 145-159.
Piatt, J. R.; Krause, P. D. (1974) Road and site characteristics that influence road salt distribution and damage to
roadside aspen trees. Water Air Soil Pollut. 3: 301-304.
Pickett, S. T. A.; Cadenasso, M. L.; Grove, J. M.; Nilon, C. Y.; Pouyat, R. V.; Zipperer, W. C.; Costanza, R. (2001)
Urban ecological systems: linking terrestrial ecological, physical, and socioeconomic components of
metropolitan areas. Annu. Rev. Ecol. Syst. 32: 127-157.
Pilinis, C.; Pandis, S. N.; Seinfield, J. H. (1995) Sensitivity of direct climate forcing by atmospheric aerosols to
aerosol size and composition. J. Geophys. Res. [Atmos.] 100: 18739-18754.
Pimentel, D.; Wilson, C.; McCullum, C.; Huang, R.; Dwen, P.; Flack, J.; Tran, Q.; Saltman, T.; Cliff, B. (1997)
Economic and environmental benefits of biodiversity. BioScience 47: 747-757'.
Pimm, S. L. (1984) The complexity and stability of ecosystems. Nature (London) 307: 321-326.
Pitchford, M. L.; Malm, W. C. (1994) Development and applications of a standard visual index. Atmos. Environ.
28: 1049-1054.
Pitchford, M. L.; McMurry, P. H. (1994) Relationship between measured water vapor growth and chemistry of
atmospheric aerosol for Grand Canyon, Arizona, in winter 1990. Atmos. Environ. 28: 827-839.
Pouyat, R. V.; McDonnell, M. J. (1991) Heavy metal accumulations in forest soils along an urban-rural gradient in
southeastern New York, USA. Water Air Soil Pollut. 57/58: 797-808.
Prescott, C. E.; Parkinson, D. (1985) Effects of sulphur pollution on rates of litter decomposition in a pine forest.
Can. J. Bot. 63: 1436-1443.
Pryor, S. C. (1996) Assessing public perception of visibility for standard setting exercises. Atmos. Environ.
30: 2705-2716.
Pryor, S.; Steyn, D. (1994) Visibility and ambient aerosols in southwestern British Columbia during REVEAL
(REgional Visibility Experimental Assessment in the Lower Fraser Valley): application of analytical
techniques for source apportionment and related visibility impacts of the REVEAL fine paniculate optical and
scene data. Victoria, BC, Canada: British Columbia Ministry of Environment, Lands and Parks. Available
from: NTIS, Springfield, VA; MIC-96-00314/INS.
Pueschel, R. F. (1993) Potential climatic effects of anthropogenic aerosols. In: Jennings, S. G., ed. Aerosol effects on
climate. Tucson, AZ: University of Arizona Press; pp. 110-132.
Rabalais, N. N. (2002) Nitrogen in aquatic ecosystems. Ambio 31: 102-112.
4-262
-------
Rae, D. A. (1983) The value to visitors of improving visibility at Mesa Verde and Great Smoky National Parks. In:
Rowe, R. D.; Chestnut, L. G., eds. Managing air quality and scenic resources at national parks and wilderness
areas. Boulder, CO: Westview Press; pp. 217-234.
Ragsdale, H. L.; Lindberg, S. E.; Lovett, G. M.; Schaefer, D. A. (1992) Atmospheric deposition and throughfall
fluxes of base cations. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient
cycling: a synthesis of the integrated forest study. New York, NY: Springer Verlag; pp. 235-253. (Billings,
W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis:
v. 91).
Ramanathan, V.; Crutzen, P. J.; Kiehl, J. T.; Rosenfeld, D. (2001) Aerosols, climate and the hydrological cycle.
Science (Washington, DC, U.S.) 294: 2119-2124.
Rapport, D. J.; Whitford, W. G. (1999) How ecosystems respond to stress: common properties of arid and aquatic
systems. BioScience 49: 193-203.
Raynal, D. J.; Joslin, J. D.; Thornton, F. C.; Schaedle, M.; Henderson, G. S. (1990) Sensitivity of tree seedlings to
aluminum: III. Red spruce and loblolly pine. J. Environ. Qual. 19: 180-187.
Realini, M.; Negrotti, R.; Appollonia, L.; Vaudan, D. (1995) Deposition of paniculate matter on stone surfaces: an
experimental verification of its effects on Carrara marble. In: Saiz-Jimenez, C., ed. The deterioration of
monuments: proceedings of the 2nd international symposium on biodeterioration and biodegradation;
January 1994; Sevilla, Spain. Sci. Total Environ. 167: 67-72.
Reiners, W. A.; Olson, R. K. (1984) Effects of canopy components on throughfall chemistry: an experimental
analysis. Oecologia63: 320-330.
Reisinger, L. M. (1990) Analysis of airborne particles sampled in the southern Appalachian Mountains. Water Air
SoilPollut. 50: 149-162.
Rennie, P. J. (1955) The uptake of nutrients by mature forest growth. Plant Soil 7: 49-95.
Reuss, J. O. (1983) Implications of the calcium-aluminum exchange system for the effect of acid precipitation on
soils. J. Environ. Qual. 12: 591-595.
Reuss, J. O.; Johnson, D. W. (1986) Acid deposition and the acidification of soils and waters. New York, NY:
Springer-Verlag. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological
studies: analysis and synthesis: v. 59).
Riggan, P. J.; Lockwood, R. N.; Lopez, E. N. (1985) Deposition and processing of airborne nitrogen pollutants in
Mediterranean-type ecosystems of southern California. Environ. Sci. Technol. 19: 781-789.
Rochette, P.; Desjardins, R. L.; Pattey, E.; Lessard, R. (1996) Instantaneous measurement of radiation and water use
efficiencies of a maize crop. Agron. J. 88: 627-635.
Rodney, D. R. (1952) The entrance of nitrogen compounds through the epidermis of apple leaves. Proc. Am. Soc.
Hortic. Sci. 59: 99-102.
Roelofs, J. G. M.; Kempers, A. J.; Houdijk, A. L. F. M.; Jansen, J. (1985) The effect of air-borne ammonium
sulphate onPinus nigra var. maritima in the Netherlands. Plant Soil 84: 45-56.
Roelofs, J. G. M.; Boxman, A. W.; VanDijk, H. F. G. (1987) Effects of airborne ammonium on natural vegetation
and forests. In: Asman, W. A. H.; Diederen, H. S. M. A., eds. Ammonia and acidification: proceedings [of a]
symposium of the European Association for the Science of Air Pollution (EURAS AP); April; Bilthoven,
The Netherlands. European Association for the Science of Air Pollution; pp. 266-276.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
aerosol. 4. Paniculate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol. 27:
2700-2711.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks.
Environ. Sci. Technol. 27: 1892-1904.
Rosenberg, N. J.; Blad, B. L.; Verma, S. B. (1983) Microclimate: the biological environment. 2nd ed. New York,
NY: John Wiley & Sons.
Rosenfeld, D. (1999) TRMM observed first direct evidence of smoke from forest fires inhibiting rainfall. Geophys.
Res. Lett. 26:3105-3108.
Rovira, A. D.; Davey, C. B. (1974) Biology of the rhizosphere. In: Carson, E. W., ed. The plant root and its
environment: proceedings of an institute; July, 1971; Blacksburg, VA. Charlottesville, VA: University Press
of Virginia; pp. 153-204.
Rueth, H.; Baron, J. S. (2002) Differences in Englemann spruce forest biogeochemistry east and west of the
Continental Divide in Colorado, USA. Ecosystems 5: 45-57.
Ruijgrok, W.; Tieben, H.; Eisinga, P. (1997) The dry deposition of particles to a forest canopy: a comparison of
model and experimental results. Atmos. Environ. 31: 399-415.
4-263
-------
Sabbioni, C. (1995) Contribution of atmospheric deposition to the formation of damage layers. In: Jimenez-Saiz, C.,
ed. The deterioration of monuments: proceedings of the 2nd international symposium on biodeterioration and
biodegradation; January 1994; Sevilla, Spain. Sci. Total Environ. 167: 49-56.
Sabbioni, C.; Zappia, G.; Gobbi, G. (1992) Carbonaceous particles on carbonate building stones in a simulated
system. In: Proceedings of the 1992 European aerosol conference; September; Oxford, United Kingdom.
J. Aerosol Sci. 23(suppl. 1): S921-S924.
Sabbioni, C.; Zappia, G.; Gobbi, G. (1996) Carbonaceous particles and stone damage in a laboratory exposure
system. J. Geophys. Res. [Atmos.] 101: 19,621-19,627.
Sabbioni, C.; Zappia, G., Ghedini, N.; Gobbi, G.; Favoni, O. (1998) Black crusts on ancient mortars.
Atmos. Environ. 32: 215-223.
Saiz-Jimenez, C. (1993) Deposition of airborne organic pollutants on historic buildings. Atmos. Environ. Part B
27: 77-85.
Salby, M. L. (1996) Fundamentals of atmospheric physics. New York, NY: Academic Press. (The International
Geophysics Series, Textbooks and Monographs, v. 61).
Samecka-Cymerman, A.; Kempers, A. J. (1999) Bioindication of heavy metals in the town Wroc/aw (Poland) with
evergreen plants. Atmos. Environ. 33: 419-430.
Samuels, H. J.; Twiss, S.; Wong, E. W. (1973) Visibility, light scattering and mass concentration of paniculate
matter: report of the California Tri-City Aerosol Sampling Project. Sacramento, CA: State of California, Air
Resources Board.
Sanyal, B.; Singhania, G. K. (1956) Atmospheric corrosion of metals: parti. J. Sci. Ind. Res. Sect. B 15: 448-455.
Saunders, P. J. W.; Godzik, S. (1986) Terrestrial vegetation-air pollutant interactions: nongaseous air pollutants.
In: Legge, A. H.; Krupa, S. V., eds. Air pollutants and their effects on the terrestrial ecosystem. New York,
NY: John Wiley & Sons; pp. 389-394. (Advances in environmental science and technology: v. 18).
Saxena, P.; Peterson, T. W. (1981) Thermodynamics of multicomponent electrolytic aerosols. J. Colloid Interface
Sci. 79: 496-510.
Saxena, P.; Mueller, P. K.; Kim, Y. P.; Seinfeld, J. H.; Koutrakis, P. (1993) Coupling thermodynamic theory with
measurements to characterize acidity of atmospheric particles. Aerosol Sci. Technol. 19: 279-293.
Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. H. (1995) Organics alter hygroscopic behavior of
atmospheric particles. J. Geophys. Res. [Atmos.] 100: 18,755-18,770.
Schiavon, N.; Chiavari, G.; Schiavon, G.; Fabbri, D. (1995) Nature and decay effects of urban soiling on granitic
building stones. In: Jimenez-Saiz, C., ed. The deterioration of monuments: proceedings of the 2nd
international symposium on biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci. Total
Environ. 167: 87-101.
Schichtel, B. A.; Husar, R. B.; Falke, S. R.; Wilson, W. E. (2001) Haze trends over the United States, 1980-1995.
Atmos. Environ. 35: 5205-5210.
Schier, G. A.; Jensen, K. (1992) Atmospheric deposition effects on foliar injury and foliar leaching in red spruce. In:
Eager, C.; Adams, M. B., eds. Ecology and decline of red spruce in the eastern United States. New York, NY:
Springer-Verlag; pp. 271-294. [Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. D.; Remmert, H., eds.
Ecological Studies, Analysis and Synthesis Ecological Studies, v. 96].
Schimek, M. G. (1981) Smoothing and regression. New York, NY: Wiley.
Schneider, S. H. (1992) Introduction to climate modeling. In: Trenberth, K. E. Climate system modeling. New York,
NY: Cambridge University Press; pp. 1-26.
Schnug, E. (1997) Significance of sulphur for the quality of domesticated plants. In: Cram, W. J.; De Kok, L. J.;
Stulen, L; Brunold, C.; Rennenberg, H., eds. Sulphur metabolism in higher plants: molecular,
ecophysiological and nutritional aspects. Leiden, The Netherlands: Backhuys Publishers; pp. 109-130.
Schonherr, J.; Huber, R. (1977) Plant cuticles are polyelectrolytes with isoelectric points around three. Plant Physiol.
59: 145-150.
Schulze, E.-D. (1989) Air pollution and forest decline in a spruce (Picea abies) forest. Science (Washington, DC)
244: 776-783.
Schuster, P. F.; Reddy, M. M.; Sherwood, S. I. (1994) Effects of acid rain and sulfur dioxide on marble dissolution.
Mater. Perform. 33: 76-80.
Schwar, M. J. R. (1998) Nuisance dust deposition and soiling rate measurements. Environ. Technol. 19: 223-229.
Science Advisory Board. (2002) A framework for assessing and reporting on ecological condition: an SAB report.
Young, T. F.; Sanzone, S., eds. Washington, DC: U.S. Environmental Protection Agency, Ecological
Processes and Effects Committee; report no. EPA-SAB-EPEC-02-009. Available at:
http://www.epa.gov/sciencel/index.html (27 January 2003).
4-264
-------
Seckmeyer, G.; McKenzie, R. L. (1992) Increased ultraviolet radiation in New Zealand (45 degrees S) relative to
Germany (48 degrees N). Nature (London) 359: 135-137.
Sehmel, G. A. (1980) Particle and gas dry deposition: a review. Atmos. Environ. 14: 983-1011.
Sehmel, G. A.; Hodgson, W. H. (1976) Predicted dry deposition velocities. In: Atmosphere-surface exchange of
paniculate and gaseous pollutants (1974): proceedings of a symposium; September 1974; Richland, WA. Oak
Ridge, TN: Energy Research and Development Administration; pp. 399-422. (ERDA symposium series: 38).
Available from: NTIS, Springfield, VA; CONF-740921.
Seinfeld, J. H. (1989) Urban air pollution: state of the science. Science (Washington, DC) 243: 745-752.
Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change.
New York, NY: John Wiley & Sons, Inc.
Sereda, P. J. (1974) Weather factors affecting corrosion of metals. In: Corrosion in natural environments: three
symposia presented at the seventy-sixth annual meeting [of the] American Society for Testing and Materials;
June 1973; Philadelphia, PA. Philadelphia, PA: American Society for Testing and Materials; pp. 7-22.
(ASTM special technical publication 558).
Shannon, J. D. (1999) Regional trends in wet deposition of sulfate in the United States and SO2 emissions from 1980
through 1995. Atmos. Environ. 33: 807-816.
Sharpies, S.; Stewart, L.; Tregenza, P. R. (2001) Glazing daylight transmittances: afield survey of windows in urban
areas. Build. Environ. 36: 503-509.
Shinn, J. H. (1978) A critical survey of measurements of foliar deposition of airborne sulfates and nitrates. Presented
at: 71st annual meeting of the Air Pollution Control Association; June; Houston, TX. Pittsburgh, PA: Air
Pollution Control Association; paper no. 78-7.2.
Shortle, W. C.; Bondietti, E. A. (1992) Timing, magnitude, and impact of acidic deposition on sensitive forest sites.
Water Air Soil Pollut. 61: 253-267.
Shortle, W. C.; Smith, K. T. (1988) Aluminum-induced calcium deficiency syndrome in declining red spruce.
Science (Washington, DC) 240: 1017-1018.
Shortle, W. C.; Smith, K. T.; Minocha, R.; Lawrence, G. B.; David, M. B. (1997) Acidic deposition, cation
mobilization, and biochemical indicators of stress in healthy red spruce. J. Environ. Qual. 26: 871-876.
Showak, W.; Dunbar, S. R. (1982) Effect of 1 percent copper addition on atmospheric corrosion of rolled zinc after
20 years' exposure. In: Dean, S. W.; Rhea, E. C., eds. Atmospheric corrosion of metals: a symposium;
May 1980; Denver, CO. Philadelphia, PA: American Society for Testing and Materials; pp. 135-162.
(ASTM special technical publication 767).
Shriner, D. S.; Henderson, G. S. (1978) Sulfur distribution and cycling in a deciduous forest watershed. J. Environ.
Qual. 7: 392-397.
Sievering, H. (1987) Dynamics of sulfur exchange at the air/forest canopy interface: a review of throughfall inferred
deposition rates. Global Biogeochem. Cycles 1: 233-249.
Sievering, H.; Rusch, D.; Marquez, L. (1996) Nitric acid, paniculate nitrate and ammonium in the continental free
troposphere: nitrogen deposition to an alpine tundra ecosystem. Atmos. Environ. 30: 2527-2537.
Simonich; S. L.; Kites, R. A. (1995) Organic pollutant accumulation in vegetation. Environ. Sci. Technol.
29: 2905-2914.
Simpson, J. W.; Horrobin, P. J., eds. (1970) The weathering and performance of building materials. New York, NY:
Wiley-Interscience.
Sinclair, T. R.; Shiraiwa, T.; Hammer, G. L. (1992) Variation in crop radiation-use efficiency with increased
diffuse-radiation. Crop Sci. 32: 1281-1284.
Sisler, J. F. (1996) Spatial and seasonal patterns and long term variability of the composition of the haze in the
United States: an analysis of data from the IMPROVE network. Fort Collins, CO: Colorado State University,
Cooperative Institute for Research in the Atmosphere; EPA air docket A096-56, document no. VI-B-09-(ee).
Sisler, J. F.; Malm, W. C. (2000) Interpretation of trends of PM25 and reconstructed visibility from the IMPROVE
network. J. Air Waste Manage. Assoc. 50: 775-789.
Sisler, J. F.; Huffman, D.; Lattimer, D. A.; Malm, W. C.; Pitchford, M. L. (1993) Spatial and temporal patterns and
the chemical composition of the haze in the United States: an anlysis of data from the IMPROVE network,
1988-1991. Fort Collins, CO: Colorado State University, Cooperative Institute for Research in the
Atmosphere (CIRA).
Skeffington, R. A.; Wilson, E. J. (1988) Excess nitrogen deposition: issues for consideration. Environ. Pollut.
54: 159-184.
Skerry, B. S.; Alavi, A.; Lindgren, K. I. (1988) Environmental and electrochemical test methods for the evaluation of
protective organic coatings. J. Coat. Technol. 60: 97-106.
4-265
-------
Slinn, W. G. N. (1977) Some approximations for the wet and dry removal of particles and gases from the
atmosphere. Water Air Soil Pollut. 7: 513-543.
Slinn, W. G. N. (1978) Parameterizations for resuspension and for wet and dry deposition of particles and gases for
use in radiation dose calculations. Nucl. Saf. 19: 205-219.
Slinn, W. G. N. (1982) Predictions for particle deposition to vegetative canopies. Atmos. Environ. 16: 1785-1794.
Sloane, C. S. (1983) Optical properties of aerosols-comparison of measurements with model calculations. Atmos.
Environ. 17: 409-416.
Sloane, C. S. (1984) Optical properties of aerosols of mixed composition. Atmos. Environ. 18: 871-878.
Sloane, C. S. (1986) Effect of composition on aerosol light scattering efficiencies. Atmos. Environ. 20: 1025-1037.
Sloane, C. S.; Wolff, G. T. (1985) Prediction of ambient light scattering using a physical model responsive to
relative humidity: validation with measurements from Detroit. Atmos. Environ. 19: 669-680.
Smith, W. H. (1973) Metal contamination of urban woody plants. Environ. Sci. Technol. 7: 631-636.
Smith, W. H. (1974) Air pollution - effects on the structure and function of the temperate forest ecosystem.
Environ. Pollut. 6: 111-129.
Smith, W. H. (1984) Pollutant uptake by plants. In: Treshow, M, ed. Air pollution and plant life. New York, NY:
John Wiley & Sons, Ltd.; pp. 417-450.
Smith, W. H. (1990a) Forests as sinks for air contaminants: soil compartment. In: Air pollution and forests:
interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag;
pp. 113-146. (Springer series on environmental management).
Smith, W. H. (1990b) Forests as sinks for air contaminants: vegetative compartment. In: Air pollution and forests:
interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag;
pp. 147-180. (Springer series on environmental management).
Smith, W. H. (1990c) Forest nutrient cycling: toxic ions. In: Air pollution and forests: interactions between air
contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag; pp. 225-268. (Springer series
on environmental management).
Smith, W. H. (1990d) Forest biotic agent stress: air pollutants and disease caused by microbial pathogens. In: Air
pollution and forests: interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY:
Springer-Verlag; pp. 366-397. (Springer series on environmental management).
Smith, W. H. (1991) Air pollution and forest damage. Chem. Eng. News 69(45): 30-43.
Smith, K. E. C.; Jones, K. C. (2000) Particles and vegetation: implications for the transfer of particle-bound organic
contaminants to vegetation. Sci. Total Environ. 246: 207-236.
Smith, W. H.; Staskawicz, B. J. (1977) Removal of atmospheric particles by leaves and twigs of urban trees:
some preliminary observations and assessment of research needs. Environ. Manage. (N. Y.) 1: 317-330.
Soulen, P. E.; Frederick, J. E. (1999) Estimating biologically active UV irradiance from satellite radiance
measurements: a sensitivity study. J. Geophys. Res. [Atmos.] 104: 4117-4126.
Spence, J. W.; Haynie, F. H. (1972) Paint technology and air pollution: a survey and economic assessment. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Programs; publication no. AP-103.
Available from: NTIS, Springfield, VA; PB-210736.
Spence, J. W.; Haynie, F.; Upham, J. B. (1975) Effects of gaseous pollutants on paints: a chamber study. J. Paint
Technol. 47: 57-63.
Spence, J. W.; Haynie, F. H.; Lipfert, F. W.; Cramer, S. D.; McDonald, L. G. (1992) Atmospheric corrosion model
for galvanized steel structures. Corrosion 48: 1009-1019.
Spiker, E. C.; Comer, V J.; Hosker, R. P.; Sherwood, S. I. (1992) Dry deposition of SO2 on limestone and marble:
the role of humidity. In: Rodrigues, J. D.; Henriques, F.; Jeremias, F. T., eds. Proceedings of the 7th
international congress on deterioration and conservation of stone; June; Lisbon, Portugal. Lisbon, Portugal:
Laboratorio Nacional de Engenharia Civil; pp. 397-406.
Spinka, J. (1971) Vliv znecisteneho ovzdusi na ovocne stromy a zeleninu [Effects of polluted air on fruit trees and
legumes]. Ziva 19: 13-15.
Stanhill, G.; Cohen, S. (2001) Global dimming: a review of the evidence for a widespread and significant reduction
in global radiation with discussion of its probable causes and possible agricultural consequences. Agric. For.
Meteorol. 107: 255-278.
Steubing, L.; Klee, R. (1970) Vergleichende Untersuchungen zur Staubfilterwirkung von Laub- und Nadelgeholzen
[Comparative investigations on the dust-filter effect of broad-leaved and coniferous trees]. Angew. Bot.
44: 73-85.
Stevens, R. K.; Dzubay, T. G.; Lewis, C. W.; Shaw, R. W., Jr. (1984) Source apportionment methods applied to the
determination of the origin of ambient aerosols that affect visibility in forested areas. Atmos. Environ.
18: 261-272.
4-266
-------
Stoddard, J. L. (1994) Long-term changes in watershed retention of nitrogen: its causes and aquatic consequences.
In: Baker, L. A., ed. Environmental chemistry of lakes and reservoirs. Washington, DC: American Chemical
Society; pp. 223-284. (Advances in chemistry series no. 237).
Stoddard, J. L.; Murdoch, P. S. (1991) Catskill Mountains. In: Charles, D. F., ed. Acidic deposition and aquatic
ecosystems: regional case studies. New York, NY: Springer-Verlag; pp. 237-271.
Storm, G. L.; Fosmire, G. J.; Bellis, E. D. (1994) Persistence of metals in soil and selected vertebrates in the vicinity
of the Palmerton zinc smelters. J. Environ. Qual. 23: 508-514.
Strandberg, H.; Johansson, L.-G. (1997) The formation of black patina on copper in humid air containing traces of
SO2. J. Electrochem. Soc. 144: 81-89.
Strickland, T. C.; Holdren, G. R., Jr.; Ringold, P. L.; Bernard, D.; Smythe, K.; Fallen, W. (1993) A national critical
loads framework for atmospheric deposition effects assessment: I. method summary. Environ. Manage.
(N. Y.) 17: 329-334.
Swietlicki, E.; Zhou, J. C.; Berg, O. H.; Martinsson, B. G.; Frank, G.; Cederfelt, S. L; Desek, U.; Berner, A.;
Birmilli, W.; Wiedensohler, A.; Yuskiewicz, B.; Bower, K. N. (1999) A closure study of sub-micrometer
aerosol particle hygroscopic behaviour. Atmos. Res. 50: 205-240.
Swietlik, D.; Faust, M. (1984) Foliar nutrition of fruit crops. Hortic. Rev. 6: 287-355.
Sydberger, T.; Ericsson, R. (1977) Laboratory testing of the atmospheric corrosion of steel. Werkst. Korros.
28: 154-158.
Taback, H. J.; Brienza, A. R.; Macko, J.; Brunetz, N. (1979) Fine particle emissions from stationary and
miscellaneous sources in the south coast air basin: appendix. Sacramento, CA: California Air Resources
Board; contract no. A6-191-30.
Talbot, J. J. (1979) A review of the potential biological impacts of cooling tower salt drift. Atmos. Environ.
13: 395-405.
Tang, I. N. (1980) Deliquescence properties and particle size change of hygroscopic aerosols. In: Willeke, K., ed.
Generation of aerosols and facilities for exposure experiments. Ann Arbor, MI: Ann Arbor Science
Publishers, Inc.; pp. 153-167.
Tang, I. N. (1997) Thermodynamic and optical properties of mixed-salt aerosols of atmospheric importance.
J. Geophys. Res. [Atmos.] 102: 1883-1893.
Tang, I. N.; Munkelwitz, H. R. (1994) Water activities, densities, and refractive indices of aqueous sulfates and
sodium nitrate droplets of atmospheric importance. J. Geophys. Res. [Atmos.] 99: 18,801-18,808.
Tanner, R. L. (1981) An ambient experimental study of phase equilibrium in the atmospheric system: aerosol H+,
NH4+, SO42", NO3" ~NH3(g), HNO3(g). In: Fogler, H. S., ed. Chemical reactors: based on a symposium
sponsored by the Division of Industrial and Engineering Chemistry at the second chemical congress of the
North American continent (180th ACS national meeting); August 1980; Las Vegas, NV. Washington, DC:
American Chemical Society; pp. 527-532. (ACS symposium series 168).
Taylor, G. E., Jr.; Hanson, P. J.; Baldocchi, D. D. (1988) Pollutant deposition to individual leaves and plant
canopies: sites of regulation and relationship to injury. In: Heck, W. W.; Taylor, O. C.; Tingey, D. T., eds.
Assessment of crop loss from air pollutants. New York, NY: Elsevier Applied Science; pp. 227-257'.
Taylor, G. E., Jr.; Johnson, D. W.; Andersen, C. P. (1994) Air pollution and forest ecosystems: a regional to global
perspective. Ecol. Appl. 4: 662-689.
Tegen, I.; Lacis, A. A.; Fung, I. (1996) The influence on climate forcing of mineral aerosols from disturbed soils.
Nature (London) 380: 419-422.
Thornton, F. C.; Schaedle, M.; Raynal, D. J. (1987) Effects of aluminum on red spruce seedlings in solution culture.
Environ. Exp. Bot. 27: 489-498.
Tilman, D. (1996) Biodiversity: population versus ecosystem stability. Ecology 77: 350-363.
Tilman, D. (2000) Causes, consequences and ethics of biodiversity. Nature (London) 405: 208-211.
Tilman, D.; Downing, J. A. (1994) Biodiversity and stability in grasslands. Nature (London) 367: 363-365.
Tjepkema, J. D.; Cartica, R. J.; Hemond, H. F. (1981) Atmospheric concentration of ammonia in Massachusetts and
deposition on vegetation. Nature (London) 294: 445-446.
Tolley, G. A.; Randall, A.; Blomquist, G.; Fabian, R.; Fishelson, G.; Frankel, A.; Hoehn, J.; Krumm, R.; Mensah, E.;
Smith, T. (1986) Establishing and valuing the effects of improved visibility in eastern United States.
Washington, DC: U.S. Environmental Protection Agency.
Tombach, I.; Thurston, S. A. (1994) The quality of the SCENES measurements: the roles of data quality goals and
evolving technology. In: Aerosols and atmospheric optics: radiative balance and visual air quality:
proceedings of the international specialty conference, volume A; September; Snowbird, UT. Pittsburgh, PA:
Air & Waste Management Association; pp. 21-32. (A&WMA publication VTP-41).
4-267
-------
Tong, S. T. Y. (1991) The retention of copper and lead paniculate matter in plant foliage and forest soil.
Environ. Int. 17: 31-37.
Treshow, M. (1980) Pollution effects on plant distribution. Environ. Conserv. 7: 279-286.
Trettin, C.C.; Johnson, D.W.; Todd, D.E., Jr. (1999) Forest nutrient and carbon pools at Walker Branch watershed:
changes during a 21-year-period. Soil Sci. Soc. Amer. J. 63: 1436-1448.
Trijonis, J.; Thayer, M.; Murdoch, J.; Hagemen, R. (1985) Air quality benefit analysis for Los Angeles and
San Francisco based on housing values and visibility. Sacramento, CA: California Air Resources Board;
ARE contract number A-2-088-32. Available: ftp://ftp.arb.ca.gov/carbis/research/apr/past/a2-088-32.pdf
[29 April, 2004].
Trijonis, J. C.; Pitchford, M.; McGown, M. (1987) Preliminary extinction budget results from the RESOLVE
program. In: Bhardwaja, P. S., ed. Visibility protection: research and policy aspects, an APCA international
specialty conference; September 1986; Grand Teton National Park, WY. Pittsburgh, PA: Air Pollution
Control Association; pp. 872-883. (APCA transactions series no. TR-10).
Trijonis, J. C.; Malm, W. C.; Pitchford, M.; White, W. H.; Charlson, R.; Husar, R. (1991) Visibility: existing and
historical conditions-causes and effects. In: Irving, P. M., ed. Acidic deposition: state of science and
technology, volume III: terrestrial, materials, health and visibility effects. Washington, DC: The U.S. National
Acid Precipitation Assessment Program. (State of science and technology report no. 24).
Turner, J.; Kelly, J. (1981) Relationships between soil nutrients and vegetation in a north coast forest, New South
Wales. Aust. For. Res. 11: 201-208.
Turner, J.; Lambert, M. J. (1980) Sulfur nutrition of forests. In: Shriner, D. S.; Richmond, C. R.; Lingberg, S. E.,
eds. Atmospheric sulfur deposition: environmental impact and health effects, proceedings of the second life
sciences symposium; October 1979; Gatlinburg, TN. Ann Arbor, Ml: Ann Arbor Science Publishers, Inc.;
pp. 321-333.
Turner, J.; Lambert, M. J.; Gessel, S. P. (1977) Use of foliage sulphate concentrations to predict response to urea
application by Douglas-fir. Can. J. For. Res. 7: 476-480.
Twomey, S. (1974) Pollution and the planetary albedo. Atmos. Environ. 8: 1251-1256.
Tyler, G. (1972) Heavy metals pollute nature, may reduce productivity. Ambio 1: 52-59.
U.S. Code. (1990) Clean Air Act, §169B, visibility. U. S. C. 42: §7492.
U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air quality
standards. U. S. C. 42: §§7408-7409.
U.S. Environmental Protection Agency. (1979) Protecting visibility: an EPA report to Congress. Research Triangle
Park, NC: Office of Air Quality Planning and Standards; EPA report no. EPA-450/5-79-008. Available from:
NTIS, Springfield, VA; PB80-220320.
U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
U.S. Environmental Protection Agency. (1986 ) Air quality criteria for lead. Research Triangle Park, NC: Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
EPA-600/8-83/028aF-dF. 4v. Available from: NTIS, Springfield, VA; PB87-142378.
U.S. Environmental Protection Agency. (1987) Assessing the risks of trace gases that can modify the stratosphere.
Washington, DC: U.S. Environmental Protection Agency, Office of Air Radiation; report no. EPA
400/1-87/001A-H.
U.S. Environmental Protection Agency. (1993) Air quality criteria for oxides of nitrogen. Research Triangle Park,
NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
nos. EPA/600/8-9!/049aF-cF. 3v. Available from: NTIS, Springfield, VA; PB95-124533, PB95-124525, and
PB95-124517.
U.S. Environmental Protection Agency. (1995a) Interim findings on the status of visibility research. Research
Triangle Park, NC: Office of Research and Development; report no. EPA/600/R-95/021.
U.S. Environmental Protection Agency. (1995b) Testing of meteorological and dispersion models for use in regional
air quality modeling. Research Triangle Park, NC: Office of Air Quality Planning and Standards; report no.
EPA-454/R-95-005. Available from: NTIS, Springfield, VA; PB95-215638.
U.S. Environmental Protection Agency. (1996a) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
U.S. Environmental Protection Agency. (1996b) Air quality criteria for ozone and related photochemical oxidants.
Research Triangle Park, NC: Office of Research and Development; report nos. EP A/600/AP-93/004aF-cF. 3v.
Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, andPB96-185608. Available:
http://cfpub2.epa.gov/ncea/.
4-268
-------
U.S. Environmental Protection Agency. (1997a) Nitrogen oxides: impacts on public health and the environment.
Washington, DC: Office of Air and Radiation; August. Available:
www.epa.gov/ttncaaal/tl/reports/noxrept.pdf [1999, November 24].
U.S. Environmental Protection Agency. (1997b) Summary of the paniculate monitoring program. Washington, DC:
Office of Air and Radiation.
U.S. Environmental Protection Agency. (1998) Monitoring PM2 5 in ambient air using designated reference of class I
equivalent methods. In: Quality Assurance Handbook for Air Pollution Measurement Systems, V. II, Part II,
Section 2.12. Research Triangle Park, NC: Research and Development, Environmental Monitoring Systems
Laboratory.
U.S. Environmental Protection Agency. (1999a) Protection of visibility. C. F. R. 40: §51.300-51.309.
U.S. Environmental Protection Agency. (1999b) Visibility monitoring guidance. Research Triangle Park, NC:
Emissions, Monitoring, and Analysis Division.
U.S. Environmental Protection Agency. (2000a) Deposition of air pollutants to the great waters. Third report to
Congress. [Executive Summary]. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office
of Air Quality Planning and Standards; report no. EPA-453/R-00-005.
U.S. Environmental Protection Agency. (2000b) Latest findings on national air quality: 1999 status and trends.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards; report no. EPA-454/F-00-002.
U.S. Environmental Protection Agency. (2001a) National air quality and emissions trends report, 1999. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA 454/R-01-004. Available:
http://www.epa.gov/oar/aqtrnd99/ [7 May 2003].
U.S. Environmental Protection Agency. (200Ib) Latest findings on national air quality: 2000 status and trends.
Research Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/K-01-002.
Available: http://www.epa.gov/air/aqtrndOO/brochure/OObrochure.pdf (22 July 2003).
Ulrich, B. (1983) Interaction of forest canopies with atmospheric constituents: SO2, alkali and earth alkali cations
and chloride. In: Ulrich, B.; Pankratn, J., eds. Effects of accumulation of air pollutants in forest ecosystems:
proceedings of a workshop; May 1982; Gottingen, Federal Republic of Germany. Dordrecht, The
Netherlands: D. Reidel Publishing Company; pp. 33-45.
United Nations Environment Programme (UNEP). (1986) Report of the international conference on the assessment
of the role of carbon dioxide and of other greenhouse gases in climate variations and associated impacts;
October 1985; Villach, Austria. Geneva, Switzerland: World Meteorological Organization; WMO no. 661.
United Nations Environment Programme (UNEP). (1998) Environmental effects of ozone depletion: 1998
assessment. J. Photochem. Photobiol. B 46: 1-4.
United Nations Environment Programme (UNEP). (2000) Environmental effects of ozone depletion: interim
summary. Available at: http://www.gcrio.org/ozone/unep2000summary.html [9 April 2002].
Unsworth, M. H. (1984) Evaporation from forests in cloud enhances the effects of acid deposition. Nature (London)
312:262-264.
Unsworth, M. H.; Wilshaw, J. C. (1989) Wet, occult and dry deposition of pollutants on forests. Agric. For.
Meteorol. 47: 221-238.
Van Aalst, R. M. (1982) Dry deposition of NOX. In: Schneider, T.; Grant, L., eds. Air pollution by nitrogen oxides.
Amsterdam, The Netherlands: Elsevier Scientific Publishing Company; pp. 263-270.
Van Beelen, P.; Doelman, P. (1997) Significance and application of microbial toxicity tests in assessing
ecotoxicological risks of contaminants in soil and sediment. Chemosphere 34: 455-499.
VanBreemen, N.; VanDijk, H. F. G. (1988) Ecosystem effects of atmospheric deposition of nitrogen in The
Netherlands. In: Dempster, J. P.; Manning, W. J., eds. Excess nitrogen deposition. Environ. Pollut.
54: 249-274.
Van Breemen, N.; Burrough, P. A.; Velthorst, E. J.; Van Dobben, H. F.; De Wit, T.; Ridder, T. B.; Reijnders, H. F.
R. (1982) Soil acidification from atmospheric ammonium sulphate in forest canopy throughfall. Nature
(London) 299: 548-550.
VanBreemen, N.; Mulder, J.; Driscoll, C. T. (1983) Acidification and alkalinization of soils. Plant Soil 75: 283-308.
Van de Hulst, H. C. (1957, reprint 1981) Light scattering by small particles. New York, NY: Wiley. (Structure of
matter series).
Van derLeun, J. C.; Tang, X.; Tevini, M. (1995) Environmental effects of ozone depletion: 1994 assessment. Ambio
24: 138.
Van derLeun, J. C.; Tang, X.; Tevini, M., eds. (1998) Environmental effects of ozone depletion: 1998 assessment.
J. Photochem. Photobiol. B 46(1-3).
4-269
-------
Van Dijk, H. F. G.; Roelofs, J. G. M. (1988) Effects of excessive ammonium deposition on the nutritional status and
condition of pine needles. Physiol. Plant. 73: 494-501.
Van Borland, R.; Dentener, F. J.; Lelieveld, J. (1997) Radiative forcing due to tropospheric ozone and sulfate
aerosols. J. Geophys. Res. [Atmos.] 102: 28,079-28,100.
Van Egmond, K.; Bresser, T.; Bouwman, L. (2002) The European nitrogen case. Ambio 31: 72-78.
Van Miegroet, H.; Cole, D. W. (1984) The impact of nitrification on soil acidification and cation leaching in red
alder ecosystem. J. Environ. Qual. 13: 586-590.
Van Miegroet, H.; Johnson, D. W.; Todd, D. E. (1993) Foliar response of red spruce saplings to fertilization with Ca
and Mg in the Great Smoky Mountains National Park. Can. J. For. Res. 23: 89-95.
Venkatram, A.; Pleim, J. (1999) The electrical analogy does not apply to modeling dry deposition of particles.
Atmos. Environ. 33: 3075-3076.
Viles, H. A. (1990) The early stages of building stone decay in an urban environment. Atmos. Environ. Part A
24: 229-232.
Virginia Climate Advisory. (2000) A closer look at visibility. Charlottesville, VA: University of Virginia, Virginia
State Climatology Office. (Virginia Climate Advisory: 00/8). Available:
http://climate.virginia.edu/advisory/2000/adOO-8.htm [29 September, 2004].
Viskari, E.-L.; Karenlampi, L. (2000) Roadside Scots pine as an indicator of deicing salt use - a comparative study
from two consecutive winters. Water Air Soil Pollut. 122: 405-419.
Vitousek, P. M.; Mooney, H. A.; Lubchenco, J.; Melillo, J. M. (1997) Human domination of Earth's ecosystems.
Science (Washington, DC) 277: 494-499.
Vogt, K. A.; Dahlgren, R.; Ugolini, F.; Zabowski, D.; Moore, E. E.; Zasoski, R. (1987a) Aluminum, Fe, Ca, Mg, K,
Mn, Cu, Zn and P in above- and belowground biomass. I. Abies amabilis and Tsuga mertensiana.
Biogeochemistry 4: 277-294.
Vogt, K. A.; Dahlgren, R.; Ugolini, F.; Zabowski, D.; Moore, E. E.; Zasoski, R. (1987b) Aluminum, Fe, Ca, Mg, K,
Mn, Cu, Zn and P in above- and belowground biomass. II. Pools, and circulation in a subalpme Abies
amabilis stand. Biogeochemistry 4: 295-311.
VonGlasow, R.; Bott, A. (1999) Interaction of radiation fog with tall vegetation. Atmos. Environ. 33: 1333-1346.
Vora, A. B.; Bhatnagar, A. R. (1987) Comparative study of dust fall on the leaves in high pollution and low
pollution area of Ahmedabad: V. Caused foliar injury. J. Environ. Biol. 8: 339-346.
Vose, J. M.; Swank, W. T. (1990) Preliminary estimates of foliar absorption of 15N-labeled nitric acid vapor (HNO3)
by mature eastern white pine (Pinus stmbus). Can. J. For. Res. 20: 857-860.
Vuilleumier, L.; Harley, R. A.; Brown, N. J.; Slusser, J. R.; Kolinski, D.; Bigelow, D. S. (2001) Variability in
ultraviolet total optical depth during the Southern California Ozone Study (SCOS97). Atmos. Environ.
35: 1111-1122.
Waggoner, A. P.; Weiss, R. E. (1980) Comparison of fine particle mass concentration and light scattering extinction
in ambient aerosol. Atmos. Environ. 14: 623-626.
Waggoner, A. P.; Weiss, R. E.; Ahlquist, N. C.; Covert, D. S.; Will, S.; Charlson, R. J. (1981) Optical characteristics
of atmospheric aerosols. In: White, W. H.; Moore, D. J.; Lodge, J. P., Jr., eds. Plumes and visibility:
measurements and model components: proceedings of the symposium; November 1980; Grand Canyon
National Park, AZ. Atmos. Environ. 15: 1891-1909.
Wagrowski, D. M.; Hites, R. A. (1997) Polycyclic aromatic hydrocarbon accumulation in urban, suburban, and rural
vegetation. Environ. Sci. Technol. 31: 279-282.
Wall, D. H. (1999) Biodiversity and ecosystem functioning. BioScience 49: 107-108.
Wall, D. H.; Moore, J. C. (1999) Interactions underground: soil biodiversity, mutualism, and ecosystem processes.
BioScience 49: 109-117.
Walton, J. R.; Johnson, J. B.; Wood, G. C. (1982) Atmospheric corrosion initiation by sulphur dioxide and
paniculate matter: II. characterisation and corrosivity of individual paniculate atmospheric pollutants.
Br. Corros. J. 17: 65-70.
Wang, W. C.; Yung, Y. L.; Lacis, A. A.; Mo, T.; Hansen, J. E. (1976) Greenhouse effects due to man-made
perturbations of trace gases. Science (Washington, DC) 194: 685-690.
Waring, R. H. (1987) Nitrate pollution: a particular danger to boreal and subalpine coniferous forests. In:
Fujimori, T.; Kimura, M., eds. Human impacts and management of mountain forests: [proceedings of a
symposium]. Ibaraki, Japan: Forestry and Forest Products Research Institute; pp. 93-105.
Waring, R. H.; Schlesinger, W. H. (1985) The carbon balance of trees. In: Forest ecosystems: concepts and
management. Orlando, FL: Academic Press, Inc.; pp. 7-37.
Watson, J. G. (2002) Visibility: science and regulation. J. Air Waste Manage. Assoc. 52: 628-713.
4-270
-------
Watson, J. G.; Chow, J. C. (1993) Ambient air sampling. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement:
principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 622-639.
Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ.
19: 241-266.
Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
(1990) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
Webb, A. H.; Bawden, R. J.; Busby, A. K.; Hopkins, J. N. (1992) Studies on the effects of air pollution on limestone
degradation in Great Britain. Atmos. Environ. Part B 26: 165-181.
Wedding, J. B.; Carlson, R. W.; Stukel, J. J.; Bazzaz, F. A. (1975) Aerosol deposition on plant leaves. Environ. Sci.
Technol. 9: 151-153.
Wedin, D. A.; Tilman, D. (1996) Influence of nitrogen loading and species composition on the carbon balance of
grasslands. Science (Washington, DC) 274: 1720-1723.
Weinbaum, S. A.; Neumann, P. M. (1977) Uptake and metabolism of 15N-labeled potassium nitrate by French prune
(Prunus domestica L.) leaves and the effects of two surfactants. J. Am. Soc. Hortic. Sci. 102: 601-604.
Wellburn, A. R. (1990) Tansley review no. 24: Why are atmospheric oxides of nitrogen usually phytotoxic and not
alternative fertilizers? New Phytol. 115: 395-429.
Wells, B. W. (1939) A new forest climax: the salt spray climax of Smith Island, North Carolina. Torrey Bot. Club
Bull. 66: 629-634.
Wells, B. W.; Shunk, I. V. (1937) Seaside shrubs: wind forms vs. spray forms. Science 85: 499.
Wells, B. W.; Shunk, I. V. (1938) Salt spray: an important factor in coastal ecology. Torrey Bot. Club Bull.
65: 485-492.
Wenny, B. N.; Schafer, J. S.; DeLuisi, J. J.; Saxena, V. K.; Barnard, W. F.; Petropavlovskikh, I. V.; Vergamini, A. J.
(1998) A study of regional aerosol radiative properties and effects on ultraviolet-B radiation. J. Geophys. Res.
[Atmos.] 103: 17,083-17,097.
Wesely, M. L. (1989) Parameterization of surface resistances to gaseous dry deposition in regional-scale numerical
models. Atmos. Environ. 23: 1293-1304.
Wesely, M. L.; Hicks, B. B. (1977) Some factors that affect the deposition rates of sulfur dioxide and similar gases
on vegetation. J. AirPollut. Control Assoc. 27: 1110-1116.
Wesely, M. L.; Hicks, B. B. (2000) A review of the current status of knowledge on dry deposition. Atmos. Environ.
34: 2261-2282.
Wesely, M. L.; Eastman, J. A.; Stedman, D. H.; Yalvac, E. D. (1982) An eddy-correlation measurement of NO2 flux
to vegetation and comparison to O3 flux. Atmos. Environ. 16: 815-820.
Wesely, M. L.; Cook, D. R.; Hart, R. L. (1983) Fluxes of gases and particles above a deciduous forest in wintertime.
Boundary Layer Meteorol. 27: 237-255.
Wesely, M. L.; Cook, D. R.; Hart, R. L.; Speer, R. E. (1985) Measurements and parameterization of paniculate
sulfur dry deposition over grass. J. Geophys. Res. [Atmos.] 90: 2131-2143.
Wesselink, L. G.; Meiwes, K.-J.; Matzner, E.; Stein, A. (1995) Long-term changes in water and soil chemistry in
spruce and beech forests, Soiling, Germany. Environ. Sci. Technol. 29: 51-58.
Westman, W. E. (1977) How much are nature's services worth? Measuring the social benefits of ecosystem
functioning is both controversial and illuminating. Science (Washington, DC) 197: 960-964.
Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
White, W. H. (1991) Contributions to light extinction. In: Irving, P. M., ed. Acidic deposition: state of science and
technology, v. Ill: terrestrial, materials, health and visibility effects; report 24. Washington, DC: National
Acid Precipitation Assessment Program; pp. 85-102.
White, W. H.; Macias, E. S. (1990) Light scattering by haze and dust at Spirit Mountain, Nevada. In: Mathai, C. V.,
ed. Visibility and fine particles: an A&WMA/EPA international specialty conference; October 1989; Estes
Park, CO. Pittsburgh, PA; Air & Waste Management Association; pp. 914-922. (A&WMA transactions series
no. TR-17).
White, W. H.; Macias, E. S.; Nininger, R. C.; Schorran, D. (1994) Size-resolved measurements of light scattering by
ambient particles in the southwestern U.S. A. Atmos. Environ. 28: 909-921.
Williams, M. W.; Baron, J. S.; Caine, N.; Sommerfeld, R.; Sanford, R. (1996) Nitrogen saturation in the
Rocky Mountains. Environ. Sci. Technol. 30: 640-646.
Wilson, E. O. (1997) The creation of biodiversity. In: Raven, P. H., ed. Nature and human society: the quest for a
sustainable world: proceedings of the 1997 forum on biodiversity; October; sponsored by the National
Research Council. Washington, DC: National Academy Press.
Wiman, B. (1981) Aerosol collection by Scots pine seedlings: design and application of a wind tunnel method.
Oikos 36: 83-92.
4-271
-------
Wiman, B. L. B. (1985) A non-stationary model for sulphur aerosol depletion and deposition in a young spruce
stand. Tellus Ser. B 37: 230-242.
Wiman, B. L. B.; Agren, G. I. (1985) Aerosol depletion and deposition in forests~a model analysis. Atmos. Environ.
19: 335-347.
Wiman, B. L. B.; Lannefors, H. O. (1985) Aerosol characteristics in a mature coniferous forest-methodology,
composition, sources and spatial concentration variations. Atmos. Environ. 19: 349-362.
Wiman, B. L. B.; Agren, G. L; Lannefors, H. O. (1985) Aerosol concentration profiles within a mature coniferous
forest-model versus field results. Atmos. Environ. 19: 363-367.
Winner, W. E.; Atkinson, C. J. (1986) Absorption of air pollution by plants, and consequences for growth.
Trends Ecol. Evol. 1: 15-18.
Winner, W. E.; Bewley, J. D. (1978a) Terrestrial mosses as bioindicators of SO2 pollution stress: synecological
analysis and the index of atmospheric purity. Oecologia 35: 221-230.
Winner, W. E.; Bewley, J. D. (1978b) Contrasts between bryophyte and vascular plant synecological responses in an
SO2-stressed white spruce association in central Alberta. Oecologia 33:311-325.
Witherspoon, J. P.; Taylor, F. G., Jr. (1971) Retention of 1-44 |_i simulated fallout particles by soybean and sorghum
plants. Health Phys. 21: 673-677.
Wittenburg, C.; Dannecker, W. (1992) Dry deposition and deposition velocity of airborne acidic species upon
different sandstones. In: Proceedings of the 1992 European aerosol conference; September; Oxford,
United Kingdom. J. Aerosol Sci. 23(suppl. 1): S869-S872.
Wolfe, A. H.; Patz, J. A. (2002) Reactive nitrogen and human health: acute and long-term implications. Ambio
31: 120-125.
Wolff, G. T. (1984) On the nature of nitrate in coarse continental aerosols. Atmos. Environ. 18: 977-981.
Wolff, G. T.; Collins, D. C.; Rodgers, W. R.; Verma, M. H.; Wong, C. A. (1990) Spotting of automotive finishes
from the interactions between dry deposition of crustal material and wet deposition of sulfate. J. Air Waste
Manage. Assoc. 40: 1638-1648.
Woodcock, A. H. (1953) Salt nuclei in marine air as a function of altitude and wind force. J. Meteorol. 10: 362-371.
Woodwell, G. M. (1970) Effects of pollution on the structure and physiology of ecosystems: changes in natural
ecosystems caused by many different types of disturbances are similar and predictable. Science (Washington,
DC) 168: 429-433.
World Health Organization. (1997) Nitrogen oxides. 2nded. Geneva, Switzerland: World Health Organization.
(Environmental health criteria 188).
World Meteorological Organization. (1988) Developing policies for responding to climatic change: a summary of
the discussions and recommendations of workshops; September-October 1987; Villach, Austria; and
November 1987; Bellagio, Austria. Geneva, Switzerland: World Meteorological Organization; report no.
WMO/TD; no. 225. [World Climate Impact Programme series report no. WCIP-1].
World Meteorological Organization. (1999) Scientific assessment of ozone depletion: 1998. Geneva, Switzerland:
World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
World Resources Institute. (2000) World resources 2000-2001: people and ecosystems: the fraying web of life.
Washington, DC: World Resources Institute.
Wu, P.-M.; Okada, K. (1994) Nature of coarse nitrate particles in the atmosphere-a single particle approach. Atmos.
Environ. 28: 2053-2060.
Wyttenbach, A.; Tobler, L.; Bajo, S. (1989) Na, Cl and Br in needles of Norway spruce (P. abies) and in the aerosol
adhering to the needles. Toxicol. Environ. Chem. 19: 25-33.
Xu, J. R.; Balik, C. M. (1989) Infrared attenuated total reflectance study of latex paint films exposed to aqueous
sulfur dioxide. J. Appl. Polym. Sci. 38: 173-183.
Yanai, R. D.; Siccama, T. G.; Arthur, M.. A.; Federer, C. A.; Friedland, A. J. (1999) Accumulation and depletion of
base cations in forest floors in the northeastern United States. Ecology 80: 2774-2787.
Yerrapragada, S. S.; Jaynes, J. H.; Chirra, S. R.; Gauri, K. L. (1994) Rate of weathering of marble due to dry
deposition of ambient sulfur and nitrogen dioxides. Anal. Chem. 66: 655-659.
Yocom, J. E.; Grappone, N. (1976) Effects of power plant emissions on materials. Palo Alto, CA: Electric Power
Research Institute; report no. EPRI/EC-139. Available from: NTIS, Springfield, VA; PB-257 539.
Yocom, J. E.; Upham, J. B. (1977) Effects of economic materials and structures. In: Stern, A. C., ed. Air pollution.
3rd ed. Volume II. New York, NY: Academic Press; pp. 93-94.
Yoshida, L. C.; Allen, E. B. (2001) Response to ammonium and nitrate by a mycorrhizal annual invasive grass and a
native shrub in southern California. Am. J. Bot. 88: 1430-1436.
Young, P. (1996) Pollution-fueled "biodeterioration" threatens historic stone: researchers are probing the role
microorganisms play in the decay of buildings and artwork. Environ. Sci. Technol. 30: 206A-208A.
4-272
-------
Young, J. R.; Ellis, E. C.; Hidy, G. M. (1988) Deposition of air-borne acidifiers in the western environment.
J. Environ. Qual. 17: 1-26.
Youngs, W. D.; Rutzke, M.; Gutenmann, W. H.; Lisk, D. J. (1993) Nickel and vanadium in foliage in the vicinity of
an oil-fired power plant. Chemosphere 27: 1269-1272.
Zappia, G.; Sabbioni, C.; Gobbi, G. (1993) Non-carbonate carbon content on black and white areas of damaged
stone monuments. Atmos. Environ. Part A 27: 1117-1121.
Zappia, G.; Sabbioni, C.; Pauri, M. G.; Gobbi, G. (1994) Mortar damage due to airborne sulfur compounds in a
simulation chamber. Mater. Struct. 27: 469-473.
Zappia, G., Sabbioni, C.; Riontino, C.; Gobbi, G.; Favoni, O. (1998) Exposure tests of building materials in urban
atmosphere. Sci. Total Environ. 224: 235-244.
Zdanovskii, A. B. (1948) New methods of calculating solubilities of electrolytes in multicomponent systems. Zh.
Fiz. Khim. 22: 1475-1485.
Zepp, R. G.; Callaghan, T. V.; Erickson, D. J. (1998) Effects of enhanced solar ultraviolet radiation on
biogeochemical cycles. J. Photochem. Photobiol. B 46: 69-82.
Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
27: 1593-1607.
Zhang, L.; Gong, S.; Padro, J.; Barrie, L. (2001) A size-segregated particle dry deposition scheme for an
atmospheric aerosol module. Atmos. Environ. 35: 549-560.
Zipperer, W. C.; Wu, J.; Pouyat, R. V.; Pickett, S. T. A. (2000) The application of ecological principles to urban and
urbanizing landscapes. Ecol. Appl. 10: 685-688.
Zuburtikudis, L; Triantafyllou, A. (2001) Environmental pollution and its impact on the monuments and works of art
in Greece. J. Environ. Prot. Ecol. 2: 68-72.
4-273
-------
Appendix 4A
Common and Latin Names
Alder, hazel
Alder, red
Bean, common
Beech
Birch, yellow
Blackhaw
Brush box
Ceanothus, hoaryleaf
Chaparral
Coachwood
Corn
Dogwood, flowering
Elm
Fir, balsam
Fir, Douglas
Fir, fraser
Grape
Grass, red brome
Grass, purple moor
Greenbriar
Gum, sweet
Haw, black
Heather, Scottish
Hickory
Alnus serrulata (Aiton) Willdenow
Alnus rubra Bong.
Phaseolus vulgar is L.
Fagus sylvatica L.
Betula alleghaniemis Britt.
Viburnum prunifolium L.
Lophostemon confertus (R. BR.) P.G. Wilson & Waterhouse
Ceanothus crassifolius Tony
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 vulgar is Salisb.
Cory a spp.
4A-1
-------
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
Pine, lodgepole x
jack pine
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 saliva L.
Avenafatua L
Diosporos virginiana L.
Pinus strobus L.
Pinus banksiana Lamb.
Pinus taeda L.
Pinus contorta Loud.
Pinus sylvestris L.
Pinus elliotti Englem.
Pinus contorta (Douglas ex Loud) x P. banksiana Lamb.
4A-2
-------
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
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 vulgaris 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
4A-2
-------
Ascomycotina
Chaetomium sp.
Fungi Imperfect!
Aureobasidium pullulans
Cladosporium sp.
Epicoccum sp.
Pestalotiopsis
Phialophora verrucosa
Pleurophomella = Sirodothis
4A-4
-------
5. HUMAN EXPOSURE TO PARTICULATE MATTER
AND ITS CONSTITUENTS
5.1 INTRODUCTION
5.1.1 Purpose
The U.S. Environmental Protection Agency's (EPA's) regulatory authority for particulate
matter (PM) applies primarily to ambient air and those sources that contribute to ambient air PM
concentrations. Most of the epidemiological studies discussed in Chapter 8 relate measured
community levels of airborne pollutants to population-based health statistics. Of necessity, these
studies rely on some simplifying assumptions regarding exposures. One such assumption is that
air pollutant concentrations measured at a community (or population-oriented) monitoring site
(or the average concentration of several such sites) can serve as a surrogate for the average
personal exposure to ambient PM for the population. However, total personal exposure to PM
includes both ambient and nonambient components; and both components may contribute to
adverse health effects. Thus, a major emphasis must be to develop an understanding of exposure
to PM from sources that contribute to ambient air pollution. Ultimately, it will be necessary to
account for both ambient and nonambient components of personal exposure in order to fully
understand the relationship between PM and health effects. In addition, knowledge of an
individual's personal exposure to ambient, nonambient, and total PM would provide useful
information for studies where health outcomes are tracked individually.
Exposure to environmental contaminants can be interpreted in a number of ways. For
airborne PM, an individual's exposure is ideally based on measurements of PM concentrations
in the air in the individual's breathing zone as the individual moves through space and time.
However, epidemiological studies frequently use the ambient concentration as a surrogate for
exposure. Therefore, understanding exposure is important because adverse health effects
associated with elevated PM concentrations occur at the individual level. Human exposure data
and models provide the link between ambient concentrations (from monitoring data or estimated
with atmospheric transport models) and lung deposition and clearance models that enable
estimates of the source-air concentration-exposure-dose relationship for input into dose-response
5-1
-------
assessments for PM from ambient sources. Personal exposure includes contributions from many
different types of particles, from many sources, and in many different environments.
The goal of this chapter is to provide current information on the development of human
exposure data and models. This includes information on (a) relationships between PM measured
at ambient sites and personal exposures to PM from both ambient and nonambient sources, and
(b) factors that affect these relationships. The human exposure data and models presented in this
chapter provide critical links between ambient monitoring data and PM dosimetry as well as
between the toxicological studies and epidemiologic studies presented in other chapters. The
specific objectives of this chapter are:
(1) To provide an overall conceptual framework of exposure science as applied to PM,
including the identification and evaluation of factors that determine personal PM
exposure;
(2) To provide a concise summary and review of recent data (since 1996) and findings from
pertinent PM exposure studies;
(3) To characterize quantitative relationships between ambient air quality measurements
(mass, chemical components, number, etc. as determined at a community monitoring site)
and total personal PM exposure and its ambient and nonambient components; and
(4) To evaluate the implications of using ambient PM concentrations as a surrogate for
personal exposure in epidemiologic studies of PM health effects.
5.1.2 Particulate Matter Mass and Constituents
Current EPA PM regulations are based on mass as a function of aerodynamic size.
However, the EPA also measures the chemical composition of PM in both mo'nitoring and
research studies. The composition of PM is variable and, as discussed in Chapters 7 and 8,
health effects may be related to PM characteristics other than mass. Because PM from ambient
air and other microenvironments may have different physical and chemical characteristics, PM
from such different sources may also have different health effects. Ultimately, to understand and
control health effects caused by PM exposures from all sources, it is important to quantify and
understand exposure to those chemical constituents from various sources that are responsible for
adverse health effects.
The National Research Council (NRC) recognized the distinction between measuring
exposure to PM mass and to chemical constituents in the document Research Priorities for
5-2
-------
Airborne Paniculate Matter I: Immediate Priorities and a Long-range Research Portfolio
(National Research Council, 1998). Specifically, under Research Topic 1, Outdoor Measures
versus Actual Human Exposures, the NRC recommended evaluation of "the relationships
between concentrations of particulate matter and gaseous co-pollutants measured at stationary
outdoor air monitoring sites and the contributions of these concentrations to actual personal
exposures . . ." for PM mass. The NRC Research Topic 2 recommends evaluating exposures to
biologically important constituents and specific characteristics of PM that cause responses in
potentially susceptible subpopulations and the general population. It also was recognized by the
NRC that "a more targeted set of studies under this research topic (#2) should await a better
understanding of the physical, chemical, and biological properties of airborne particles
associated with the reported mortality and morbidity outcomes" (National Research Council,
1999). The NRC also stated that the later studies "should be designed to determine the extent to
which members of the population contact these biologically important constituents and size
fractions of concern in outdoor air, outdoor air that has penetrated indoors, and air pollutants
generated indoors" (National Research Council, 1999). Thus, exposure studies should include
contributions from all sources. The emphasis in this chapter on PM mass reflects the current
state of the science. Where available, data also have been provided on chemical constituents
although in most cases the data are limited. As recognized by the NRC, a better understanding
of exposures to PM chemical constituents from multiple sources will be required to more fully
identify, understand, and control those sources of PM contributing to adverse health effects and
to accurately define the relationship between PM exposure and health outcomes due to either
short- or long-term exposures.
5.1.3 Relationship to Past Documents
Early versions of PM criteria documents did not emphasize total human exposure, but
rather focused almost exclusively on outdoor air concentrations. For instance, the 1969 Air
Quality Criteria for P articulate Matter (National Air Pollution Control Administration, 1969)
did not discuss either exposure or indoor concentrations. The 1982 EPA PM Air Quality
Criteria Document (1982 PM AQCD), however, provided some discussion of indoor PM
concentrations reflecting an increase in microenvironmental and personal exposure studies (U.S.
Environmental Protection Agency, 1982). The new data indicated that personal activities, along
5O
-------
with PM generated by personal and indoor sources (e.g., cigarette smoking), could lead to high
indoor levels and high personal exposures to total PM. Some studies reported indoor
concentrations that exceeded PM concentrations found in the air outside the monitored
microenvironments or at nearby monitoring sites. Between 1982 and 1996, many more studies
of personal and indoor PM exposure demonstrated that, in most inhabited domestic
environments, indoor PM mass concentrations and personal PM exposures of the residents were
greater than ambient PM mass concentrations measured simultaneously (e.g., Sexton et al., 1984;
Spengler et al., 1985; Clayton et al., 1993). As a result, the National Research Council (1991)
recognized the potential importance of indoor sources of contaminants (including PM) in
causing health outcomes.
The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996) reviewed the human
PM exposure literature through early 1996, mainly to evaluate the use of ambient air monitors as
surrogates for PM exposure in epidemiological studies. Many of the studies cited showed poor
correlations between personal exposure or indoor measurements of PM mass and outdoor or
ambient site measurements. Conversely, Janssen et al. (1995) and Tamura et al. (1996a) showed
that in the absence of major nonambient sources, total PM exposures to individuals tracked
through time were highly correlated with ambient PM concentrations. Analyses of these latter
two studies led to the consideration of ambient and nonambient exposures as separate
components of total personal exposure. As a result, the 1996 PM AQCD (U.S. Environmental
Protection Agency, 1996) distinguished between ambient and nonambient PM personal exposure
for the first time. This chapter builds on the work of the 1996 PM AQCD by further evaluating
the ambient and nonambient components of PM and by reporting research that evaluates the
relationship between ambient concentrations and total, ambient, and nonambient personal
exposure.
5.1.4 Chapter Structure
The chapter is organized to provide information on the principles of exposure, to review
the existing literature, and to summarize key findings and limitations in the information.
Specific sections are described below:
• Section 5.2 discusses the basic concepts of exposure including definitions, methods for
estimating exposure, and methods for estimating ambient air components of exposure.
5-4
-------
Section 5.3 presents PM mass data, including a description of the key available studies,
the relationship of PM exposures with ambient concentrations, and factors that affect
the relationship.
Section 5.4 presents data on PM constituents, including a description of the key available
studies, the relationship with ambient concentrations, and factors that affect the
relationship.
Section 5.5 then discusses the implications of using ambient PM concentrations in
epidemiologic studies of PM health effects.
Section 5.6 summarizes key findings and limitations of the information.
5.2 BASIC CONCEPTS OF EXPOSURE
5.2.1 The Concept of Exposure
"There is reasonable agreement that human exposure [to a substance] means contact with
the chemical or agent." However, contact can be either with "(a) the visible exterior of the
person (skin and openings into the body such as mouth and nostrils), or (b) the so-called
exchange boundaries where absorption takes place (skin, lung, gastrointestinal tract)" (Federal
Register, 1986). In its 1992 Guidelines for Exposure Assessment (U.S. Environmental
Protection Agency, 1992), EPA defined exposure as "taking place at the visible external
boundary, as in (a) above, [concluding that this definition] is less ambiguous and more consistent
with nomenclature in other scientific fields." The 1992 Guidelines stated:
"Under this definition, it is helpful to think of the human body as having a hypothetical
outer boundary separating inside the body from outside the body. This outer boundary
of the body is the skin and the openings into the body such as the mouth, the nostrils,
and punctures and lesions in the skin. As used in these Guidelines, exposure to a
chemical is the contact of that chemical with the outer boundary. An exposure
assessment is the quantitative or qualitative evaluation of that contact; it describes the
intensity, frequency, and duration of contact, and often evaluates the rates at which the
chemical crosses the boundary (chemical intake or uptake rates), the route by which it
crosses the boundary (exposure route; e.g., dermal, oral, or respiratory), and the resulting
amount of the chemical that actually crosses the boundary (a dose) and the amount
absorbed (internal dose)" (U.S. Environmental Protection Agency, 1992).
When applied to PM exposure by the inhalation route, the concentration of interest is that
of PM in the air that will enter the respiratory system, not the average concentration of inspired
5-5
-------
and exhaled air as measured at the mouth or nostrils. Therefore, a measurement of inhalation
exposure to PM is based on measurements of the PM concentration near the breathing zone,
but not affected by exhaled air.
5.2.2 Components of Exposure
The total exposure of an individual over a discrete period of time includes exposures to
many different types of particles from various sources while in different microenvironments.
Duan (1982) defined a microenvironment as "a [portion] of air space with homogeneous
pollutant concentration." It also has been defined as a 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 microenvironment and surrounding ones (Mage, 1985). In general,
people pass through a series of microenvironments including outdoor, in-vehicle, and indoor
microenvironments as they go through space and time. Thus, total daily exposure for a single
individual to PM must be expressed as the sum of various exposures for the microenvironments
that the person occupies in the day (modified from National Research Council, 1991).
In a given microenvironment, particles may originate from a wide variety of sources.
For example, in an indoor microenvironment, PM may be generated by (1) indoor activities,
(2) outdoor PM entering indoors, (3) the chemical interaction of outdoor air pollutants and
indoor air or indoor sources, (4) transport from another indoor microenvironment, or (5) personal
activities. All of these disparate sources should be accounted for when estimating total human
exposure to PM.
An analysis of personal exposure to PM mass (or to its constituent compounds) requires
definition and discussion of several classes of particles and exposure. In this chapter, PM
metrics will be described in terms of exposure or as an air concentration. Particulate matter also
may be described according to both its source (i.e., ambient, nonambient) and the
microenvironment in which exposure occurs. Table 5-1 provides a summary of the terms used
in this chapter, the notations used for these terms, and their definitions. These terms are used
throughout this chapter and provide the terminology for evaluating personal exposure to total
PM and to PM from ambient and nonambient sources.
Exposures are significant only if they are associated with a biologically relevant duration
of contact with a substance of concern (Lioy, 1990; National Research Council, 1991).
5-6
-------
TABLE 5-1. TYPES OF PARTICIPATE MATTER USED IN EXPOSURE
AND CONCENTRATION VARIABLES
General Definitions
Term
Notation
Definition
Concentration
Personal Exposure
C General term for the average air concentration over some specified time
period, used with subscript to indicate concentration of a specific type of
PM, usually expressed in ug/m3 units.
E General term for the average personal exposure over some specified time
period, used with subscript to indicate exposure to a specific type of PM,
quantified as the concentration at the oral/nasal contact boundary.
Types of Particulate Matter
Term
Subscript
Definition
Ambient PM
Ambient-Outdoor PM
Indoor PM
Ambient-Indoor PM
Indoor-generated PM
Indoor-reaction PM
Personal Cloud PM
Personal PM
(of a subject)
a 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. (C without subscript is used for Ca. Ea is used for
exposure to ambient PM while outdoors.)
o Ambient PM in an outdoor microenvironment. (C0 is frequently
considered to be equal to, or at least represented by, C.)
i All PM found indoors.
ai Ambient PM that has infiltrated indoors (i.e., has penetrated indoors and
remains suspended). E^ is used for ambient exposure while indoors.
ig PM generated indoors.
ir PM formed indoors by pollutants from outdoors reacting with indoor-
generated pollutants.
pc PM contributing to personal exposure but not contained in indoor or
outdoor measurements of PM, usually related to personal activities.
s PM at the oral/nasal contact zone as the subject moves through space
and time.
Term
Concentration and Exposure Variable Used without Subscripts
Notation Definition
Ambient Concentration C
Total Personal T
Exposure
Ambient Exposure A
Nonambient Exposure N
Concentration measured at a community ambient air monitoring site
(or the average of several such sites).
Total personal exposure as measured by a personal exposure monitor
(PEM).
Personal exposure to the ambient component of total personal exposure,
i.e., personal exposure to that PM measured at an ambient air community
monitoring site. Includes exposure to C and C^ but not to resuspended
ambient PM previously deposited indoors.
Personal exposure to nonambient PM.
5-7
-------
TABLE 5-1 (cont'd). TYPES OF PARTICIPATE MATTER USED IN EXPOSURE
AND CONCENTRATION VARIABLES
Relationships among Concentration and Exposure Variables for a Two-Compartment Model
(Indoors and Outdoors)
T = A + N
A = yC + (y-1) 0^, where y = fraction of time outside
Application of this concept to PM exposure is complicated by a lack of understanding of the
biological mechanisms of PM toxicity. It is not certain whether the relevant duration is the
instantaneous exposure to a peak concentration, or hourly, daily, or long-term exposure for
months or years (or possibly all of the above). Similarly, it is not certain as to how PM toxicity
depends on particle size or particle composition; whether number, surface area, or mass is the
appropriate metric; or how PM toxicity may be influenced by conditions that might increase
susceptibility such as age, preexisting disease conditions (chronic obstructive pulmonary disease
[COPD], asthma, diabetes, etc.), exposure to infectious agents, exposure to heat or cold, stress,
etc. A person's exposure is clearly influenced by the distribution of many variables and
parameters. A measurement at a single point in space and time along each distribution cannot
adequately describe a person's exposure. Thus, it is important to think of exposure as a path
function with the instantaneous exposure varying as the PM concentration and composition
varies as the person moves through space and time.
The 1997 NAAQS were developed largely on the basis of evidence from epidemiologic
studies that found relatively consistent associations between outdoor PM mass concentrations
and observed health effects. Thus, an emphasis in this chapter is on the relationship between PM
concentrations measured at ambient sites and personal exposures to the PM measured at those
ambient sites (National Research Council, 1998), i.e., ambient PM exposure. Although this is an
emphasis, it should be kept in mind that every particle that deposits in the lung becomes part of
the dose delivered to an individual. It is likely that the nonambient component of total exposure
also exerts health effects not necessarily detected by community time-series epidemiologic
studies.
5-S
-------
5.2.3 Quantification of Exposure
Quantification of inhalation exposure to PM or any other air pollutant starts with the
concept of the variation in the concentration of the air pollutant in the breathing zone (but
unperturbed by exhaled breath) as measured by a personal exposure monitor as a person moves
through space and time. The relationships among the various types of exposure quantities can be
easily seen in the hypothetical exposure time profile shown in Figure 5-1. The peak exposure,
instantaneous exposure, and average exposure have units of concentration, for PM, usually
|ig/m3. The integrated exposure has units of concentration x time.
c
o
C
d>
O
c
o
o
Peak
Exposure
Averaged
Exposure
Instantaneous
Exposure
\
Area = Integrated
Exposure
Time
Figure 5-1. Hypothetical exposure time profile: pollutant exposure as a function of time
showing how the averaged exposure, integrated exposure, and peak exposure
relate to the instantaneous exposure. (*2 ~ t\ = T.)
Source: Adapted from Duan (1989); Zartarian et al. (1997).
5-9
-------
The integrated exposure may be written as:
\Cs(t)dt, (5-1)
/=*,
where £ is the integrated personal exposure during the time period tl to t2 and Cs is the
instantaneous exposure of the subject as the subject moves through space and time (Lioy, 1990;
National Research Council, 1991; Georgopoulos and Lioy, 1994). Because Cs is a function of
both space and time, it cannot generally be correctly approximated by a measurement at a single
point in space or at a single time.
For most of the discussion in this chapter, we will be interested in the average exposure
given as:
i
7 \Cs(i)dt, (5-2)
*i * t
where E is the average exposure during the time period tl to t2. Most studies report 24-h
averages, although some studies measure 12-h or 2- or 3-day averages.
Equations 5-1 and 5-2 apply to a specific individual moving through space and time on a
specific path. When treating populations, the distribution of the values of the variables and
parameters must be considered. Georgopoulos and Lioy (1994) showed how to modify these
equations to consider the probability distributions.
5.2.4 Methods To Estimate Personal Exposure
Personal exposure may be estimated using either direct or indirect approaches. Direct
approaches measure the contact of the person with the chemical concentration in the exposure
media over an identified period of time. Direct measurement methods include personal exposure
monitors (PEMs) for PM that are worn continuously by individuals as they encounter various
microenvironments while performing their daily activities. Indirect approaches use models and
available information on concentrations of chemicals in microenvironments, the time individuals
spend in those microenvironments, and personal PM-generating activities to estimate personal
5-10
-------
exposure. This section describes methods used to directly measure personal exposures and
microenvironmental concentrations as well as the models used to estimate exposure. Several
approaches to estimating personal exposure to ambient PM are also described.
5.2.4.1 Direct Measurement Methods
5.2.4.1.1 Personal Exposure Monitoring Methods
In theory, personal exposure to total PM is defined as the concentration of PM in inhaled
air entering the nose or mouth. Practically, it is measured by sampling PM with a PEM worn by
a person and sampling from a point near the breathing zone (but not affected by exhaled breath).
Personal exposure monitors (PEMs) for PM use measurement techniques similar to those
employed for measuring ambient PM. Most PEMs rely on filter-based mass measurement of a
particle size fraction (PM10 or PM2 5) usually integrated over either a 24- or 12-h period at flow
rates of 2 to 4 L/min using battery-operated pumps. Because PEMs must be worn by study
participants, they must be quiet, compact, and battery-operated. These requirements limit the
type of pumps and the total sample volume that can be collected. Generally, small sample
volumes limit personal exposure measurements to PM mass and a few elements detected by
X-ray fluorescence (XRF). In most studies, PM2 5 and PM10 have not been collected
concurrently; thus, very few data are available by which to estimate personal exposure to coarse
thoracic PM (i.e., PM10_2 5) exposures.
One of the earliest versions of a PEM was developed by the Harvard School of Public
Health for use in the Particle Total Exposure Assessment Methodology (PTEAM) study (Clayton
et al., 1993; Thomas et al., 1993; Ozkaynak et al., 1996a). This 4 L/min sampler used a sharp
cut impactor with 2.5 or 10 jim 50% cut points (Marple et al., 1987). A version of this device is
manufactured by MSP, Inc. (Minneapolis, MN). Several exposure studies have used the MSP
PEM (usually the 2.5 jim PEM version) with a flow rate of 2 L/min (Kamens et al., 1991;
Williams et al., 2000a,b; Landis et al., 2001). Modifications to the MSP PEM, including a
"scalper" nozzle to exclude particles > 4 jim, are described in Pellizzari et al. (1999). The
penetration curve for the MSP scalper and the MSP 2.5 jim impactor (for 2 L/min flow rate) are
given in Rodes et al. (2001).
More recently the Harvard School of Public Health has developed a personal particle
sampler that can operate as a PM2 5 or a PMX sampler and that can be used in a personal
5-11
-------
multipollutant sampler (Sioutas et al., 1996; Demokritou et al., 2001). The PM25 version
(4 L/min) has been used in an exposure study (Liu et al., 2002). Evaluations of other PEMs have
been reported (Williams et al., 1999; Lanki et al., 2002).
Other methods used for ambient PM have also been adapted for use as PEMs. For
example, a personal nephelometer that measures light scattering has been worn by subjects and
used in personal exposure studies to obtain real-time PM measurements (Quintana et al., 2000;
Rea et al., 2001; Magari et al., 2002; Lanki et al., 2002). Light scattering instruments are most
sensitive to particles in the accumulation-mode size range. Recent developments in light
scattering devices have made possible the short-term measurement of personal and indoor
exposures (Liu et al., 2002). These devices include the Radiance nephelometer (Radiance
Research, Seattle, WA), the personal DataRAM or pDR (Thermo MIE Inc., Smyrna, GA), and
the DustTrak aerosol monitor (TSI Inc., St. Paul, MN; Brauer, 1995; Howard-Reed et al., 2000;
Quintana et al., 2000, 2001; Chang et al., 2000; Magari et al., 2002). The Radiance
nephelometer is an active monitor with a 530-nm wavelength defined by an optical filter. The
pDR uses a wavelength of 880 nm, making it more sensitive than the nephelometer to particles
> 1 jim in diameter but less sensitive to particles < 0.3 jim diameter. The pDR (usually used in
the passive mode, without a size fractionation inlet) has been widely used in exposure studies
because of its light weight (no pump required), data logging capability, and easy handling
(Quintana et al., 2000, 2001; Williams et al., 2000a; Muraleedharan and Radojevic, 2000; Rea
et al., 2001). A portable condensation nuclei counter (with a lower size limit of 20 nm diameter)
has also been used in exposure studies (Abraham et al., 2002).
Simultaneous measurements of both gravimetric mass and light scattering coefficients have
provided assessment of both the nephelometer and pDR in outdoor and indoor air (Brauer, 1995;
Quintana et al., 2000; Scheff and Wadden, 1979; Waggoner and Weiss, 1980; Sioutas et al.,
2000). Less information is available to assess their use in personal environments. Liu et al.
(2002) compared the pDR and the Radiance Nephelometer against gravimetric measurements
from both the Harvard impactor for PM2 5 (HI2 5) and the Harvard personal environmental
monitor (HPEM2 5) in indoor, outdoor, and personal settings at residences and on the person of
elderly subjects living across the metropolitan Seattle area.
Ten-minute averages for collocated indoor nephelometers and pDRs at diverse residential
sites have produced varied coefficients of determination across homes (R2 = 0.75 to 0.96). The
5-12
-------
differences between the nephelometer and the pDR increased during cooking hours. When
regressed against 24-h gravimetric measurements made with the HI2 5, the Radiance
nephelometer showed higher coefficients of determination indoors (R2 = 0.81 [non-cooking
days] and 0.93 [days with cooking]) and outdoors (R2 = 0.88 [outdoors]) than the pDR indoors
(R2 = 0.84 [non-cooking days] and 0.77 [days with cooking]). A comparison of the pDR versus
the HPEM2 5, when both were carried by subjects, gave lower coefficients of determination
(R2 = 0.44 [non-cooking days] and 0.60 [days with cooking].
Liu et al. (2002) also report:
"The pDR measurements generally overestimated gravimetric PM2 5 measurements,
with a mean indoor pDR/HI2 5 radio of 1.56 + 0.07. As the pDR uses a calibration
dust density of 2.6 g/cm3, overestimating gravimetric measurements by a factor of
1.56 gives an estimated indoor particle density of 1.67g/cm3. This estimated indoor
particle density is within the range for sulfates (1.76), nitrate (2.1), organic carbon
(1.4), and elemental carbon (2.0) (Ames et al., 2000). Somewhat different results were
obtained for the comparison with the personal gravimetric measurements. When both
the pDR and HPEM25 were carried by subjects, the mean personal pDR/HPEM25 ratio
was 1.27 + 0.08. This overestimation factor leads to an estimated density of the
personal aerosol of 2.05 g/cm3. This larger density seem reasonable if the personal
aerosol is largely suspended crustal material, as has been found by Ozkaynak et al.
(1996a). These overestimation factors were similar but slightly lower than the mean
indoor pDR/HI25 ratio 1.41 reported by Quintana et al. (2000) and the mean personal
pDR/PEM ratio 1.49 by Howard-Reed et al. (2000)."
Howard-Reed et al. (2000) compared pDR readings, averaged over 24 h with
corresponding gravimetric measurements taken with MSP and PM2 5 PEMs (2 L/min). Both
personal samplers were worn by participants in a Baltimore exposure study. Howard-Reed et al.
(2000) reported a coefficient of determination of R2 = 0.66 and a regression equation of pDR
(|ig/m3) = 1.10 x PEM (|ig/m3) + 5.84 for 34 24-h measurements. Lanki et al. (2002), using
the pDR in a flow-through mode and a PM2 5 PEM of their own design, found R2 = 0.66 and
pDR = 1.85 x PEM + 4.52. Sioutas et al. (2000) have also evaluated the pDR in a flow-through
mode of operation (2 L/min, with a diffusion dryer in the inlet). They found a significant
increase in pDR to gravimetric mass for relative humidity (RH) above 60%. The ratio was 2:1 at
RH = 80%. They subsequently used a diffusion drying tube in the inlet. With this system, they
were able to study the sensitivity of the pDR to particle size by comparing the pDR reading to
the mass measured by the multi-orifice, uniform deposit detector (MOUDI) as a function of the
5-13
-------
particle mass mean diameter (MMD) as determined by the MOUDI. As the MMD increases
from 0.3 to 1.1 |im, the pDR-to-MOUDI ratio increases from approximately 0.7 to about 1.6.
As the MMD increases to 1.5 jam, the ratio decreases to about 1.0.
Chang et al. (2001) evaluated the DustTrak against a MSP PEM in a roll-around system.
Particles were passed through at 2.5 jim PEM size-selective inlet and a Nafion diffusion dryer
before entering the DustTrak. Hourly PM2 5 concentrations measured by the DustTrak were
averaged and compared to 12-h integrated MSP PEM PM2 5 measurements. The DustTrak
concentrations were higher than the PEM by a factor of about 2. The use of a DustTrak
correction factor (2.07 in summer and 2.02 in winter) resulted in good agreement between the
two methods, with R2 = 0.87 in the summer (35 pairs) and R2 = 0.81 in the winter (42 pairs).
5.2.4.1.2 MicroenvironmentalMonitoring Methods
Direct measurements of microenvironmental PM concentrations, which are used with
models to estimate personal exposure to PM, also employ methods similar to those used for
ambient PM. These methods differ from PEMs in that they are stationary with respect to the
microenvironment (i.e., a stationary PEM). Microenvironmental monitoring methods include
filter-based mass measurements of particle size fractions (PM10, PM25) usually integrated over
either a 24- or 12-h period such as the Federal Reference Methods (FRMs) for PM2 5 and PM10.
Flow rates vary between various devices from 4 to 20 L/min. Larger sample volumes allow
more extensive chemical characterization to be conducted on microenvironmental samples.
Because more than one pumping system can be used in a microenvironment, PM2 5 and PM10 can
be collected simultaneously.
The Harvard School of Public Health has developed a 10 L/min sampler, known as the
Harvard Impactor (HI2 5, Air Diagnostics and Engineering, Inc., Naples, ME), which has been
used as an indoor or outdoor monitor in exposure studies (Liu et al., 2002). Babich et al. (2000)
reported that the HI2 5 showed excellent agreement (R2 = 0.99) with the PM2 5 Federal Reference
Method (FRM) for 81 24-h samples in Riverside and Bakersfield, CA. The HI also showed good
precision (4.8%) for 243 24-h collocated samples over eight studies. Other continuous ambient
PM measurement methods that have been utilized for microenvironmental monitoring include
the Tapered Element Oscillating Microbalance (TEOM) and nephelometers. Various continuous
5-14
-------
techniques for counting particles by size (Climet, LASX, SMPS, APS) also have also been used.
Measurement techniques are discussed in Chapter 2.
5.2.4.2 Indirect or Modeling Methods
5.2.4.2.1 Personal Exposure Models
As a relatively new field, exposure modeling for PM mass (PM2 5 and PM10_2 5) and PM
chemical constituents faces significant methodological challenges and input data limitations.
Exposure models typically use one of two general approaches: (1) a time-series approach that
estimates individuals' microenvironmental exposures sequentially over time or (2) a time-
averaged approach that estimates individuals' microenvironmental exposures using average
microenvironmental concentrations and the total time spent in each microenvironment.
Although the time-series approach to modeling personal exposures provides the appropriate
structure for accurately estimating personal exposures (Esmen and Hall, 2000; Mihlan et al.,
2000), a time-averaged approach typically is used when the input data needed to support a time-
series model are not available. However, the time-varying dose profile of an exposed individual
can only be modeled by using the time-series approach (McCurdy, 1997, 2000).
Even though the processes that lead to exposure are nonlinear in nature, personal exposure
models are often used to combine microenvironmental concentration data with human activity
pattern data to estimate personal exposures. Time-averaged models can be used to estimate
personal exposure for an individual or for a defined population. Total personal exposure models
estimate exposures for all of the different microenvironments in which a person spends time, and
total average personal exposure is calculated from the sum of these microenvironmental
exposures:
(5-3)
where Ej is the personal exposure in each microenvironment, j (Duan, 1982). Example
microenvironments include outdoors, indoors at home, indoors at work, and in transit. Each
microenvironmental exposure, Ep is calculated from the average concentration in
microenvironment 7, Cp weighted by the time spent in microenvironment j, tp Tis the sum
5-15
V * V
E = /. EI = —\ J, Cjtj\,
-------
of tj over ally. This model has been applied to concentration data in a number of studies
(Ott, 1984; Ott et al., 1988, 1992; Miller et al., 1998; Klepeis et al., 1994; Lachenmyer and
Hidy, 2000).
Many exposure studies employ 24-h average measurements of concentration indoors and
outdoors and use these concentrations with time spent indoors and outdoors, as shown in
Equation 5-3. It is important to note that although measurement data may be an average
concentration over some time period (i.e., 24 h), significant variations in PM concentrations can
occur during that time period. Thus, an error may be introduced if real-time concentrations are
highly variable and an average concentration for a microenvironment is used to estimate
exposure when the individual is in that microenvironment for only a fraction of the total time.
This may create large errors if the indoor (e.g., in a house) 24-h average includes significant time
periods when there are no people in the house, because the indoor concentrations are increased
by the activities of people. In an effort to overcome these errors, the Air Pollution Exposure
Distribution within Urban Populations in Europe (EXPOLIS) study (Kousa et al., 2002) turned
outdoor samplers off when the subject was indoors and the indoor sampler off with the subject
was outdoors. This provides a better estimate of Ea and E; to compare with the PEM
measurement and allows a better calculation of Epc. However, it does not provide data that can
be used to regress C or C0 with C;. This problem could be overcome with continuous
monitoring, and the extent of the error could be calculated. Also, shorter time interval
information might be useful in examining relationships between acute effects and short-term
exposures.
Microenvironmental concentrations used in the exposure models can be measured directly
or estimated by one or more microenvironmental models. Microenvironmental models vary in
complexity from a simple indoor/outdoor (I/O) ratio to a multi-compartmental mass-balance
model. A discussion of microenvironmental models is presented in Section 5.3.4.2.2.
On the individual level, the time spent in the various microenvironments is obtained from
time-activity diaries that are completed by the individual. For population-based estimates, the
time spent in various microenvironments is obtained from human activity databases. Many of
the largest human activity databases have been consolidated by the EPA's National Exposure
Research Laboratory (NERL) into one comprehensive database called the Consolidated Human
Activity Database (CHAD). CHAD contains over 22,000 person-days of 24-h activity data from
5-16
-------
11 different human activity-pattern studies (McCurdy et al., 2000). Population cohorts with
diverse characteristics can be constructed from the activity data in CHAD and used for exposure
analysis and modeling (McCurdy, 2000). These databases can also be used to estimate
inhalation rates based on activity levels, age, gender, and weight for dosimetry calculations.
A human activity database may contain information on human location, activity, and exertion
rate as well as information such as the presence of combustion sources (e.g., wood fireplaces,
smokers). However, in exposure studies, "activity" usually refers to a person's location in space,
i.e., in what microenvironment at what times. In dosimetry, "activity" is used as an indication of
the level of physical exertion and is used to estimate breathing rate and the extent of mouth,
nose, or combined breathing. Table 5-2 is a summary of the human activity studies in CHAD.
Methodologically, personal exposure models can be divided into three general types:
(1) statistical models based on empirical data obtained from one or more personal monitoring
studies, (2) simulation models based upon known or assumed physical relationships, and
(3) physical-stochastic models that include Monte Carlo or other techniques to explicitly address
variability and uncertainty in model structure and input data (Ryan, 1991; Macintosh et al.,
1995). The attributes, strengths, and weaknesses of these model types are discussed by Ryan
(1991), National Research Council (1991), Frey and Rhodes (1996), and Ramachandran and
Vincent (1999). A review of the logic of exposure modeling is found in Klepeis (1999).
Personal exposure models that have been developed for PM are summarized in Table 5-3.
The regression-based models (Johnson et al., 2000; Janssen et al., 1997; Janssen et al., 1998a)
were developed specifically to account for the observed difference between personal exposure
and microenvironmental measurements and are based on data from a single study, limiting their
utility for broader purposes. Other types of models in Table 5-3 were limited by a lack of data
for the various model inputs. For example, ambient PM monitoring data are not generally of
adequate spatial and temporal resolution for these models. Lurmann and Korc (1994) assumed a
constant relationship between coefficient of haze (COH) and PM10 and used site-specific COH
information to stochastically develop a time series of 1-h PM10 data from every sixth day
24-h PM10 measurements. A mass-balance model typically was used for indoor
microenvironments when sufficient data were available, such as for a residence. For most other
microenvironments, I/O ratios were used because of the lack of data required for a mass-balance
model. In addition, only the deterministic model, PMEX, included estimation of inhaled dose
5-17
-------
oo
TABLE 5-2. ACTIVITY PATTERN STUDIES INCLUDED IN THE CONSOLIDATED HUMAN
ACTIVITY DATABASE (CHAD)
Study Name
Baltimore, MD
CARB: Adolescents
and Adults
CARB: Children
Cincinnati (EPRI)
Denver (EPA)
Calendar Time Period
of the Study
Jan-Feb 1997
Jul-Aug 1998
Oct 1987-Sept 1988
Apr 1989-Feb 1990
Mar-Apr and Aug 1985
Nov 1982-Feb 1983
Diary
Age1
65+
12-94
0-11
0-86
18-70
Days2
391
1762
1200
2614
805
Type3
Diary; 15-min
blocks
Retrospective
Retrospective
Diary
Diary
Time4
24-h Standard
24-h Standard
24-h Standard
24 h; nominal
7 pm-7 am
24 h; nominal
7 pm-7 am
Rate5
No
No
No
Yes
No
Documentation or
Reference
Williams et al. (2000a,b)
Robinson et al. (1991)
Wiley etal. (1991a)
Wiley etal. (1991b)
Johnson (1989)
Akland et al. (1985)
Johnson (1984)
Notes
Multiple days, varying from 5-15;
part of a PM25 PEM study
3 consecutive days; 186 P-D
removed7
Part of CO PEM6 study; 2 consec.
days; 55 P-D removed7
Los Angeles: Elem. Oct 1989
School Children
Los Angeles: Sept-Oct 1990
High School Adoles.
National: NHAPS-A8 Sept 1992-Oct 1994
10-12
13-17
43 Diary
24-h Standard
Yes Spier etal. (1992)
Yes Spier etal. (1992)
7 P-D removed7
23 P-D removed7
National: NHAPS-B8
U. Michigan:
Children
Valdez, AK
Washington, DC
(EPA)
As above
Feb-Decl997
Nov 1990-Oct 1991
Nov 1982-Feb 1983
0-93 4723 Retrospective 24-h Standard
0-93 4663 Retrospective 24-h Standard
0-13 5616 Retrospective 24-h Standard
11-71 401 Retrospective Varying 24-h
period
No9 Klepeis etal. (1995) A national random-probability
Tsang and Klepeis (1996) survey
No9 As above
As above
No Institute for Social Research 2 days of data: one is a weekend
(1997) day
18-98 699 Diary
24 h; nominal
7 pm-7 am
No Goldstein etal. (1992)
No Akland et al. (1985)
Hartwell et al. (1984)
4 P-D removed7
Part of a CO PEM6 study; 6 P-D
removed7
1 All studies included both genders. The age range depicted is for the subjects actually included; in most cases, there was not an upper limit for the adult studies. Ages are inclusive.
Age 0 = babies < 1 year old.
2 The actual number of person-days of data in CHAD after the "flagging" and removal of questionable data. See the text for a discussion of these procedures.
3 Retrospective: a "what did you do yesterday" type of survey; also known as an ex-post survey. Diary: a "real-time" paper diary that a subject carried as he or she went through the day.
4 Standard = midnight to midnight.
5 Was activity-specific breathing rate data collected?
6 PEM = a personal monitoring study. In addition to the diary, a subject carried a small CO or PM2 5 monitor throughout the sampling period.
7 P-D removed = The number of person-days of activity pattern data removed from consolidated CHAD because of missing activity and location information; completeness criteria are listed in
the text.
8 National Human Activity Pattern Study; A = the air version; B = the water version. The activity data obtained on the two versions are identical.
9 A question was asked regarding which activities (within each 6-h time block in the day) involved "heavy breathing," lifting heavy objects, and running hard.
-------
TABLE 5-3. PERSONAL EXPOSURE MODELS FOR PARTICIPATE MATTER
Study
Citation
Model
Name
Model Type Microenvironments or Predictors Output
Notes
vo
Time-series Models
Hayes and PMEX
Marshall
(1999)
Johnson et al.
(2000)
None
Klepeis et al. None
(1994)
Lurmann and
Korc(1994)
Koontz and
Niang (1998)
REHEX-II Stochastic
Deterministic Indoors: residential, work, school
Outdoors: near roadway, other
Motor vehicle
Regression- Auto travel, roadside, ETS, food
based prep., grilling, high ambient PM
Stochastic ETS, cooking, cleaning, attached
garage, wood burning
12 residential with different
sources, restaurant/bar,
nonresidential indoors,
in transit, outdoors
CPIEM Stochastic Indoors: residence, office,
industrial plant, school, public
building, restaurant/lounge, other.
Outdoors, in vehicle
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 etal., 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
SIM
Clayton et al.
(1999a)
Janssen et al.
(1997)
None
Janssen et al. None
(1998a)
Stochastic
Regression-
based
Regression-
based
Smoking parent, ETS exposure,
outdoor physical activity
Distribution of annual
PM2 5 exposures
Accounts for difference between
personal and microenvironmental
PM,n
Number of cigarettes smoked, hours Accounts for difference between
of ETS exposure, residence on busy personal and microenvironmental
road, time in vehicle PM,n
Based on 3-day ambient measurements
Children only
Adults only
-------
TABLE 5-3 (cont'd). PERSONAL EXPOSURE MODELS FOR PARTICIPATE MATTER
Study
Citation
Model
Name
Model Type Microenvironments or Predictors Output
Notes
Time-averaged Models (cont'd)
Ott et al. RCS Statistical
(2000)
Burke et al.
(2001)
SHEDS-
PM
Stochastic
to
o
Chao and
Tung (2001)
None
Mass balance
with empirical
corrections
Not separated
Outdoors, indoors: residence,
office, stores, school, in vehicle,
restaurant/lounge
Indoors in unoccupied residences in
Hong Kong
Distribution of PM10 exposure
for population
PM2 5 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
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 PM2 5 exposure
from active smoking or exposure in subways
Model makes corrections for nonideal mixing
(residence with multiple compartments with
limited intermixing)
-------
from activity-specific breathing rate information. Data from recent PM personal exposure and
microenvironmental measurement studies should help in the future to facilitate the development
of improved personal exposure models for PM.
An integrated human exposure source-to-dose modeling system that will include exposure
models to predict population exposures to environmental pollutants, such as PM, currently is
being developed by the EPA/NERL. A first-generation population exposure model for PM,
called the Stochastic Human Exposure and Dose Simulation (SHEDS-PM) model, recently has
been developed. The SHEDS-PM model uses a 2-stage Monte Carlo sampling technique
previously applied by Macintosh et al. (1995) for benzene exposures. This technique allows for
separate characterization of variability and uncertainty in the model predictions to predict the
distribution of total exposure to PM for the population of an urban/metropolitan area and to
estimate the contribution of ambient PM to total PM exposure. Results from a case study using
data from Philadelphia, PA have been reported (Burke et al., 2001). Work is underway to link
exposure modeling with dosimetry so as to provide estimates of integrated PM doses for
different regions of the lung. In the future, both exposure and dose metrics generated for various
subgroups of concern should aid in evaluation of PM health effects.
5. 2. 4. 2. 2 Microenvironmental Models
The mass-balance model has been used extensively in exposure analysis to estimate PM
concentrations in indoor microenvironments (Calder, 1957; Sexton and Ryan, 1988; Duan, 1982,
1991; McCurdy, 1995; Johnson, 1995; Klepeis et al., 1995; Dockery and Spengler, 1981; Ott,
1984; Ott et al., 1988, 1992, 2000; Miller et al., 1998; Mage et al., 1999; Wilson et al., 2000).
The mass balance model describes the infiltration of particles from outdoors into the indoor
microenvironment, the removal of particles in indoor microenvironments, and the generation of
particles from indoor sources:
(5-4)
where V = volume of the well-mixed indoor air (m3);
C; = concentration of indoor PM (|ig/m3);
v = volumetric air exchange rate between indoors and outdoors (m3/h);
5-21
-------
P = penetration ratio, the fraction of ambient (outdoor) PM that is not removed
from ambient air during its entry into the indoor volume;
C = concentration of PM in the ambient air (|ig/m3);
k = removal rate (h"1); and
Qi = rate of generation of particles from indoor sources (|ig/h).
Qi contains a variety of indoor, particle-generating sources, including combustion or
mechanical processes; condensation of vapors formed by combustion or chemical reaction;
suspension from bulk material; and resuspension of previously deposited PM. The removal rate,
k, includes dry deposition to interior surfaces by diffusion, impaction, electrostatic forces, and
gravitational fallout. It may include other removal processes, such as filtration by forced air
heating, ventilation, or air conditioning (HVAC) or by independent air cleaners. All parameters
except V are functions of time. P and k also are functions of particle aerodynamic diameter, Da,
air exchange rate, v, and house characteristics such as the surface to volume ratio, type of
surface, etc. All variables in Equation 5-3 will have distributions within the population and, in
some cases, may vary by a factor of 5 to 10. It is important to determine the distribution of these
variables. Sensitivity and uncertainty analyses are necessary when attempting to explain model
results.
In addition to the mass-balance model, some single-source or single-microenvironment
models exist. However, most are used to estimate personal exposures to environmental tobacco
smoke (ETS). These models include both empirically based statistical models and physical
models based on first principles: some are time-averaged, whereas others are time-series. These
models evaluate the contribution of ETS to total PM exposure in an enclosed microenvironment
and can be applied as activity-specific components of total personal exposure models. Examples
of ETS-oriented personal exposure models include Klepeis (1999), Klepeis et al. (1996, 2000),
Mage and Ott (1996), Ott (1999), Ott et al. (1992, 1995), and Robinson et al. (1994).
5.2.4.3 Methods for Estimating Personal Exposure to Ambient Particulate Matter
In keeping with the various components of PM exposure described in Section 5.3.2,
personal exposure to PM can be expressed as the sum of exposure to particles from different
sources summed over all microenvironments in which exposure occurs. Total personal exposure
may be expressed as:
5-22
-------
T — F+F+F+F+F a o
i - r,a -i- nai -i- nig -t- r,ir -i- npc p-5)
or
T = A + N, (5-6)
where T is the total personal exposure to ambient and nonambient PM, Ea is personal exposure to
ambient PM while outdoors, Eai is personal exposure to ambient PM that has infiltrated indoors
while indoors, Eig is personal exposure to indoor-generated PM, Eir is exposure to indoor-
reaction PM, and Epc is personal exposure to PM from personal activity (personal cloud). T can
also be expressed as A + N where A is ambient PM exposure (Ea + E^) and N is nonambient PM
exposure (Eig + Eir + Epc). Although personal exposure to ambient and nonambient PM cannot be
measured directly, they can be calculated or estimated from other measurement data.
Approaches for estimating these components of PM exposure are described in the next section.
5.2.4.3.1 Mass Balance Approach
Ambient-Indoor Concentrations of Particulate Matter
The mass balance model described above (Equation 5-4) has been used to estimate PM
concentrations in indoor microenvironments. This model also may be used to estimate ambient-
indoor (C^ and indoor-generated (Cig) PM concentrations. The mass balance model can be
solved for Cai and Cig assuming equilibrium conditions, i.e., all variables remain constant
(Ott et al., 2000; Dockery and Spengler, 1981; Koutrakis et al., 1992) and no indoor reaction
PM (Cir). By substituting a = v/F, where a = the number of air exchanges per hour substituting,
JCai + dCig for dC{ in Equation 5-4, and assuming that JCai and dCig = 0, i.e., ambient-indoor
PM (Cai) and indoor-generated PM (Cig) are at equilibrium, Cai and Cig are given by
Equations 5-7 and 5-8.
Pa
c = c ^__^ (5.7)
(5-8)
5-23
-------
Equations 5-7 and 5-8 assume equilibrium conditions and, therefore, are valid only when the
parameters P, k, a, C, and Qt are not changing rapidly and when the Cs are averaged over several
hours. It should be understood that equilibrium is a simplification of indoor microenvironments
that are occupied by residents. This assumption of equilibrium may only represent a virtual set
of individuals or populations at risk. Under certain conditions (e.g., air-conditioned homes,
homes with HVAC or air cleaners that cycle on and off, or ambient pollutants with rapidly
varying concentrations), nonequilibrium versions of the mass-balance model (Ott et al., 2000;
Freijer and Bloemen, 2000; Isukapalli and Georgopoulos, 2000) are likely to provide a more
accurate estimate of Cai and Cig. However, the equilibrium model provides a useful, although
simplified, example of the basic relationships (Ott et al., 2000).
Equation 5-7 may be rearranged further to give Cai/C, the equilibrium fraction of ambient
PM that is found indoors, defined as the infiltration factor, FINF, (Dockery and Spengler, 1981).
C
The penetration ratio, P, and the decay rate, k, can be estimated using a variety of techniques.
A discussion of these variables and estimation techniques is given in Section 5.4.3.2.2. Both
P and k are a function of particle aerodynamic diameter, air exchange rate, and housing
characteristics. FINF will also be a function of these parameters; and, as a result, FINF may vary
substantially within a population. Distributions of this parameter should be estimated to
understand the uncertainty and variability associated with estimating exposure to PM of ambient
origin. The distribution of daytime FINF, as estimated from PTEAM data, is shown in Figure
5-2a (Wilson et al., 2000).
Personal Exposure to Ambient Particulate Matter
Personal exposure to ambient PM, A, may be estimated using ambient-indoor PM
concentration, Cai, from the mass balance model, ambient outdoor PM concentrations, C, and
information on the time an individual spent in the various microenvironments. For a two-
compartment model, A may be expressed as
5-24
-------
(A
O
+3
TO
0)
(A
-D
O
0)
.D
E
D
(A
C
O
T3
O
•s
E
D
20-
10-
5-
a
r
o;
]
o r
I — i
)?
5 f
13
or
I — i
3
s r
—
14
or
)4
5 r
—
)5
0 (
—
15
5 f
16
0
Tfi
5f
i — i
17
n f
17
5 f
18
0 (
T8
f
5 0.900.951.00
Fraction of Ambient PM10 Found Indoors (F,NF = Caj/C)
25'
20-
10-
5-
0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00
Fraction of Ambient PM10 Found in Total Personal PM10 Exposure
(a = A/C)
Figure 5-2. Distribution of FINF (a) and a (b) for daytime as estimated from PTEAM study
data.
Source: Wilson et al. (2000).
5-25
-------
Pa
(5-10)
where y is the fraction of time that an individual spent outdoors, and (1 -y) is the fraction of
time spent indoors.
It is convenient to express personal exposure to ambient PM, A, as the product of the
ambient PM concentration, C, and a personal exposure or attenuation factor. Following the
usage in several recent papers (Zeger et al., 2000; Dominici et al., 2000; Ott et al., 2000), the
symbol a will be used for this attenuation factor. Equation 5-10 can be rearranged to obtain an
expression for a:
Pa
a + k
(5-11)
Substituting equation 5-9 in equation 5-11 gives a relationship for a in terms of the infiltration
factor FINF and the fraction of time spent in the various microenvironments:
a =y+(l-y)Fim.
(5-12)
Thus, personal exposures to ambient PM, A, may be calculated from measurable quantities:
A= aC.
(5-13)
The factor a can be measured directly or calculated from measured or estimated values of the
parameters a, &, and P and the time spent in various microenvironments from activity pattern
diaries (Wilson et al., 2000). Because a depends on housing and lifestyle factors, air exchange
rate, and PM deposition rate, it could vary to a certain extent from region to region and from
season to season. Consequently, predicted exposures based on these physical modeling concepts
provide exposure distributions derived conceptually as resulting from housing, lifestyles, and
meteorological considerations. For any given population the coefficient a may represent
5-26
-------
substantial intra- and inter-personal variability based on personal activities, housing
characteristics, particle size, and composition. The distribution of daytime a as estimated from
PTEAM data is shown in Figure 5-2b. Note that the distribution of a is shifted to higher values
compared to FINF because of the inclusion of time outdoors in a. Distributions of a should be
determined using population studies in order to evaluate the uncertainty and variability
associated with model exposures.
The mass-balance model has been used to separate indoor concentrations into ambient and
nonambient components. This approach, based on Equation 5-5 as given in Duan (1982) and
called superposition of component concentrations, has been applied using multiple
microenvironments to estimate exposures to carbon monoxide (Ott, 1984; Ott et al., 1988, 1992),
volatile organic compounds (Miller et al., 1998), and particles (Koutrakis et al., 1992; Klepeis
et al., 1994). However, in these studies and in most of the exposure literature, the ambient and
nonambient components are added to yield a personal exposure from all sources of the pollutant.
The use of the mass-balance model, ambient concentrations, and exposure parameters to
estimate exposure to ambient PM and exposure to indoor-generated PM separately as different
classes of exposure has been discussed in Wilson and Suh (1997) and in Wilson et al. (2000).
5.2.4.3.2 The Sulfate Ratio Technique for Estimating Ambient PM Exposure
The ratio of personal exposure to ambient concentration for sulfate has been recommended
as a technique to estimate a (Wilson et al., 2000). If sulfate has no indoor sources, then As = Ts.
(Superscript S refers to sulfate; superscript 2.5 to PM2 5.) As can be seen in Equation 5-11, P and
k depend on particle size, but a and_y do not. It has been known since the mid-1970s that sulfate
and accumulation-mode mass have similar size distributions (Whitby, 1978). Therefore, if the
coarse mode contribution to PM2 5 is small, so that sulfate and PM2 5 would have similar size
distributions, then TS/CS = A2'5/C2'5 = as = a2'5. Sulfate is formed in the ambient air via
photochemical oxidation of gaseous sulfur dioxide arising from the primary emissions from the
combustion of fossil fuels containing sulfur. It also arises from the direct emissions of sulfur-
containing particles from nonanthropogenic sources (e.g., volcanic activity, windblown soil).
In the indoor environment, the only common sources of sulfate may be resuspension by human
activity of deposited PM containing ammonium sulfates or soil sulfates that were tracked into
the home. However, resuspended PM will be mostly larger than PM25. In some homes, an
5-27
-------
unvented kerosene heater using a high-sulfur fuel may be a major contributor during winter
(Leaderer et al., 1999). The use of matches to light cigarettes or gas stoves can also be a source
of sulfates.
Studies that have used the sulfate ratio technique to estimate a and ambient PM exposures
are discussed in Section 5.4.3.1. When there are no indoor sources of accumulation-mode
sulfates, one may deduce that the ambient-to-personal relationship found for sulfates probably
would be the same as that for other PM with the same size range and physical/chemical
properties. This assumption has been validated for several homes in Boston (Sarnat et al., 2002).
For particle sizes within the accumulation-mode size range, the ratio Cai/C was similar for sulfate
and PM2 5 as estimated from SMPS measurements. However, ambient PM with different
physical or chemical characteristics than sulfate will not behave similarly to sulfate. Sulfate has
been used as a marker of outdoor air in the indoor microenvironments (Jones et al., 2000; Ebelt
et al., 2000). However, the personal exposure of sulfate (Ts = As) should not be taken as an
indicator or surrogate for ambient PM2 5 exposure (A2 5) unless it has been previously determined
that PM2 5 and sulfate concentrations are highly correlated. This may be the case in some air
sheds with high sulfate concentrations, but will not be true in general.
5.2.4.3.3 Source-Apportionment Techniques
Source apportionment techniques provide a method for determining personal exposure to
PM from specific sources. If a sufficient number of samples are analyzed with sufficient
compositional detail, it is possible to use statistical techniques to derive source category
signatures, identify indoor and outdoor source categories, and estimate their contribution to
indoor and personal PM. Daily contributions from sources that have no indoor component can
be used as tracers to generate exposure to ambient PM of similar aerodynamic size or directly as
exposure surrogates in epidemiological analyses. Studies that have used source-apportionment
are discussed in Section 5.4.3.3 (i.e., Ozkaynak and Thurston, 1987; Yakovleva et al., 1999; Mar
et al. 2000; Laden et al., 2000).
5.2.4.3.4 Comparison and Validation of Methods
There are as yet no published comparisons or validations of the various methods for
estimating personal exposure to ambient PM on an average or individual daily basis.
5-28
-------
5.3 SUMMARY OF PARTICULATE MATTER MASS DATA
5.3.1 Types of Particulate Matter Exposure Measurement Studies
A variety of field measurement studies have been conducted to quantify personal exposure
to PM mass, to measure microenvironmental concentrations of PM, to evaluate relationships
between personal exposure to PM and PM air concentrations measured at ambient sites, and to
evaluate factors that affect exposure. In general, exposure measurement studies are of two types
depending on how the participants are selected for the study. In a,probability study, participants
are selected using a probability sampling design where every member of the defined population
has a known, positive probability of being included in the sample. Probability study results can
be used to make statistical inferences about the target population. In a purposeful or
nonprobability design, any convenient method may be used to enlist participants; and the
probability of any individual in the population being included in the sample is unknown
(National Research Council, 1991). Participants in purposeful samples may not have the same
characteristics that would lead to exposure as the rest of the unsampled population. Thus, results
of purposeful studies apply only to the subjects sampled on the days that they were sampled and
not to other subjects or other periods of time. Although such studies may report significant
differences, confidence intervals, and/? values, they do not have inferential validity (Lessler and
Kalsbeek,1992). Purposeful studies, however, may have generalizability (external validity).
The extent of generalizability is a matter of judgement based on study participant characteristics.
Purposeful studies of PM personal exposure can provide data with which to develop
relationships based on important exposure factors and can provide useful information for
developing and evaluating either statistical or physical/chemical human exposure models.
Regardless of the sampling design (probability or purposeful), there are three general
categories of study design that can be used to measure personal exposure to PM and evaluate the
relationship between personal PM exposure levels and ambient PM concentrations measured
simultaneously: (1) longitudinal, in which each subject is measured for many days; (2) pooled,
in which each subject is measured for only one or two days, different days for different subjects;
and (3) daily-average, in which many subjects are measured on the same day. Only one study,
in which 14 subjects were measured for 14 days, provided sufficient data for a comparison of
longitudinal and daily-average data (Lioy et al., 1990). Longitudinal and pooled studies are
discussed in Section 5.4.3.1.1.
5-29
-------
5.3.2 Available Data
5.3.2.1 Personal Exposure Data
Table 5-4 gives an overview of the personal exposure studies that have been reported since
the 1996 PM AQCD. In addition, major studies that were reported before that time also have
been included to provide a comprehensive evaluation of data in this area. Table 5-4 gives
information on the sampling and study designs, the study population, the season, number of
participants, PM exposure metric, and the PM size fraction measured.
Although there are a number of studies listed in the table, the data available with which to
evaluate longitudinal relationships and the factors that influence these are limited. Few studies
are based on probability sampling designs that allow study results to be inferred to the general
population and to develop distributional data or exposures and the factors that affect exposure.
Unfortunately, none of these probability studies used a longitudinal study design. This limits our
ability to provide population estimates and distributional data on the relationship between
personal PM exposures and ambient site measurements. In addition, most of the probability
studies of PM exposure were conducted during a single season; thus, variations in ambient
concentrations, air exchange rates, and personal activities are not accounted for across seasons.
In these cases, study results are only applicable to a specific time period. Longitudinal studies,
on the other hand, generally have small sample sizes and use a purposeful sampling design.
Some studies did not include ambient site measurements to allow comparisons with the exposure
data. Approximately half of these studies monitored PM2 5. Only one or two studies measured
both PM10 and PM2 5 to provide information on PM10_2 5.
Four large-scale probability studies that quantify personal exposure to PM under normal
ambient source conditions have been reported in the literature. These include the EPA's
PTEAM study (Clayton et al., 1993; Ozkaynak et al., 1996a,b); the Toronto, Ontario study
(Clayton et al., 1999a; Pellizzari et al., 1999); the EXPOLIS study (Jantunen et al., 1998, 2000;
Oglesby, et al., 2000; Gotshi et al., 2002; Kousa et al., 2002); and a study of a small, highly
polluted area in Mexico City (Santos-Burgoa et al., 1998). A fifth study conducted in Kuwait
during the last days of the oil-well fires (Al-Raheem et al., 2000) is not reported here because the
ambient PM levels were not representative of normal ambient source conditions.
5-30
-------
TABLE 5-4. SUMMARY OF RECENT PM PERSONAL EXPOSURE STUDIES
Study Design
Study Location and No. of
Population Subjects
Study Period
Age Days per PM Exposure"
(Years) Subject Metrics
PM Size Co-pollutant
Measured' Metrics Reference
Probability Studies
Pooled0
Pooled
Pooled
Pooled
Pooled
Pooled
Indianapolis, IN 240
Riverside, CA, 178
PTEAM
Basel, Switzerland, 50
EXPOLIS
Toronto, Canada 732
Mexico City 66
Baltimore, MD 20 Adults
nonsmoke exposed 21 Children
adults, children, COPD 15 COPD
patients
1996
Fall 1990
1997
Sept 1995-
Aug 1996
1992
Summer 1998,
winter 1999
16-? One 72-h P, I, A, O
sample/
subject
10-70 1 (12 h) P, I, O, A
1 (48 h) P, I, P
16+ 3 P, I, O, A
< 65 1 P, I, O
A 75 12-days P, A
Ch9-13 8-daysfor
COPD 65 children in
summer
PM2 5, PM10 Mn, Al, Ca Pellizzari et al.
(2001)
PM10 Clayton et al.
(1993)
Ozkaynak et al.
(1996a,b)
PM25 VOC, CO, Oglesbyetal.
N02, S, K, Pb, (2000)
Br, Ca Jantunen et al.
(1998)
PM25(12mo) Clayton etal.
PM10 (3 mo) (1999a),
Pellizzari et al.
(1999)
PM10 Santos-Burgoa
et al. (1998)
PM2 5 O3, NO2, SO2, Sarnat et al.
PM10 VOCs, ED/OC, (2001)
CO
Purposeful Studies
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Wageningen, 13
Netherlands, school
children
Amsterdam (Am), 41 (Am)
Helsinki (His), elderly 49 (His)
angina or coronary
heart disease
Baltimore, elderly 21
healthy and COPD
Fresno I 516
Fresno II (elderly)
1995
Winter 1998,
spring 1999
July-Aug 1998
Feb 1999,
Apr-May 1999
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, 0, A
PM2 5, PM10 Janssen et al.
(1999a)
PM25 Janssen et al.
(2000)
PM2 5, PM10 CO, O3, NO2, Williams et al.
SO2 (2000a,b)
PM2 5, PM10 CO, O3 Evans et al.
PM2 5, PM10 (2000)
-------
TABLE 5-4 (cont'd). SUMMARY OF RECENT PM PERSONAL EXPOSURE STUDIES
to
Study
Design
Study Location and
Population
No. of
Subjects
Age Days per PM Exposure" PM Size
Study Period (Years) Subject Metrics Measured'
Co-pollutant
Metrics Reference
Purposeful Studies (cont'd)
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Pooled
Pooled
Longitudinal
Longitudinal
|_iE diary
Los Angeles, elderly
COPD subjects
Boston, COPD subjects
Vancouver, British
Columbia, COPD
Amsterdam and
Wageningen, Neth.,
school children
Amsterdam, adults
Baltimore, elderly
Baltimore, elderly,
COPD, children
Tokyo, Japan,
elderly housewives
Osaka, Japan
Milan, Italy, office
workers
Banska Bystrica,
Slovakia
Wageningen, NL
Mpala, Kenya
30
18
16
45
37
15
56
18
26
100
49
13
252
Summer/fall 56-83 4 P, I, O PM25
1996
Winter 1996- 12 P, I, O, A PM25, PM10
1997,
summer 1996
April-Sept 1998 54-86 7 P, A PM25, PM10
1994-1995 10-12 4-8 P, A, School PM10
1994 51-70 5-8 P, I, A PM10
Summer 1998, 75 ±6. 8 12 P PM25, PM10
spring 1999
Summer 1998, Adults: 75 ± 6.8 12 P, I, O, A PM25
winter 1999 Children: 9-13
COPD: 65 ±6.6
1992 3 P, 1, 0, A SPM
Fall 1990-1995 Multiple days P, I, O PM2, PM2 10,
PM>10
Spring/summer 1 P, Home, PM10
and winter, year Office,
not stated Commuting
1997-1998 15-59 1 P, I, O, A PM10, PM25
Mar- June 1995 12-14 5-8 P, A, I at school PM25, PM10
1996-1998 5-75 2 years I Undefined
Optical MIE
Linn et al.
(1999)
Rojas-Bracho
et al. (2000)
Ebelt et al.
(2000)
Janssen et al.
(1997)
Janssen et al.
(1998a)
O3, NO2, Sarnat et al.
S02 VOCs (2000)
O3, NO2, Sarnat et al.
S02, CO, EC, (2000)
OC, VOC
NO2 Tamura et al.
(1996a)
Tamura et al.
(1996b)
NO2, CO, Carrer et al.
VOCs (1998)
SO42~, Braueretal.
nicotine (2000)
None Janssen et al.
(1999a)
CO Ezzati and
Kammen
(2001)
-------
TABLE 5-4 (cont'd). SUMMARY OF RECENT PM PERSONAL EXPOSURE STUDIES
OJ
Study
Design
Longitudinal
Pooled
Pooled
Pooled
Pooled
Pooled
Study Location and
Population
London, UK
Zurich, CH
Minneapolis/St. Paul,
MN Volunteers
Birmingham,
UK - healthy adults,
children and
susceptibles
Santiago, Chile
children
Copenhagen, DK
non-smoking students
No. of
Subjects
10
10
32
1 1 healthy
adults,
1 8 susceptible
8 in 1998
20 in 1999
68 subjects
Study Period
1997
1998
Spring, summer,
fall 1999
Season and year
not given
Winters
1998 & 1999
Winter 1999,
spring, summer,
fall 2000
Age
(Years)
9-11
Adults
24-64
Adults,
Adults > 65
Child < 10
10-12
20-33
Days per PM Exposure"
Subject Metrics
5 day/season P, I, O
3 seasons
12h/dayfor P, I, O
3 days
2-15 P, I, 0
Adults and P, I, O
children 10,
susceptibles
5, daytime
only
5 P, I, O
2 P, A
PM Size
Measured'
PM2.5, PM10
Pollen
PM25
PM10
PM2.5, PM10
PM2 5, BS
(from PM2 5
filter)
Co-pollutant
Metrics Reference
Wheeler et al.
(2000)
Riediker et al.
(2000)
Adgate et al.
(2002)
CO, NO2 Harrison et al.
(2002)
NO2, 03 Rojas-Bracho
et al. (2002)
S0rensen
et al. (2003)
a All based on gravimetric measurements.
P = personal, I = indoors, O = outdoors, A = ambient.
Pooled; data from many subjects with only a few days per subject.
Longitudinal; one subject measured for many days.
-------
Recent longitudinal exposure studies have focused on potentially susceptible
subpopulations such as the young and elderly with preexisting respiratory and heart diseases
(hypertension, COPD, and congestive heart disease). This is in keeping with epidemiological
studies that indicate mortality associated with high levels of ambient PM25 is greatest for elderly
people with cardiopulmonary disease (U.S. Environmental Protection Agency, 1996).
Longitudinal studies were conducted in the Netherlands by Janssen (1998) and Janssen et al.
(1997, 1998a,b, 1999b,c) on purposefully selected samples of adults (50 to 70 years old) and
children (10 to 12 years old). School children have also been studied in Chile (Rojas-Bracho
et al., 2002). Several additional studies have focused on nonsmoking elderly populations in
Amsterdam and Helsinki (Janssen et al., 2000), Tokyo (Tamura et al., 1996a), Baltimore, MD
(Liao et al., 1999; Williams et al., 2000a,b,c), and Fresno, CA (Evans et al., 2000). These
cohorts were selected because of the low incidence of indoor sources of PM (such as combustion
or cooking). This should allow an examination of the relationship between personal and ambient
PM concentrations without the large influences caused by smoking, cooking, and other indoor
particle-generating activities. The EPA has a research program focused on understanding PM
exposure characteristics and relationships. Within the program, longitudinal studies are being
conducted on elderly participants with underlying heart and lung disease (COPD, patients with
cardiac defibrillator, and myocardial infarction), an elderly environmental justice cohort, and
asthmatics. These studies are being conducted in several cities throughout the United States and
over several seasons (Rodes et al., 2001; Conner et al., 2001; Landis et al., 2001; Rea et al.,
2001).
A series of studies by Phillips et al. (1994, 1996, 1997a,b, 1998a,b, 1999) examined
personal ETS exposure in several European cities. Participants varied by age and occupation.
Respirable suspended particulate (RSP) concentrations were reported. These studies are not
included in Table 5-4 because of their focus on ETS exposure (which is not the focus of this
chapter). A small personal exposure study in Zurich, Switzerland was reported by Monn et al.
(1997) for PM10. This study also is not listed in Table 5-4 because indoor and outdoor
measurements were not taken simultaneously with the personal measurements and other details
of the study were not published.
5-34
-------
5.3.2.2 Microenvironmental Data
Usually, personal PM monitoring is conducted using integrated measurements over a
12- or 24-h period. As such, total PM exposure estimates based on PEM measurements do not
capture data from individual microenvironments. Recent studies have examined PM
concentrations in various microenvironments using a number of different types of instruments
ranging from filter-based to continuous particle monitors. Details on the instruments used,
measurements collected, and findings of these studies according to microenvironment
(residential indoor, nonresidential indoor, and traffic-related) are summarized in Table 5-5.
Those studies which collected microenvironmental data as part of a personal exposure
monitoring study are summarized in Table 5-4 above. In general, the studies listed in Table 5-5
are relatively small, purposeful studies designed to provide specific data on the factors that affect
microenvironmental concentration of PM from both ambient and nonambient sources.
Recently published studies have used various types of continuous monitors to examine
particle concentrations in specific microenvironments and resulting from specific activities.
Continuous particle monitors such as the scanning mobility particle sizer (SMPS), aerodynamic
particle sizer (APS), and Climet have been used to measure particle size distributions in
residential microenvironments (Abt et al., 2000a; Long et al., 2000, 2001a; Wallace et al., 1997;
Wallace, 2000a; McBride et al., 1999; Vette et al., 2001; Wallace and Howard-Reed, 2002).
These studies have been able to assess the infiltration factor for ambient particles to indoor
microenvironments, as well as penetration factors and deposition rates. Continuous instruments
are also a valuable tool for assessing the impact of particle resuspension caused by human
activity. A semi-quantitative estimate of PM exposure can be obtained using personal
nephelometers that measure PM using light-scattering techniques. Recent PM exposure studies
have used condensation nuclei counters (1-s averaging time) and personal nephelometers (1-min
averaging time) to measure PM continuously (Howard-Reed et al., 2000; Quintana et al., 2000;
Magari et al., 2002; Lanki et al., 2002) in various microenvironments. These data have been
used to identify the most important ambient and nonambient sources of PM, to provide an
estimate of source strength, and to compare modeled time activity data and PEM 24-h mass data
to nephelometer measurements (Rea et al., 2001).
5-35
-------
TABLE 5-5. SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
Reference
Study Description
Instrument(s)
Size Fraction (urn)
Summary of Measurements
Notes
Residential Indoor: Nonsmoking Homes
Abt et al.
(2000a)
Boston, MA
Long et al.
(2000a)
Boston, MA
Anuszewski
etal. (1998)
Seattle, WA
Leaderer et al.
(1999)
Southwest VA
Wallace et al.
(1997); Wallace
(2000b)
Reston, VA
Howard-Reed
et al. (2000)
Fresno, CA
Baltimore, MD
2 homes,
2 seasons,
6 days
9 homes,
2 seasons
9 homes,
18 days
58 homes, summer
SMPS
APS
SMPS
APS
Nephelometer (Radiance)
0.02-10 urn
1 home,
4 years
15 participants
SMPS
Climet
PAHs
Black carbon
Nephelometer (personal
MIE)
PEM
PM10
PM2,
6 size bins;
100 size channels
0.01-0.4 urn
0.1-10 urn
PM,,
Detailed traces of PM in various size classes for
different air exchange rates (< Ih"1 to > 2 h"1).
Continuous PM distributions and size
distributions obtained for indoor and outdoor
air using SMPS and APS monitors.
Simultaneous indoor and outdoor PM measured
continuously; 1-h avgtime, I/O = 0.98; air
exchange rate: 0.7-1.7 h"1.
24 h mean:
Regional air 26.0 ±11.5 ng/m3 (n = 47);
Outdoor homes 28.0 ± 17.7 ug/m3 (n = 43);
Indoor w/ AC 28.9 ± 18.7 ug/m3 (n = 49);
Indoor w/o AC 33.3 ± 14.2 ug/m3 (n = 8).
24 h mean:
Regional air 20.2 ± 9.9 ug/m3 (n = 50);
Outdoor homes 21.8 ± 14.8 ug/m3 (n = 43);
Indoor w/ AC 18.7 ± 13.2 ug/m3 (n = 49);
Indoor w/o AC 21.1 ±7.5 ug/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.
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 generated by
indoor events were ultrafine particles by
volume.
Homes contained asthmatic children,
heavy wood burning. Dominant source of
fine particles was outdoor air.
Epidemiologic 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-um particles linked to outdoor
concentrations, frying, broiling; 0.5- to
2.5-um particles related to cooking
events; > 2.5-um 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.
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
Reference
Study Description
Instrument(s)
Size Fraction (um)
Summary of Measurements
Notes
Residential Indoor: Nonsmoking Homes (cont'd)
Rea et al. (2001) 15 participants
Baltimore, MD
Fresno, CA
Quintana et al.
(2000)
San Diego, CA
Chang et al.
(2000)
Baltimore, MD
Lioy et al.
(1999)
Ezzati and
Kammen (2001)
Mpala, Kenya
Asthmatic children indoor
and outdoor
9 homes
1 person performing
predetermined activities
10 vacuum cleaners
55 Native huts
2-years
Chao and Tung 5 unoccupied homes
(2001) measured indoors and
Hong Kong outdoors, along with air
exchange rates
Nephelometer (personal
MIE)
PEM
Nephelometer (personal
MIE)
Harvard impactors
TEOM
"Roll-around" monitor
(RAS)
(PM25, CO, VOC, 03N02
SO2)
MiniRam (MIE)
DustTrak (TSI)
0.1-10 um
PM25 and PM10
0.1-10 urn
PM25 and PM10
PM2,
0.3-0.5 urn
Not specified.
Optical device detects
particles 1-10 um, but it
is not PM10
PM2 5 real time,
calibrated against an
Andersen Mark II
54 ±31% of average daily PM25 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.
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.
One-hour personal O3 exposures were
significantly lower in indoor than outdoor
microenvironments. One-hour personal CO
exposures were highest in vehicles. Personal
and ambient PM2 5 correlations were strongest
for outdoor microenvironments and those with
high air exchange rates (i.e., vehicles).
Vacuum cleaners ranged in collection efficiency
from 29-99%. Substantial fine particle
emissions from motors with emission rates from
0.028-128.8 ng/min.
Measured PM surrounding wood fires in
unvented 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 residual 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.
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 and outdoor measurements
collected using passive, active, and active
heated nephelometers for comparison to
PM mass measurements.
One-hour personal exposures measured
simultaneously. Personal and ambient
concentrations were compared.
Exposures were related to ARI.
Developed an excellent model for ambient
PM infiltration in the absence
of anthropogenic indoor sources.
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
Reference
Study Description
Instrument(s)
Size Fraction (um)
Summary of Measurements
Notes
Residential Indoor: Nonsmoking Homes (cont'd)
Fischer et al.
(2000)
Amsterdam, NL
Measured traffic related
differences of PM and
VOCs, indoor/outdoor in 18
paired homes at varying
distances from traffic
Harvard Impactors
PM2 5 and PM10. EC was
measured by reflectance
of the PM2 5 filters. PAH
also measured as
indicator of diesel
traffic.
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.
"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.
300 m from traffic
Measured PM indoors and
outdoors at 16 homes while
residents were absent.
Air exchange rate estimated,
not measured
Measured ultrafine PM in
various home and traffic
microenvironments
Comparison of indoor and
outdoor PM in homes in
Coachella Valley, CA
Measured Black Smoke
(BS) by reflection from
PM2 5 filters for 4 European
cities, indoor and outdoor.
Measured PM2 5 for sulfate
analysis in and out of
6 Boston homes
Measured ultrafine, fine and
coarse PM in one Reston,
VA tri-level home for
18 months with air
exchange and meteorologic
data
Harvard Impactors
PM25andPM10and
PAH. EC measured by
filter reflectance.
Scanning mobility particle Submicron PM,
sizer, aerodynamic particle supramicron PM,
sizer, and a TSI Dust Irak PM2 5
TSI 8525
1 |_im optical diameter
USC Personal PM 5 Lpm PM25 and PM10.25
XPOLIS PM monitor and
EEL 43 reflectometer.
Harvard Impactors with
SMPS and APS particle
counters
TSI 3071 SMPS
TSI 3320 APS
Climet 500-1
MIEpDR-1000
PM,
PM,
lOnmto > 10 |-im
Median ratio = 1 (no indoor combustion
sources).
For supra- and submicron particles, indoor =
outdoor for normal ventilation conditions of > 2
air changes/hour.
1-second readings collected while noting
presence of sources.
12.5 = 4.3 + 0.74 O
110-2.5 = 3 + 0.30
I EC = 0.840 EC
I OC> O OC.
PM measured I and O for two nights while
subject was home (5 pm - 8 am). 183 sampled
PM25 and BS were compared for these samples.
The ratio of I/O PM2 5, sulfate and size
sub-fractions are developed as a function of
season and air exchange rate.
Analysis of particle counts as a function of air
exchange rate and meteorological variables.
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.
The TSI 8525 is a useful instrument for
screening UFPM in microenvironments.
Thirteen volunteers had two to four
23-h measurements I and O analyzed for
elemental carbon (EC) and organic carbon
(OC). Some unoccupied homes
measured.
BS is a supplementary measurement that
can be made on PM2 5 filters as an
indication of EC.
Sulfur is primarily of outdoor origin and
can be used to track ambient PM of
similar (0.06-0.5 |_im) AD, but different
relations exist for ultrafine PM and
coarse-mode PM < 2.5 |_im AD.
Wind speed has little influence on AER.
The home sometimes acts as
1 compartment and sometimes as multiple
compartments.
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
Reference
Study Description
Instrument(s)
Size Fraction (um)
Summary of Measurements
Notes
Residential Indoor:
Brauer et al.
(1996)
Mexico
Jenkins et al.
(1996 a,b)
16 U.S. Cities
McBride et al.
(1999)
NA
Vette et al.
(2001)
Fresno, CA
Douce et al.
(2001)
Other Home Types
22 rural Mexican homes
(smoking and nonsmoking)
Smoking and nonsmoking
homes
Combustion source
(incense) and walking
(1 room, carpeted)
Detached
semioccupied residence
ETS measured by
4 methods in smokers
homes and offices
Inertial impactor PM10
Radiance
nephelometer PM25
Fluoropore membrane Particle phase ETS
filters markers
Met-One laser particle
counter
SMPS 0.01-2.5 urn
LASX
37 mm filter and XAD-4 PM5
resin backup
Indoor PM25: 132-555 ug/m3,
PM10: 282-768 ug/m3.
Outdoor PM25: 37 ug/m3.
PM10: 68ng/m3;I/OPM25: 1.8-12.4;
PM10: 4.7-10.0.
Mean PM3 5 concentrations were 17-20 ug/m3
in smoking homes over nonsmoking homes.
Ratios of particle counts at 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.
Samples extracted and analyzed for UV
absorption, fluorescence, solanesol, and
scopoletin.
Variety of cooking fuels used.
Nephelometer data were highly correlated
with PM25 and PM10 indoors (r = »0.87-
0.95).
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.
Solanesol is best method of 4, but needs
study of solanesol stability on filter.
Nonresidential Microenvironments
Bohadana et al.
(2000)
Donham et al.
(2000)
San Francisco,
CA
Klepeis et al.
(1996)
San Francisco,
CA
Nieuwenhuijsen
etal. (1999)
Teschke et al.
(1999)
Manufacturing plant,
woodworkers
34 poultry workers
Airport lounge, ETS
Agricultural activities
Wood production,
wood finishing,
wood construction workers
Not given
NIOSH Method 0600 PM5
monitors
probed respirators
TSI 8510 piezobalance PM35
PM4
PM -~50
443 personal time-weighted average
occupations samples of airborne dust.
Total dust sampled indoor respiratory masks.
Personal monitoring: 630 ± 980 ng/m3
(n = 210) ranging from 10-7730 ng/m3.
Estimated cigarette emission rate of
1.43 mg/min/cigarette.
Average respirable fraction: 4.5 mg/m3.
1,632 observations from 1979-1997.
Arithmetic mean exposure: 7.93 mg/m3.
Geometric mean exposure: 1.86 mg/m3.
Respirable dust constituted about 10% of
total dust measured.
Personal exposures to ETS can be
modeled in these types of
microenvironments.
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
Reference
Study Description
Instrument(s)
Size Fraction (um)
Summary of Measurements
Notes
40% nonsmoking) PM2 5
11-163 ug/m3; PM10 24-89 ug/m3
Unrestricted smoking: PM25 47-253 ug/m3;
PM10 51-268 ug/m3.
Indoor PM10: 30-470 ug/m3
Outdoor PM10: 20-617 ug/m3
Daily 9 a.m - 6 p.m./weekday for 9 months.
Measured inside and outside the various
microenvironments tested, weekdays morning
and afternoon. At least 3 visits, several 10-min
averages.
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.
Temp, RH and solar intensity influence
In/Out. Wind speed has no effect.
10-min outdoor averages are poor
predictors of indoor values.
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
Reference
Study Description
Instrument(s)
Size Fraction (um)
Summary of Measurements
Notes
Traffic-Related Microenvironments (TRMs)
Praml and
Schierl (2000)
Munich,
Germany
Monn et al.
(1997)
Switzerland
Trams and buses, rural and
urban
Spatial scale from a city
street
Rodes et al. In-vehicle, various road
(1998) types, 2-h trips
Saevanenso,
Los Angeles, CA
Continuous millipore PM10
polycarbonate filter
Harvard PM10
microenvironment monitor
PM25
PM10
Gradient in distance from
roadway
Harvard impactor
Roorda-Knape
etal. (1998)
van Vliet et al.
(1997)
Netherlands
Houseman et al. Indoor and outdoor vehicles TSI DustTrak
(2002) buses, subways
Boston, MA
Brauer et al. Commuting environments APC-1000
(1999)
Vancover, BC
Janssen et al.
(1997)
Netherlands
Background and roadway
PM25
PM10
Black smoke
PM10
PM25
PM10
n = 201 4-h trips, mean concentration
155 ug/m3 range: 13-686 ug/m3
I/O: 2.8.
48- or 72-h avg times; horizontal distance from
street: 0, 15, 50, and 80 m; vertical distance
from street: 20 m.
Mean PM10 27.3 ± 3.0 ug/m3.
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.
PM monitoring at 50, 100, 150, and 300 m
from roadway; 1-week avg time.
Vehicle concentrations ranged from
33-170 ng/m3. Outdoor vehicle concentrations
ranged from 40-144 ug/m3.
Bus concentrations: 17-268 ug/m3;
outdoor 10-203 ug/m3.
Subway: 28-174 ug/m3; outdoor 8-203 ug/m3.
PM < 5: greatest concentrations by
combustion powered vehicles.
PM > 5: greatest concentrations by
bicycling and buses.
PM25 background: 21-35 ug/m3; roadway
23-43 ug/m3.
PM10 background: 13-32 and 29-62 ug/m3;
roadway 16-56 and 30-75 ug/m3.
Tram > circular bus route > radial bus
route.
Day > night.
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.
Air exchange rates measured at various
ventilation settings and speeds.
Monitoring vehicle followed a diesel bus
or truck.
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 ug/m3.
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
to
Reference
Study Description
Traffic-Related Microenvironments (TRMs) (cont
Adams et al.
(2001)
London, UK
Almetal. (1999)
Kuopio, Finland
Chan et al.
(2002)
Hong Kong
Hoek et al.
(2001)
Hoek et al.
(2002)
Jinsart et al.
(2002)
Bangkok
Lena et al.
(2002)
New York, NY
Zhu et al. (2002)
Los Angeles, CA
PM by volunteers in TRM
in London, UK
9-km commuter route,
rush hours
1/mo
PM measured in TRM in
Hong Kong
BS in Netherlands
interpolated to outdoor
locations relative to traffic
at subject's homes
PM Exposure of traffic
police at intersection post in
Bangkok, Thailand
PM2 5 and EC in the Bronx,
NY in area of high Diesel
truck traffic
Ultrafme near Interstate
Highway in Los Angeles,
CA with heavy Diesel
Instrument(s) Size Fraction (urn)
'd)
16 L/min personal PM2 5
monitor. Porous foam
size selector
Climet 6 channels
TSI 8520 DustTrac PM25 and PM10
Calibrated to Partisol
PM25
HiVol PM10
Black Smoke measured Not Recorded
from filters
Sibata personal single PM2 5 and PM10
nozzle
2.5 L/min
PM25 @ 4 L/min PM25
3 Lpm quartz filter for EC
EEL for BS
CPC TSI 3022A 6 nm-220 nm
SMPS TSI 3936
BC aethalometer
Summary of Measurements
Volunteers rode/cycled along fixed routes
repetitively.
Windows closed, vents open
air exchange rate 36-47 h"1.
Repetitive sampling over 8 fixed routes.
GIS used to interpolate background and regional
ambient BS with added increment for distance
to traffic < 50 m and < 100 m.
Sampled 12-h while on duty at post at
intersection.
Sampled 10-h at sidewalk locations while
counting cars and trucks for 3 -weeks in
summer of 1999.
200 m upwind and 17, 20, 30, 90, 150, and
300 m downwind.
Notes
Personal exposures were generally double
those at fixed-site ambient station.
Subway exposures are maximal.
Morning commutes were generally higher
than afternoon commutes; relationships
determined between PM and wind speed
and vehicle speed.
Highly variable by mode.
Tram exposures are maximal.
May be useful technique. Estimates not
validated with BS measurements at
interpolated loci.
Exposures of both PM25 and PM10 higher
than ambient PM measured at station with
p gauge. Not comparable to U.S.
conditions.
EC is a large component of diesel PM2 5
and varies with truck traffic.
Ultrafme PM decreased exponentially
from the freeway and was equal to upwind
at 300 m.
traffic
-------
5.3.2.3 Traffic-Related Microenvironments
There has been increasing interest in the possible role of traffic-related pollutants.
Distance to roadways has been used as a surrogate for exposure to traffic-related pollutants
(Hoek et al., 2001), and this exposure indicator was subsequently used in an epidemiological
study (Hoek et al., 2002). A traffic model, using traffic volume, direct exhaust emissions rate,
and a reentrainment rate has been used to estimate concentrations of traffic-related emissions at
several schools in east Los Angeles (Korenstein and Piazza, 2002). Personal exposure studies
have been made in a variety of commuting situations including vehicle traffic (Adams et al.,
2001; Chan et al., 2002). Other studies have measured various indicators of traffic near
roadways (Lena et al., 2002), inside vehicles in traffic (Abraham et al., 2002), and in several
types of traffic-related microenvironments (Levy et al., 2002). Table 5-5 above provides a brief
description of these studies, instruments used, measurements made, and key findings.
5.3.2.4 Reanalyses of Previously-Reported Particulate Matter Exposure Data
Papers that have reanalyzed and interpreted the data collected in previous PM exposure
studies are summarized in Table 5-6. These reanalyses are directed toward understanding the
personal cloud, the variability in total PM exposure, and the personal exposure-to-ambient
concentration relationships for PM. Brown and Paxton (1998) determined that the high
variability in personal exposure to PM makes the personal-to-ambient PM relationship difficult
to predict. Wallace (2000b) used data from a number of studies to test two hypotheses: elderly
COPD patients have (1) smaller personal clouds and (2) higher correlations between personal
exposure and ambient concentrations compared to healthy elderly, children, and the general
population. The analysis by Wallace (2000a) and three subsequent longitudinal studies
(Williams 2000a,b,c; Ebelt et al., 2000; Sarnat et al., 2000) supported hypothesis 1 but not
hypothesis 2. Ozkaynak and Spengler (1996) showed that at least 50% of personal PM10
exposure for the general population comes from ambient particles. Wilson and Suh (1997)
concluded that fine and coarse particles should be treated as separate classes of pollutants
because of differences in characteristics and potential health effects. Wilson et al. (2000) gave a
review of what they call the "exposure paradox" and determined that personal PM needs to be
divided into different classes according to source type and that correlations between personal and
ambient PM will be higher when nonambient sources of PM are removed from the personal PM
5-43
-------
TABLE 5-6. PAPERS REPORTING REANALYSES OF PARTICIPATE MATTER EXPOSURE STUDIES
Reference
Study Cited
Objectives/Hypotheses
Findings
Wallace PTEAM (Ozkaynak et al., 1990;
(2000a) Spengler et al., 1989; Wiener 1988, 1989;
Wiener et al., 1990)
THEES (Lioy et al., 1990)
Amsterdam COPD (Janssen et al., 1997,
1998a)
Boston COPD (Rojas-Bracho et al., 2000)
Ozkaynak
and Spengler
(1996)
Dockery and Spengler (1981)
PTEAM (Ozkaynak et al., 1996a,b)
Netherlands (Janssen et al., 1995)
Brown and THEES (Lioy et al., 1990)
Paxton PTEAM pilot (Wallace, 1996)
(1998) Boston and Nashville COPD (Rojas-Bracho
et al., 2000); Bahadori et al., 1998)
Wilson and Philadelphia (Burton et al., 1996;
Suh (1997) Suggs and Burton, 1983)
EPA AIRS database
Wilson et al. New Jersey (Lioy et al., 1990)
(2000) Japan (Tamura et al., 1996a)
PTEAM (Clayton et al., 1993;
Ozkaynak etal., 1996a,b)
Netherlands (Janssen, 1998a;
Suh etal., 1992)
Examines the differences between pooled and longitudinal
correlations in personal and ambient (or outdoor) data for
PM25 and PM10.
Discusses the personal cloud for PM2 5 and PM10.
Hypothesizes that COPD patients have (1) smaller personal
clouds (supported) and (2) higher correlations of personal
exposure with outdoor concentrations because of reduced
mobility (not supported).
Uses statistical modeling techniques to examine the
relationship between ambient PM concentrations and personal
exposures. Data analysis involves use of air exchange rates,
penetration factors, and June 25, 2004 ratios, as well as
examining exposure in various microenvironments (traveling,
working, outdoors, indoors) activities (exposure to smoke,
cooking), and source strengths.
Cross-sectional and longitudinal regression analysis on
data sets.
Determines the utility of fine and coarse PM concentrations as
indicators of time-series epidemiology with regard to day-to-
day variability, area uniformity, and June 25, 2004 PM ratios.
Necessary to treat personal exposure to ambient PM and
personal exposure to nonambient PM as separate components
of total personal PM exposure.
Synoptic review of the "exposure paradox": low correlations
between personal exposure and ambient PM concentrations in
spite of the existence of statistical association between ambient
PM and epidemiologic health effects.
Uses personal exposure equation, mass balance, regression
analysis, and deductive logic.
Median longitudinal correlation coefficient is much higher than
the pooled correlation coefficient for the same data sets.
Personal cloud for PM10: 3-67 ug/m3; PM2 5 6-27 ug/m3.
Personal cloud for elderly COPD was much smaller (PM10:
6-11 ng/m3; PM2 5 " 6 g/m3) than for other healthy populations
(PM10: 27-56 ug/m3; PM25: 11-27 ug/m3) of elderly, children,
and the general population. However, correlations of personal
exposure with ambient concentrations were not higher for
elderly COPD than for other groups.
The important components of personal exposures are received
during contact with indoor sources, mainly in homes and work
places.
Ambient aerosols contribute about 50% or more to the personal
PM10 exposures of the general population.
The contribution of ambient aerosols to the total toxicity of
inhaled particles is significant.
Individual personal PM exposure is subject to high variability,
which makes the personal-to-ambient PM relationship difficult
to predict.
Fine and coarse particles should be considered separate classes
of pollutants.
Fixed-site ambient fine-particle measurements likely give a
reasonable indication of the variability in the concentration of
ambient fine particles across the community. Coarse-particle
measurements most likely will not.
Personal PM exposure needs to be divided into different
classes according to source type: exposure to ambient PM
(outdoor and indoors) and exposure to nonambient PM (indoor
source and personal activity).
Correlations are higher between personal exposure and ambient
PM concentrations when PM exposures from nonambient
sources are removed.
-------
TABLE 5-6 (cont'd). PAPERS REPORTING REANALYSES OF PARTICIPATE MATTER EXPOSURE STUDIES
Reference
Study Cited
Objectives/Hypotheses
Findings
Mage et al. Japan (Tamura et al., 1996a)
(1999) State College (Suh et al., 1995)
Netherlands (Janssen et al., 1997, 1998a,
1999a)
New Jersey (Lioy et al., 1990)
PTEAM (Clayton et al., 1993;
Ozkaynaketal., 1996a,b)
Mage (1998) PTEAM (Clayton et al., 1993;
Ozkaynaketal., 1993, 1996a,b)
Monn (2001) Multiple Literature Review
Rotko et al.
(2000a)
Rotko et al.
(2000b)
Rodes et al.
(2001)
Jantunen et al. (1998)
Carrer et al. (1997)
Koistinen et al. (1999)
Rotko et al. (2000a),
Jantunen et al. (1998)
EPA Baltimore and Fresno 1 and 2
Examines the influence of nonambient PM on total PM
concentrations and how it may confound the outdoor/personal
PM relationship. Missing data and outlier values created using
an algorithm. Linear regression analysis of subsequent data
sets.
Uses a reduced-form mass-balance model to predict the
average fraction of ambient PM to which the average person
is exposed.
Objective review of literature published since 1996 as an
implicit update to the 1996 U.S. EPA PM AQCD. Emphasis
on European studies.
Compares exposure relationships between the six EXPOLIS
European cities (Athens, Basel, Grenoble, Helsinki, Milan,
Prague).
Determines sociodemographic influences of exposure in
Helsinki.
Investigates relationships between the different retirement
centers and identify most likely factors influencing personal
and indoor concentrations.
Variation in daily personal exposure for subjects with similar
lifestyles and no ETS exposure are driven by variations in
ambient PM concentrations.
Exposure to ambient PM is highly correlated in time with
ambient PM concentrations measured at a community site.
Indoor PM does not confound the relationship between daily
mortality and ambient PM.
On average, a person is exposed to > 75% of ambient PM2 5 and
> 64% of ambient PM10 measured by the community monitor.
"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" (Dockery 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 PM2 5
exposures than higher (professional) occupational status.
Mean personal exposure PM2 5 was higher than their apartment
concentrations. Personal cloud of 3 |_ig/m3 for PM25 was
negligible but cloud for PM10 was 20 |_ig /m3. Indoor PM2 5
data were less than ambient concentrations.
-------
concentration. Mage (1998) conducted analysis using the PTEAM data and showed that the
average person in PTEAM (Riverside, C A in the fall) was exposed to > 75% of ambient PM2 5
and > 64% of ambient PM10. Mage et al. (1999) used an algorithm to fill in missing data and
outliers to analyze data sets and showed that variation in daily personal exposures for subjects
with similar activity patterns and no ETS exposure are driven by variation in ambient PM
concentrations.
5.3.3 Factors Influencing and Key Findings on Particulate Matter
Exposures
5.3.3.1 Relationship of Personal/Microenvironmental Particulate Matter with Ambient
Particulate Matter
Understanding the relationship between ambient site measurements and personal exposure
to PM is important for several reasons. First, it allows us to examine the extent to which
ambient measurements for PM and various PM constituents can serve as valid surrogates for
exposure to ambient PM or ambient constituents of PM in epidemiological studies. Second, it
provides information that may improve surrogate exposure measurements and, hence, increase
the power of epidemiologic studies. Finally, because compliance with the NAAQS is based on
ambient monitoring, it can be used to understand the effect of regulation on exposures to PM and
its constituents and can, therefore, help link the effect of regulations to health outcomes. Many
of the studies summarized in Table 5-4 have analyzed this relationship using measurements of
personal PM exposures and ambient PM concentrations. Of main interest are the PM
concentrations measured in ambient, indoor, and outdoor air; personal exposure measurements;
the statistical correlations between measurements; and the attenuation and/or infiltration factors
developed for personal exposure and indoor microenvironments. Attenuation and infiltration
factors are discussed in Section 5.3.4.3.1. Information on correlation analysis is provided below.
5.3.3.1.1 Types of Correlations
The three types of correlation data that will be discussed in this section are longitudinal,
"pooled," and daily-average correlations. Longitudinal correlations are calculated when data
from a study includes measurements over multiple days for each subject (longitudinal study
design). Longitudinal correlations describe the temporal relationship between daily personal PM
exposure or microenvironment concentration and daily ambient PM concentration for each
5-46
-------
individual subject. The longitudinal correlation coefficient, r, may differ for each subject.
An analysis of the variability in r across subjects can be performed with this type of data.
Typically, the median r is reported along with the range across subjects in the study. Pooled
correlations are calculated when a study involves one or only a few measurements per subject
and when different subjects are studied on subsequent days. Pooled correlations combine
individual subject/individual day data for the correlation calculation. Pooled correlations
describe the relationship between daily personal PM exposure and daily ambient PM
concentration across all subjects in the study. For some studies, the multiple days of
measurements for each subject were assumed to be independent (after autocorrelation and
sensitivity analysis) and combined together in the correlation calculation (Ebelt et al., 2000).
Daily-average correlations are calculated by averaging exposure across subjects for each day.
Daily-average correlations then describe the relationship between the daily average exposure and
daily ambient PM concentration. The term cross-sectional is used to refer to both pooled and
daily average correlations, so the meaning of this term must be determined from context.
Pooled correlations have been simulated from longitudinal data by using a random-
sampling procedure to select a random day from each subject's measurements for use in the
correlation. This procedure was repeated many times, and statistics (such as the mean and
standard deviation of the pooled correlation coefficient) were reported (Janssen et al., 1997,
1998a, 1999c).
The type of correlation analysis can have a substantial effect on the resultant r value. Mage
et al. (1999) mathematically demonstrated that very low correlations between personal exposure
and ambient concentrations could be obtained when people with very different nonambient
exposures are pooled, even though their individual longitudinal correlations are high. The
longitudinal studies conducted by Tamura et al. (1996a) and Janssen et al. (1997, 1998a, 1999c)
determined that the longitudinal correlations between personal exposure and ambient PM
concentrations were higher than the correlations obtained from a pooled data set. Wallace
(2000a) reviewed a number of longitudinal studies and found that the median longitudinal
correlation coefficient was higher than the pooled correlation coefficient for the same data (see
Tables 1 and 2, Wallace, 2000a).
Mage et al. (1999) examined three longitudinal exposure data sets where several subjects
were measured each day. They showed that by averaging daily exposures across subjects, daily-
5-47
-------
average correlations could be obtained. These were all higher than the median longitudinal
correlations. Williams et al. (2000a,b) and Evans et al. (2000) have also reported higher
correlation coefficients for daily-average correlations compared to longitudinal correlations. The
higher correlations found between daily-average personal exposures and ambient PM
concentrations, as opposed to lower correlations found between individual exposures and
ambient PM levels, have been attributed to the statistical process of averaging (Ott et al., 2000).
Personal exposures include contributions from nonambient as well as ambient PM
concentrations. When several subjects are measured on the same day, the mean variability due
to variations in nonambient exposures is reduced due to averaging. Therefore, the correlation
between personal exposure and ambient concentrations increases as the number of subjects
measured daily increases. Ott et al. (2000), using the theory on which their Random Component
Superposition (RCS) model is based, predict expected correlations above 0.9 for the PTEAM
study and above 0.70 for the New Jersey study (Lioy et al., 1990) if 25 subjects had been
measured daily in each study.
5.3.3.1.2 Correlation Data from Personal Exposure Studies
Measurement data and correlation coefficients for the personal exposure studies described
in Section 5.4.2.1 are summarized in Table 5-7. All data are based on mass measurements. The
studies are grouped by the type of study design, longitudinal or pooled. For each study in
Table 5-7, summary statistics for the total personal PM exposure measurements are presented
as well as statistics for residential indoor, residential outdoor, and ambient PM concentrations
when available. The correlation coefficients, r, between total personal PM exposures and
ambient PM concentrations also are presented and classified as longitudinal or pooled
correlations. When reported, p-values for the correlation coefficients are included. Correlation
coefficients between personal, indoor, outdoor, and ambient also are reported when available.
5.3.3.1.3 Correlations Between Personal Exposures, Indoor, Outdoor, and
Ambient Measurements
Longitudinal and pooled correlations between personal exposure and ambient or outdoor
PM concentrations varied considerably between study and study subjects. Most studies report
longitudinal correlation coefficients that range from < 0 to «1, indicating that an individual's
5-48
-------
TABLE 5-7. PERSONAL MONITORING STUDIES FOR PARTICIPATE MATTER: MEASURED CONCENTRATIONS
AND CORRELATION COEFFICIENTS
Measured Concentration Levels (jig/m3)
Size Avg. Sample
Fraction Time Statistic Size1
Residential
Personal Indoor
Residential
Outdoor Ambient
Type3
Personal-Ambient 2
Correlation Coefficients (r)
Value (Range)
Notes
Other Correlation
Coefficients (r)
Value
Type3 (Range)
Longitudinal Studies
Ebelt et al. (2000) - Vancouver, BC
PM25
24 h x±SD 106
Range
18.2 ± 14.6
2-91
11.4±4.1
4-29
Median L
P
0.48 (-0.68-0.83)
0.15
n= 16 COPD subjects
Evans et al. (2000) -Fresno, CA
PM25
PM,5
Janssen et al.
PM10
Janssen et al.
PM10
Janssen et al.
PM2.5
PM2.5
PM25
24 h x 24
Range
24 h x 12
Range
(1997) -Netherlands
24 h x±SD 301
Range
(1998a) -Netherlands
24 h x ± SD 262
Range
(1999c) -Netherlands
24 h x ± SD 77
Range
24 h x±SD 55
Range
24 h x ± SD 22
Range
13.3 9.7
1-24 4-17
11.1 8.0
7-16 4-12
105.2 ±28.7
57-195
61.7±18.3 35.0±9.4
38-113 19-65
28.3 ± 11.3
19-60
24.4 ±4.9
19-33
37.0 ± 17.4
21-60
20.5 21.7
4-52 6-37
10.1 8.6
5-20 4-16
38.5 ±5.6
25-56
41. 5 ±4.3
32-50
17.1 ±2.8
14-22
17.1 ±2.6
15-22
17.1 ±3.7
14-21
P
P
Median L
Median L
Median L
MeanP
MeanP
MeanP
Median L
Median P
Median P
Median L
Median P
Median L
Median P
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
0.50 (-0.41-0.92)
0.50 (0.07-0.83)5
0.34(-0.09-0.67)5
0.86 (-0.11-0.99)
0.41 (-0.28-0.93)5
0.92
0.825
Fresno- 1 study
Fresno-2 study
n = 45 school children
With nonsmoking parents
With smoking parents
All
With nonsmoking parents
With smoking parents
n = 37 adults
No ETS exposure
All
n = 13 school children
With nonsmoking parents
With smoking parents
Pp-i 0.814
PP-O 0.804
PP-, 0.954
PP-O 0.80"
Med. 0.72
Lp.; (-0.10-0.98)
Med. 0.73
L;.,, (-0.88-0.95)
-------
TABLE 5-7 (cont'd). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
CORRELATION COEFFICIENTS
Size Avg.
Fraction Time
Janssen et al. (2000)
PM2 5 24 h
Janssen et al. (2000)
PM2 5 24 h
Sample
Statistic Size1
— Netherlands
x±SD 338
Range
— Finland
x±SD 336
Range
Personal-Ambient 2
Measured Concentration Levels (jig/m3) Correlation Coefficients (r)
Residential Residential
Personal Indoor Outdoor Ambient Type3 Value (Range) Notes
24.3 ±25.7 28.6 ±41. 8 20.6 ± 4.0 Median L 0.79 (-0.41-0.98) n = 36 elderly w/CV
9-134 9-239 disease
13-31 Median L 0.85 No ETS exposures
10.8 ±4.4 11.0 ±4.0 12.6 ±2.0 Median L 0.76 (-0.12-0.97) n = 46 elderly w/CV
4-33 3-27 10-18 disease
Other Correlation
Coefficients (r)
Type3
Med.
LP-,
Med.
Med.
LP-,
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
PM2 5 24 h
PM10 24 h
x ± SD 60
Range
x±SD 59
Range
23. 8 ±15.1 23. 5 ±15.3 24.8 ±14.5 P 0.266
4-65 4-92 4-63
34.8 ±14.8 32.6 ±15.6 39.8 ±18.3 33 ± 15 P 0.226
5-85 9-105 7-97 9-??
P™
P™
0.26'
0.47'
0.32'
0.66'
Rojas-Bracho et al. (2000) -Boston
PM25 12 h
PM10 12 h
PM10.2.5 12 h
x ± SD 224
Range
x ± SD 225
Range
x ± SD 222
Range
21.6 ±13.6 17.5±14.1 14.2±11.2 Median L 0.61(0.10-0.93)' n= 17 adults
1-128 2-73 1-57
37.2 ±22.8 31.9 ±25.2 22.2 ±18.7 Median L 0.35(0.0-0.72)'
9-211 2-329 3-76
15.6 ±14.6 14.5 ±9.2 8.1 ±6.8 Median L 0.30(0.0-0.97)'
-11-103 -3-255 -2-64
Med.
LP-,
Med.
L;.0
Med.
LP-,
Med.
L;.0
Med.
LP-,
Med.
L;.0
0.87'
0.74'
0.71'
0.50'
0.42'
0.20'
-------
TABLE 5-7 (cont'd). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
CORRELATION COEFFICIENTS
Measured Concentration Levels (jig/m3)
Size Avg. Sample Residential Residential
Fraction Time Statistic Size1 Personal Indoor Outdoor Ambient
Sarnat et al. (2000) -Baltimore
PM25 24 h x±SD 3736 26.7 ±13.7 25.2 ±11. 5
x±SD 18.5 ±11.2
5.6 ±49.0
PM10 24 h x±SD 3736 33.9 ±11.7 34.0 ±12.8
x±SD 28.0 ±16.5 7.5 ±73.2
PM10.25 24 h x±SD 3736 7.2 ± 4.0 8.4 ±2.3
x±SD 9.6 ±7.9 -1.3 ±24.2
Tamura et al. (1996a) - Tokyo
PM10 48 h
Williams et al. (2000a,b) -Baltimore
PM25 24 h x 23 13.0 9.4 22.0 22.0
Range 7-25 4-19 7-52 8-59
PM10 24 h x 28 11.0 30.0 29.9
Range 4-23 13-66 13-74
PM10.25 24 h x 26 1.0 8.0 8.0
Range -3-5 -2-16 1-15
Personal-Ambient 2
Correlation Coefficients (r)
Type3 Value (Range)
Median L 0.76 (-0.21-0.95)7
Median L 0.25 (-0.38-0.81)7
P 0.898
P 0.758
P 0.508
P 0.448
Median L 0.64 (0.08-0.86)7
Median L 0.53 (-0.79-0.89)7
Median L 0.11 (-0.60-0.64)7
Median L 0.32 (-0.48-0.68)7
P 0.83
Median L 0.80 (0.38-0.98)6
P 0.894
Notes
n= 15 adults,
summer
n= 15 adults,
winter
Ventilation:
High summer
Med. summer
Low summer
Winter
Summer
Winter
Summer
Winter
n = 7 elderly adults
n = 21 elderly adults
Other Correlation
Coefficients (r)
Value
Type3 (Range)
Pp.; 0.90"
PP-O 0.954
Pi-o 0.944
P;.a 0.87"
P;.0 0.82"
P™ O-8!4
poa 0.944
P;-o 0.184
P;.a 0.084
PO-, 0.454
-------
TABLE 5-7 (cont'd). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
CORRELATION COEFFICIENTS
Size Avg.
Fraction Time Statistic
Sample
Size1
Personal-Ambient 2
Measured Concentration Levels (jig/m3) Correlation Coefficients (r)
Residential Residential
Personal Indoor Outdoor Ambient Type3 Value (Range) Notes
Other Correlation
Coefficients (r)
Value
Type3 (Range)
Williams et al. (2000a,b) -Baltimore (cont'd)
Keeler et al. Mean
(2002) (Std)
Detroit, MI
PM1024-h
PM2524-h
Landis et al. Mean
(2001)
Baltimore, MD
PM2524-h
Sulfate24-h
Pooled Studies
Bahadori (1998) -Nashville
PM25 12 h x±SD
Range
PM10 12 h x±SD
Range
Pellizzari et al. (1999) - Toronto
PM2 5 3d x
PM10 3d x
20
asthmati
c
children
10
elderly
retirees
30
30
922
141
68.4(39.2) 52.2 25.8
(30.6) (11.8)
34.4 15.6
(21.7) (8.2)
12.8 10.2 21.0 Median L r = 0.82
(0.51-0.92)
4.5 4.0 10.2
r = 0.95
(0.74-0.97)
21.7 ±10.5 15.5 ±6.6 23.4 ±6.8 P 0.09 n = 10 COPD subjects;
10-67 5-40 3-61 daytime
33.0 ±16.9 21.6 ±10.7 32.5 ±8.1 P -0.08 n = 10 COPD subjects;
5-88 9-77 7-76 daytime
28.4 21.1 15.1 P 0.23 n = 178; n for indoor,
outdoor lower than
personal
67.9 29.8 24.3 No correlations reported
Lp-i r = 0.60
(0.41-0.85)
r = 0.95
(0.73-0.97)
PP-, 0.72
P,-o 0.31
PP-, 0.43
PI-O 0.06
PP-, 0.79
P;-o 0.33
Oglesby et al. (2000) - EXPOLIS Basel
PM2 5 48 h x ± SD
4420
23.7±17.1 19.0±11.7 P 0.07 All
17.5 ±13.0 17.7±7.1 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
P 0.74
P 0.67
PP-, 0.47
P™ 0.23
-------
TABLE 5-7 (cont'd). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
CORRELATION COEFFICIENTS
Size Avg.
Fraction Time Statistic
Measured Concentration Levels (jig/m3)
Sample Residential Residential
Size1 Personal Indoor Outdoor Ambient Type3
Personal-Ambient 2
Correlation Coefficients (r)
Value (Range) Notes
Other Correlation
Coefficients (r)
Value
Type3 (Range)
Pellizzari et al. (2001) — Indianapolis
PM2 5 72 h Median
250 23 18 18 18 P
0.102 Between the Logarithms
of concentrations
PM 0.138
Pp.i 0.923
Brauer et al. (2000) Banska Bystrica
PM25 24 h Mean
PM10
Kousa et al. (2002)
Helsinki
+ 3 cities
PM2 5 48 h
Abbreviations used:
Avg. = Averaging (time)
Cone. = Concentration
CV = Cardiovascular
d = Day
PM10 1.2212e + 10 796655535 354522326 P
summer
PM10
winter
PM2.5
summer
PM25
winter
S04
winter
P
h = Hour
i-a = Indoor-ambient correlation
i-o = Indoor-outdoor correlation
L = Longitudinal correlation
Med. = Median
ETS = Environmental tobacco smoke o-a = Outdoor-ambient correlation
PM10 Multivariate model with
R2 = 0. 17 PM10 and nicotine
0.69 Leisure time, non-ETS
exposed, Helsinki,
log-transformed
P = Pooled correlation
p-i = Personal-indoor correlation
p-o = Personal-outdoor correlation
SD = Standard deviation
Stat. = Statistic
x = Mean
Pp.i R2 = 0.15
so42-
PP.O R2 = 0.23
PP.O 0.65
Pp.; 0.83
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 = V(R2) •
7 Spearman rank correlations.
8 Calculated, r = v (R2), from R2 from a mixed model regression.
-------
activities and residence type may have a significant effect on total personal exposure to PM.
General population studies tend to show lower correlations because of the higher variation in the
levels of PM-generating activities. In contrast, the absence of indoor sources for the populations
in several of the longitudinal studies resulted in high correlations between personal exposure and
ambient PM within subjects over time for these populations. But even for these studies,
correlations varied by individual depending on their activities and the microenvironments that
they occupied.
Probability Studies
In the Toronto study (Pellizzari et al., 1999), pooled correlations were derived for personal,
indoor, outdoor, and fixed-site ambient measurements. This study involved a probability sample
of 732 participants who represented the general population of people 16 years and older. The
study included between 185 and 203 monitoring periods with usable PM data for personal,
residential indoor, and outdoor measurements. For PM10 measurements, the mean concentrations
were 67.9 |ig/m3 for personal, 29.8 |ig/m3 for indoor, and 24.3 |ig/m3 for outdoor air samples.
For PM25, the mean concentrations were 28.4 |ig/m3 for personal, 21.1 |ig/m3 for indoor air, and
15.1 |ig/m3 for outdoor air samples. A low, but significant, correlation (r = 0.23, p < 0.01) was
reported between personal exposure and ambient measurements. The correlations between
indoor concentrations and the various outdoor measurements of PM25 ranged from 0.21 to 0.33.
The highest correlations were for outdoor measurements at the residences with the ambient
measurements made at the roof site (0.88) and the other fixed site (0.82). Pellizzari et al. (1999)
state that much of the difference among the data for personal/indoor/outdoor PM:
". . . can be attributed to tobacco smoking, since all variables reflecting
smoking . . . were found to be highly correlated with the personal (and indoor)
particulate matter levels, relative to other variables that were measured . . . none
of the outdoor concentration data types (residential or otherwise) can adequately
predict personal exposures to particulate matter." (p. 729)
Using a Random Component Superposition (RCS) statistical model, Ott et al. (2000)
calculated an attenuation factor of 0.61 for personal exposure for PM10 for the Toronto study.
The mean nonambient exposure component for PM10 was estimated as 53 |ig/m3 with a standard
5-54
-------
deviation of 85 |ig/m3. Although the data were available for PM2 5, similar calculations were
not made.
PM10 data from the PTEAM study were analyzed using the same approach (Ott et al.,
2000). For PTEAM, an attenuation factor of 0.55 was calculated for personal exposure.
Infiltration factors were calculated for each residence with an average of 0.56 and a standard
deviation of 0.15. Values ranged from a minimum of 0.19 to a maximum of 0.87 showing the
substantial variability that can be seen between homes depending upon the housing
characteristics and operation of the HVAC system. The mean ± SD nonambient exposure
component for PM10 was estimated as 59 ± 46 |ig/m3.
Santos-Burgoa et al. (1998) described a 1992 study of personal exposures and indoor
concentrations to a randomly sampled population near Mexico City. The sample of 66
monitored subjects included children, students, office and industrial workers, and housewives.
None of the people monitored were more than 65 years old. The mean 24-h personal exposure
and indoor concentrations were 97 ± 44 and 99 ± 50 |ig/m3, respectively, with an rPersonal/Ambient
= 0.26 (p = 0.099). Other correlations of interest were rPersonal/Indoor = 0.47 (p = 0.002) and
rindoor/Ambient= 0-23 (p = 0.158). A strong statistical association was found between personal
exposure and socioeconomic class (p = 0.047) and a composite index of indoor sources at the
home (p = 0.039).
Correlation analysis for personal exposure has not yet been reported for EXPOLIS. Some
preliminary results (Jantunen et al., 2000) show that in Basel and Helsinki a single ambient
monitoring station was sufficient to characterize the ambient PM2 5 concentration in each city.
Using microenvironmental concentration data collected while the subjects were at home, at
work, and outdoors, they calculated the sum of the time-weighted-averages of these data and
found the results closely matched the personal PM2 5 exposure data collected by the monitors
carried by most of the subjects, although a few subjects (mostly smokers) were noticeable
exceptions.
Longitudinal Studies
A number of longitudinal studies using a purposeful sampling design have been conducted
and reported in the literature since 1996. Several of these studies (Janssen et al., 1998a, 1999b,
2000; Williams et al., 2000b; Evans et al., 2000) support the previous work by Janssen et al.
5-55
-------
(1995) and Tamura et al. (1996a) and demonstrate that, for individuals with little exposure to
nonambient sources of PM, correlations between total PM exposure and ambient PM
measurements are high. Other studies (Ebelt et al., 2000; Sarnat et al., 2000) show strong
correlations for the SO42 component of PM25 but poorer correlations for PM25 mass. Still other
studies show only weak correlations (Rojas-Bracho et al., 2000; Linn et al., 1999). Even when
strong longitudinal correlations are demonstrated for individuals in a study, the variety of living
conditions may lead to variations in attenuating factors or the fraction of ambient PM
contributing to personal exposure. Groups with similar living conditions, especially if
measurements are conducted during one season, may have similar a and, therefore, very high
correlations between personal exposure and ambient concentrations, even for pooled
correlations. However, when studies contain subjects with homes of very different ventilation
characteristics or cover more than one season, variations in a can be high across subjects, thus,
showing poor pooled correlations even in the absence of indoor sources.
Elderly Subjects. Janssen et al. (2000) continued their longitudinal studies with
measurements of personal, indoor, and outdoor concentrations of PM25 for elderly subjects with
doctor-diagnosed angina pectoris or coronary heart disease. Studies were conducted in
Amsterdam, the Netherlands and Helsinki, Finland in the winter and spring of 1998 and 1999.
In the Amsterdam study with 338 to 417 observations, mean PM2 5 concentrations were 24.3,
28.6, and 20.6 |ig/m3 for personal, indoor, and outdoor samples, respectively. If the
measurements with ETS in the home were excluded, the mean indoor concentration dropped to
16 |ig/m3, which was lower than outdoor concentrations. In the Helsinki study, the mean PM2 5
concentrations were 10.8 |ig/m3 for personal, 11.0 |ig/m3 for indoor air, and 12.6 |ig/m3 outdoor
air samples. The authors noted that for this group of subjects, personal exposure, indoor
concentrations, and ambient concentrations of PM25 were highly correlated within subjects over
time. Median Pearson's correlation coefficients between personal exposure and outdoor
concentrations were 0.79 in Amsterdam and 0.76 in Helsinki. The median Pearson's r for the
indoor/outdoor relationship was 0.85 for the Amsterdam study when homes with ETS were
excluded. The correlation for indoors versus outdoors was 0.70 for all homes.
Results from the correlation analysis can be used to estimate infiltration factors and
penetration factors for these two groups of subjects. In Amsterdam, the attenuation factor was
5-56
-------
0.43 and the infiltration factor was 0.47. Very similar results were seen in Helsinki for the
attenuation factor (0.45) and the infiltration factor (0.51).
A series of three PM personal monitoring studies involving elderly subjects was conducted
in Baltimore County, MD, and Fresno, CA. The first study was a 17-day pilot (January to
February 1997) to investigate daily personal and indoor PML5 concentrations, as well as
outdoor PM2 5 and PM10_2 5 concentrations experienced by nonsmoking elderly residents of a
retirement community located near Baltimore (Liao et al., 1999; Williams et al., 2000c). The
26 residents were aged 65 to 89 (mean = 81) years, and 69% of them reported a medical
condition such as hypertension or coronary heart disease. In addition, they were quite sedentary:
on average, less than 5 h/day were spent on ambulatory activities. Because most of the residents
ate meals in a communal dining area, the average daily cooking time in the individual
apartments was only 0.5 h (range = 0 to 4.5 h). About 96% of the residents' time was spent
indoors (Williams et al., 2000c). Personal monitoring, conducted for five subjects, yielded
longitudinal correlation coefficients between ambient concentrations and personal exposure
ranging from 0.00 to 0.90.
The main Baltimore study and the Fresno study were conducted using similar monitoring
techniques and study design. Concentrations measured in these studies are summarized in
Table 5-8. For PM2 5, personal exposure and indoor air concentrations are similar for all three
studies, even though outdoor air concentrations for Fresno in the spring are only half of those
measured for Fresno in the winter and for Baltimore. This difference is presumably due to high
penetration efficiencies in the spring in Fresno when the weather was warm and participants
tended to keep the windows and doors of their homes open. These data also show that, even
when correlations are high, the use of an ambient monitor as a surrogate for exposure in
epidemiologic studies can bias the strength of the health effect found due to differing exposure
levels.
Calculated correlation coefficients are summarized in Tables 5-9 and 5-10. In Table 5-9,
the Baltimore results show high daily average correlations for both PM2 5 and PM10. These
results primarily represent the behavior of fine-particle regional sulfate for a group of
participants who have few indoor or personal sulfate sources. However, even for this group,
there was a wide range of individual correlation coefficients. The Fresno data, on the other
hand, shows much lower daily average correlations. The residential site may have been
5-57
-------
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
PM25 Concentration
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
ftig/m3)
Outdoors
22.0 ±12.0
20.5 ±13.4
10.1 ±3.2
PM10 Concentration (jig/m3)
Personal Indoors
— 13.5 ±6.3
— 15.1±4.1
37.3 16.7 ±3.1
Outdoors
30.0 ±13.7
28.2 ±15.9
28.7 ±6.6
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
PM2 5 R2
Personal/Ambient
0.80
(0.14-0.80)a
—
0.7
Personal/Indoors
0.98
(0.20-0.99)a
—
0.77
PM10R2
Ambient/Outdoor
0.89
0.48
0.61
aRange for individual participants.
TABLE 5-10. REGRESSION ANALYSIS REPORTED FOR INDOOR/OUTDOOR
RELATIONSHIPS FOR PM2 5 IN THE BALTIMORE (Williams et al., 2000a,b,c)
AND FRESNO (Evans et al., 2000) STUDIES
Daily Average
Study
Baltimore
Fresno-Winter
Fresno-Spring
R2
0.92
0.86
0.56
Slope
0.39
nr
nr
Intercept
Oig/m3)
1.5
nr
nr
R2
0.73 ±0.16
0.55 ±0.25
0.39 ±0.21
Individual
Slope
0.43 ±0.15
0.25 ±0.17
0.49 ±0.38
Intercept
Oig/m3)
0.9 ±2.6
4.4 ±3.2
3.0 ±3.7
nr = not reported.
5-58
-------
influenced by highway traffic which could account for the low correlations between ambient and
outdoor residential monitors.
In addition, the composition of particles is not the same for these two locations. In Fresno,
ammonium nitrate represents a much larger fraction of fine-particle mass. Because this
compound is in equilibrium with its precursor gases, ammonia and nitric acid, its partitioning
between the gaseous and particulate phase can be easily disturbed while outdoor particles
infiltrate indoors. Thus, differences in June 25, 2004 ratios and correlations between Fresno
and Baltimore could be a function of differences in ventilation, indoor sources, and sampling
artifacts.
The correlation analysis in Table 5-10 shows correlation coefficients as well as the slope
(infiltration factor) and the intercept (indoor concentration due to nonambient sources) for the
Baltimore and Fresno studies. These data show strongest correlations for Baltimore, where there
are very low indoor concentrations from nonambient sources. Correlations are not as strong for
Fresno, where there are higher concentrations from nonambient sources. The infiltration factors
for Baltimore and Fresno-spring are very similar at approximately 0.5. The infiltration factors
for Fresno-winter are considerably lower.
Subjects with COPT). Linn et al. (1999) describe a 4-day longitudinal assessment of
personal PM25 and PM10 exposures (on alternate days) in 30 COPD subjects aged 56 to 83 years.
Concurrent indoor and outdoor monitoring was conducted at their residences in the Los Angeles
area during summer/autumn of 1996. PM10 data from the nearest fixed-site monitoring station to
each residence was also obtained. Pooled correlations for personal exposure to outdoor
measurements gave R2 values of 0.26 and 0.22 for PM25 and PM10, respectively. Correlations of
day-to-day changes in PM2 5 and PM10 measured outside the homes and correlated with
concurrent PM10 measurements at the nearest ambient monitoring location gave R2 values of
0.22 and 0.44, respectively. Correlations of day-to-day changes in PM mass measured indoors
correlated with outdoor measurements at the homes gave R2 values of 0.27 and 0.19 for PM10
and PM2 5, respectively.
Personal, indoor, and outdoor PM2 5, PM10, and PM2 5_10 correlations were reported by
Rojas-Bracho et al. (2000) for a study conducted in Boston, MA on 18 individuals with COPD.
Both the mean and median personal exposure concentrations were higher than the indoor
5-59
-------
concentrations, which were higher than outdoor concentrations for all three PM measurement
parameters. June 25, 2004 geometric mean ratios were 1.4 ± 1.9 for PM10, 1.3 ± 1.8 for PM25,
and 1.5 ± 2.7 for PM10_25. Median longitudinal R2s between personal exposure and ambient PM
measurements were 0.12 for PM10, 0.37 for PM25 and 0.07 for PM10_25. The relationship between
the indoor and outdoor concentrations was strongest for PM25, with a median R2 of 0.55 and with
significant R2 values for 11 homes. For PM10, the median R2 value was 0.25, with significant
values for eight homes. Only five homes had significant indoor/outdoor associations for PM10_2 5,
with an insignificant median R2 value of 0.04. The poor correlations for PM10_2 5 are a result of
poorer penetration efficiencies, higher decay rates, and spatial inhomogeneities.
Bahadori et al. (1998) reported a pilot study of PM exposure of 10 nonrandomly chosen
COPD patients in Nashville, TN during the summer of 1995. Each subject alternately carried a
personal PM25 or PM10 monitor for a 12-h daytime period (8:00 am to 8:00 pm) for
6 consecutive days. These same pollutants were monitored simultaneously indoors and outdoors
at their homes. All of the homes were air-conditioned and had low air exchange rates (mean =
0.57/h), which may have contributed to the finding that mean indoor PM2 5 was 66% of the mean
ambient PM2 5. This can be contrasted with the PTEAM study in Riverside, CA, where no air
conditioners were in use and the mean indoor PM2 5 was 98% of the mean ambient PM25
(Clayton et al., 1993). Data sets were pooled for correlation analysis. Resulting pooled
correlations between personal and outdoor concentrations were r = 0.09 for PM2 5 and -0.08
for PM10.
5.3.3.1.4 Personal Exposure to Sulfate Compared to Personal Exposure to Ambient
Particulate Matter
A study conducted in Vancouver involving 16 COPD patients aged 54 to 86 years reported
low median longitudinal (r = 0.48) and pooled (r = 0.15) correlation coefficients between
personal exposures and ambient concentrations of PM25 (Ebelt et al., 2000). However, the mean
correlation between personal exposures to sulfate and ambient concentrations of sulfate was
much higher (r = 0.96). Because there are typically minimal indoor sources of sulfate, the
relationship between ambient concentrations and personal exposures to sulfate would not be
weakened by variability in an indoor-generated sulfate component as, for example, in the case
of PM2 5 for which there are many primary indoor sources as well as some secondary indoor
5-60
-------
sources. Correlations of ambient concentrations versus personal exposures for PM2 5 and sulfate
are compared in Figure 5-3.
1.00 -
— . O-75 -
§ O-50 •
*j 0.25 -
O
o
c 0.00 -
o
.2 -0.25 -
Q -0.50 -
-0.75 -
-1 nn .
Ebelt et al. (2000)
Pearson's "r"
p.. 5
I
Sulfate
Percentile
90th Percentile
75th Percentile
Median
25th Percentile
1 0th Percentile
Sarnat et al. (2000)
Spearman's "r"
PM25
1
!
T
[
^ Summer
| | Winter
TO
I
Sulfate
PM2 5 Sulfate
PM2 5 Sulfate
Figure 5-3. Comparison of correlation coefficients for longitudinal analyses of personal
exposure versus ambient concentrations for individual subjects for PM2 5
and sulfate.
Another study, conducted in Baltimore, MD, involved 15 nonsmoking adult subjects
(> 64 years old) who were monitored for 12 days during summer 1998 and winter 1999 (Sarnat
et al., 2000). All subjects (nonrandom selection) were retired, physically healthy, and lived in
nonsmoking private residences. Each residence, except one, was equipped with central
air-conditioning; however, not all residences used air-conditioning throughout the summer. The
average age of the subjects was 75 years (± 6.8 years). Sarnat et al. (2000) reported higher
longitudinal and pooled correlations for PM2 5 during summer than winter. Similar to Ebelt et al.
(2000), Sarnat et al. (2000) reported stronger associations between personal exposure to SO42~
5-61
-------
and ambient concentrations of SO42 than for total personal PM2 5 exposure and ambient PM2 5
concentrations. The ranges of correlations are shown in Figure 5-3, along with similar data from
Ebelt et al. (2000).
The higher correlation coefficients and the narrower range of the correlation coefficient for
sulfate suggest that removing indoor-generated and personal-activity PM from total personal PM
would result in a higher correlation with ambient concentrations. If, as discussed in 5.3.4.3.1,
there are no indoor sources, a personal exposure measurement for sulfate gives the ambient
exposure of sulfate; the ratio of personal sulfate to ambient sulfate gives the attenuation
coefficient on an individual, daily basis; and the attenuation coefficient times the ambient PM2 5
concentration gives the individual, daily values of ambient PM25 exposures (Wilson et al., 2000).
This technique applies only to the nonvolatile components of fine PM, as measured by PM25.
It requires that the sulfate concentration be large enough so that it can be measured with
reasonable accuracy. It does not require that sulfate be correlated with PM2 5 or the non-sulfate
components of PM25, because the sulfate data is used to estimate the attenuation coefficient,
not PM2 5. The technique does require that there be minimal indoor sources of sulfate, as
indicated by a near-zero intercept for the regression, and that the size distribution of PM25
and sulfate be similar.
Sarnat et al. (2001) subsequently extended the Baltimore study to include 20 older adults,
21 children, and 15 individuals with COPD for a total of 56 subjects. In both studies
(Sarnat et al., 2000, 2001), personal and ambient sulfate data were used to estimate the
ambient PM2 5 exposure. They used this information in mixed-model analysis (mixed models
account for differences among individual subjects), but did not report correlations between
ambient PM2 5 exposure and ambient PM2 5 concentrations based on the pooled data set.
However, Sarnat et al. (2001) did report slopes from the mixed model analyses. The t-statistic
for the slope of ambient exposure versus ambient concentration, as compared to total personal
exposure versus ambient concentration, increased from 9.96 to 11.12 (total exposure versus
ambient concentration) for the summer period and from 4.36 to 19.88 (ambient exposure versus
ambient concentration) for the winter period.
The study conducted by Sarnat et al. (2000) also illustrates the importance of ventilation on
personal exposure to PM. During the summer, subjects recorded the ventilation status of every
visited indoor location (e.g., windows open, air-conditioning use). As a surrogate for the air
5-62
-------
exchange rate, personal exposures were classified by the fraction of time the windows were open
while a subject was in an indoor environment (Fv). Sarnat et al. (2000) reported regression
analyses of personal exposure on ambient concentration for PM2 5 and for sulfate for each of the
three ventilation conditions (Figure 5-4). The correlation between personal exposure and
ambient concentration is higher for sulfate than for PM2 5, presumably because PM2 5 has indoor
sources as well as ambient sources but sulfate has only ambient sources. As expected, the
improvement is better for the lower ventilation conditions, because under these conditions the
ambient concentration is larger. For the lowest ventilation condition, R2 improves from
0.25 to 0.72.
5.3.3.1.5 Personal Exposure to Ambient and Nonambient Particulate Matter
The utility of treating personal exposure to ambient PM, A, and personal exposure to
nonambient PM, N, as separate and distinct components of total personal exposure to PM, T,
was pointed out by Wilson and Suh (1997). The PTEAM study measured, in addition to indoor,
outdoor, and personal PM10, the air exchange rate for each home and collected information on
the time spent in various indoor and outdoor microenvironments. This information is available
for 147 12-h daytime periods. With this information and statistically estimated values ofP and
k, it is possible to estimate the daytime A and N as described in Section 5.2.4.3. Various
examples of this information have been reported (Mage et al., 1999; Wilson et al., 2000).
Graphs showing the relationships between ambient concentration and the various components of
personal exposure (T, A, and N) are shown in Figure 5-5. The correlation coefficient for the
pooled data set improves from r = 0.377 for T versus C (Figure 5-5a) to r = 0.856 for A versus C
(Figure 5-5b) because of the removal of the N, which, as shown in Figure 5-5c, is highly variable
and independent of C. The correlation between A and C is less than 1.0 because of the
day-to-day variation in the a of each individual. The regression of T on C gives O~ = 0.711
and N = 81.6 |ig/m3. The regression of A on C gives (X = 0.625. The regression of N on C
gives N = 79.2 |ig/m3. The finite intercept in the regression with A must be attributed to bias or
error in some of the measurements. No reported studies, other than PTEAM, have provided the
quantity of data on individual, daily values of T, C, C;, and a that are required to conduct an
analysis comparable to that shown in Figure 5-5. It should be noted that the PTEAM study was
conducted in southern California in the fall, when houses were open and air exchange rates were
5-63
-------
3
w
o
Q.
X
LU
15
o
o
Q.
X
Lil
o
-------
_
O w
250-
200-
UJ 5? 150-
jn S 100-
- 2
ra 50-|
.o
0-
100-
7=40.5+0.7110
r= 0.373 •
R2 = 0.142
n = 147
O w
** E
w —
TO +-
o o>
75-
50-
25-
0
200
X\ = 1.16+0.6250
/•= 0.856
R2 = 0.733
O) 150-
§.s-100-|
*. o.
2 nn J
Q_ C 00 -
-50
N = 39.6+0.0860
/•= 0.051
R2 = 0.0026
n=147 ^
* * *
* *
* * * <
* * *
**
50 100 150 200
Ambient Concentration, C (ijg/m3)
250
Figure 5-5. Regression analyses of aspects of daytime personal exposure to PM10 estimated
using data from the PTEAM study: (a) total personal exposure to PM, T,
regressed on ambient concentration, C; (b) personal exposure to ambient PM,
A, regressed on C; and (c) personal exposure to nonambient PM, N, regressed
on C.
Source: Data taken from Clayton et al. (1993). Adapted from Mage et al. (1999) and Wilson et al. (2000).
5-65
-------
high and relatively uniform. These are best case conditions for showing high correlations
between ambient site measurements and personal correlations. Such high correlations are not
usually found and would not be expected with lower and more variable air exchange rates.
The RCS model introduced by Ott et al. (2000) presents a modeling framework to
determine the average contribution of ambient PM10 and indoor-generated PM10 to average
personal exposures in large urban metropolitan areas. The model has been tested using personal,
indoor, and outdoor PM10 data from three urban areas (Riverside, CA; Toronto; and Phillipsburg,
NJ). Results suggest that it is possible to separate the average ambient and nonambient PM
contributions to personal exposures on a community-wide basis. However, as discussed in the
paper, the authors make some assumptions that require individual consideration in each
city-specific application of the model for exposure or health effects investigations. Primarily,
housing factors, air-conditioning, seasonal differences, and complexities in time-activity profiles
specific to the cohort being studied have to be taken into account prior to adopting the model to
a given situation. Finally, the exposure-based analyses presented here do not predict the relative
contribution of indoor and outdoor PM to particle mass burden to the lung as a function of
human activities and different microenvironmental sources and concentrations of PM and its
co-pollutants.
5.3.3.2 Factors That Affect Relationships Between Personal Exposures and Ambient PM
A number of factors can affect the relationship between personal PM exposures and PM
measured at ambient-site community monitors. Spatial variability in outdoor
microenvironmental concentrations of PM and variations in penetration of PM into indoor
microenvironments influence the relationship of ambient PM concentration to ambient PM
exposure. Air exchange rates and decay rates in indoor microenvironments influence the
relationship of ambient PM concentrations to both ambient and total PM exposure, whereas
personal activities that generate particles influence the relationship to total, but not ambient PM
exposure. Information on these effects is presented in detail in the following sections.
5.3.3.2.1 Spatial Variability and Correlations Over Time
Chapter 3, Section 3.2.3, presents information on the spatial variability of PM mass and
chemical components at fixed-site ambient monitors; for purposes of this chapter, this spatial
5-66
-------
variability is called an "ambient gradient." The data presented in Section 3.2.3 indicate that
ambient gradients of PM and its constituents exist to a greater or lesser degree in urban areas.
These and any other gradients that may exist between a fixed-site monitor and the outdoor
microenvironments near where people live, work, and play obviously affect their exposure.
The purpose of this section is to review the available data on ambient monitor-to-outdoor
microenvironmental concentration gradients or relationships that have been measured by
researchers since 1996. These analyses, presented below, are generally consistent with earlier
studies covered in the 1996 PM AQCD. A few outdoor-to-outdoor monitoring studies also are
included to highlight relationships among important microenvironment categories. To assess
spatial variability or gradients, the spatial correlations in the data are usually analyzed.
However, it should be noted that high temporal correlation between two monitoring locations
does not imply low spatial variability or low ambient gradients. A high temporal correlation
between two sites indicates that changes in concentrations at one site may be estimated from data
collected at another site.
In a paper on the EXPOLIS-EAS study, Oglesby et al. (2000) concluded that in Basel,
Switzerland little spatial variability exists between PM levels measured at fixed site monitors
and the participants' outdoor microenvironments. The authors reported a high correlation
between home outdoor PM2 5 levels (48-h measurements beginning and ending at 8:00 a.m.) and
the corresponding 24-h average PM4 (time-weighted values calculated from midnight to
midnight) measured at a fixed monitoring station (n = 38, rsp = 0.96, p < 0.001). They
considered each home outdoor monitor as a temporary fixed monitor and concluded that
"the PM2 5 level measured at home outdoors . . . represents the fine particle level prevailing in
the city of Basel during the 48-h measuring period."
In a study conducted in Helsinki, Finland, Buzorius et al. (1999) concluded that a single
monitor may be used to adequately describe the temporal variations in concentration across the
metropolitan area. Particle size distributions were measured using a differential mobility particle
sizer (DMPS; Wintlmayer) coupled with a condensation particle counter (CPC TSI 3010, 3022)
at four locations, including the official air monitoring station, which represented a "background"
site. The monitoring period varied between 2 weeks and 6 mo for the sites, and data were
reported for 10-min and 1-, 8-, and 24-h averages. As expected, temporal variation decreased as
the averaging time increased. The authors reported that particle number concentration varied in
5-67
-------
magnitude with local traffic intensity. Linear correlation coefficients computed for all possible
site-pairs and averaging times showed that the correlation coefficient improved with increasing
averaging time. Using wind speed and direction vectors, lagged correlations were calculated and
were generally higher than the "raw" data correlations. The authors noted that weekday
correlations were higher than weekend correlations as "traffic provides relatively uniform spatial
distribution of particulate matter" and concluded that, even for time periods of 10 min and 1 h,
sampling at one station can describe temporal variations across relatively large areas of the city
with r> 0.7.
Dubowsky et al. (1999) pointed out that, although the variation of PM2 5 mass
concentration across a community may be small, there may be significant spatial variations of
specific components of the total mass on a local scale. An example is given of a study of
concentrations of poly cyclic aromatic hydrocarbons (PAHs) at three indoor locations in a
community: (1) an urban site, (2) a semi-urban site 1.6 km away, and (3) a suburban site located
further away. The authors found the geometric mean PAH concentrations at these three
locations varied respectively as 31:19:8 ng/m3 and suggested that the local variations in traffic
density were responsible for this gradient. Note that these concentrations are 1,000 times lower
than the total PM mass concentration and that such a small gradient would not be detectable for
total PM25 mass measurements on the order of 25 |ig/m3.
The Total Human Environmental Exposure Study (THEES) reported by Waldman et al.
(1991) measured indoor, outdoor, and personal benzo(a)pyrene (BaP) levels and found that the
outdoor BaP was the same at all outdoor sites across the three sampling periods. This study
showed seasonal differences for BaP levels as well as for exposures due to indoor and outdoor
sources and individual activities.
Leaderer et al. (1999) monitored 24-h PM10, PM2 5, and sulfates during the summers of
1995 and 1996 at a regional site in Vinton, VA (6 km from Roanoke, VA). One similar 24-h
measurement was made outdoors at residences in the surrounding area at distances ranging from
1 to > 175 km from the Vinton site, at an average separation distance of 96 km. The authors
reported significant correlations for PM2 5 and sulfates between the residential outdoor values
and those measured at Vinton on the same day. In addition, the mean values of the regional site
and residential site PM25 and sulfates showed no significant differences in spite of the large
distance separations and mountainous terrain intervening in most directions. However, for the
5-68
-------
concentrations of coarse particles estimated as PM10 minus PM2 5, no significant correlations
among these sites were found (n = 30; r = -0.20).
Lillquist et al. (1998) found no significant gradient in PM10 concentrations in Salt Lake
City, UT when levels were low, but a gradient existed when levels were high. Outdoor PM10
levels were measured for a period of about 5 mo at three hospitals using a Minivol 4.01 sampler
(Airmetrics, Inc.) operating at 5 L/min and at the Utah Department of Air Quality (DAQ)
ambient monitoring station located 3 to 13 km from the hospitals.
Pope et al. (1999) monitored ambient PM10 concentrations in Provo, UT (Utah Valley)
during the same time frame the following year and reported nearly identical concentrations at
three sites separated by 4 to 12 km. Pearson correlation coefficients for the data were between
0.92 and 0.96. The greater degree of variability in the Salt Lake City PM10 data relative to the
Provo data may be related to the higher incidence of windblown crustal material in Salt Lake
City.
Vakeva et al. (1999) found significant vertical gradients in submicron particles existed in
an urban street canyon of Lahti, Finland. Particle number concentrations were measured using a
TSI screen diffusion battery and a condensation particle counter at 1.5 and 25 m above the street
at rooftop level. The authors found a five-fold decrease in concentration between the two
sampling heights and attributed the vertical gradient to dilution and dispersion of pollutants
emitted at street level.
White (1998) suggested that the higher random measurement error for the coarse PM
fraction compared to the error for the fine PM fraction may be responsible for a major portion of
the apparent greater spatial variability of coarse ambient PM concentration compared to fine
ambient PM concentration in a community (e.g., Burton et al., 1996; Leaderer et al., 1999).
When PM2 5 and PM10 are collected independently and the coarse fraction is obtained by
difference (PM10_2 5 = PM10-PM2 5), the expected variance in the coarse fraction is influenced by
the variances of the PM10 and PM25 measurements. When a dichotomous sampler collects PM25
and PM10_2 5 on two separate filters, the coarse fraction is expected to have a larger error than the
fine fraction. There is a possible error caused by loss of mass below the cut-point size and a gain
of mass above the cut-point size that is created by the asymmetry of the product of the
penetration times PM concentration about the cut-point size. Because a dichotomous PM
5-69
-------
sampler collects coarse mass using an upper and lower cut point, it is expected to have a larger
variance than for fine mass collected using only one cut point.
Wilson and Suh (1997) concluded that PM25 and PM10 concentrations are correlated more
highly across Philadelphia than are PM10_2 5 concentrations. Ambient monitoring data from 1992
to 1993 was reviewed for PM2 5, PM10_2 5, and PM10, as well as for PM25 and PM10_25 dichotomous
data for 212 site-years of information contained in the AIRS database (U.S. Environmental
Protection Agency, 2000). The authors also observed that PM10 was frequently correlated more
highly with PM2 5 than with PM10_2 5. The authors noted that PM2 5 constitutes a large fraction
of PM10 and that this is the likely reason for the strong agreement between PM25 and PM10.
Similar observations were made by Keywood et al. (1999) in six Australian cities. The authors
reported that PM10 was more highly correlated with PM2 5 than with coarse PM (PM10_2 5)
suggesting that "variability in PM10 is dominated by variability in PM2 5."
Lippmann et al. (2000) examined the site-to-site temporal correlations in Detroit (1981 to
1994) and found the ranking of median site-to-site correlations was O3 (0.83), PM10 (0.78), TSP
(0.71), NO2 (0.70), CO (0.50), and SO2 (0.49). The authors explain that O3 and a fraction of TSP
and PM10 (e.g., sulfate) are secondary pollutants that would tend to be distributed more
uniformly spatially within the city than primary pollutants such as CO and SO2, which are more
likely to be influenced by local emission sources. Lippman et al. (2000) concluded ". . . spatial
uniformity of pollutants may be due to area-wide sources, or to transport (e.g., advection) of
fairly stable pollutants into the urban area from upwind sources. Relative spatial uniformity of
pollutants would therefore vary from city to city or region to region."
Goswami et al. (2002) used data collected at outdoor monitors of homes in a large
exposure study in Seattle, WA to analyze the spatial variability of outdoor PM2 5 concentrations.
The day-to-day variability between sites was 10 times higher than the spatial variability between
sites. However, differences between sites was sufficient to potentially contribute to
measurement error. An examination of the spatial characteristics of the monitoring sites showed
that the most representative monitoring sites were located at elevations of 80 to 120 m above sea
level and at distances of 100 to 300 m from the nearest arterial road.
5-70
-------
5.3.3.2.2 Physical Factors Affecting Indoor Microenvironmental Particulate Matter
Concentrations
Several physical factors can affect ambient particle concentrations in the indoor
microenvironment, including air exchange, penetration, and particle deposition. Combined,
these factors are critical variables that describe ambient particle dynamics in the indoor
microenvironments and, to a large degree, significantly affect an individual's personal exposure
to ambient particles while indoors. The relationship between the concentrations of ambient
particles outdoors, C, and ambient particles that have infiltrated indoors, Cai, is given by
C
where P is the penetration factor; a is the air exchange rate; and k is the particle deposition rate.
(As discussed in Section 5.3.2.3.1, use of this model assumes equilibrium conditions and
assumes that all variables remain constant.) Particle penetration is a dimensionless quantity that
describes the fraction of ambient particles that effectively penetrates the building shell. "Air
exchange" is a term used to describe the rate at which the indoor air in a building or residence is
replaced by outdoor air. The dominant processes governing particle penetration are air exchange
and deposition of particles as they traverse through cracks and crevices and other routes of entry
into the building. Although air exchange rates have been measured in numerous studies, very
few field data existed prior to 1996 to determine size-dependent penetration factors and particle
deposition rates. All three parameters (P, a, and k) vary substantially depending on building
type, region of the country, and season. In the past several years, researchers have made
significant advancements in understanding the relationship between particle size and penetration
factors and particle deposition rates. This section highlights studies that have been conducted to
better understand physical factors affecting indoor particle dynamics.
Air Exchange Rates
The air exchange rate, a, in a residence varies depending on a variety of factors, including
geographical location, age of the building, the extent to which window and doors are open, and
season. Murray and Burmaster (1995) used measured values of a from households throughout
the United States to describe empirical distributions and to estimate univariate parametric
5-71
-------
probability distributions of air exchange rates. Figure 5-6 shows the results classified by season
and region. In general, a is highest in the warmest region and increases from the coldest to the
warmest region during all seasons. Air exchange rates also are quite variable within and
between seasons, as well as between regions (Figure 5-6). Data from the warmest region in
summer should be viewed cautiously, because many of the measurements were made in southern
California in July when windows were more likely to be open than in other areas of the country
where air-conditioning is used. Use of air-conditioning generally results in lowering air
exchange rates. In a separate analysis of these data, Koontz and Rector (1995) suggested that a
conservative estimate for air exchange in residential settings would be 0.18 h"1 (10th percentile)
and a typical air exchange would be 0.45 h-1 (50th percentile).
3 -
re
K.
7000,
colder region = 5500 to 6999, warmer region = 2500 to 4999, and warmest
region <, 2500 heating-degree days.
Source: Adapted from Murray and Burmaster (1995).
5-72
-------
These data provide reasonable experimental evidence that a varies by season in locations
with distinct seasons. As a result, infiltration of ambient particles may be more efficient during
warmer seasons when windows are likely to be opened more frequently and air exchange rates
are higher. This suggests that the fraction of ambient particles present in the indoor
microenvironment would be greater during warmer seasons than colder seasons. For example,
in a study conducted in Boston, MA, participants living in non-air-conditioned homes kept the
windows closed except during the summer (Long et al., 2000). This resulted in higher and more
variable air exchange rates in summer than during any other season (Figure 5-7). During
nighttime periods when indoor sources are negligible, the indoor/outdoor concentration ratio or
infiltration factor may be used to determine the relative contribution of ambient particles in the
indoor microenvironment. Particle data collected during this study were used to determine the
indoor/outdoor concentration ratios by particle size (Figure 5-8). For these nine homes in
Boston the fraction of ambient particles penetrating indoors was higher during summer when air
exchange rates were higher than during fall when air exchange rates were lower (Long et al.,
200 la).
^ 8
.C
7
d>
G>
O e
X 5
Q
-------
o
•*->
o
c
o
1.1
1.0-
0.9-
0.8-
0.7-
0.6-
0.5-
0.4-
0.3-
0.2-
0.1 -
0.0
0.1-
T T
Summer Fall
CO
p
q'
CN
S 8 8 5
odd'
i i i CO
CO •<*• CD O
O O O o
o o o
if) CM
cq ^; up
9999
m c\i co ^r
T odd
OJ
CO
CN
in CD O
-- T
CD
Particle Diameter (|jm)
Figure 5-8. 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. (2001a).
Wallace et al. (2002) conducted a year-long investigation of air rates in an occupied house
to establish the effects of temperature, wind velocity, use of exhaust fans, and window-opening
behavior. Air exchange rates were calculated by periodically injecting a tracer gas (SF6) into the
return air duct and measuring the concentration in 10 indoor locations sequentially every minute
by a gas chromatograph equipped with an electron capture detector. Temperature and relative
humidity were measured outdoors and in multiple indoor locations. Wind speed and direction in
the horizontal plane were also measured. Use of the thermostat-controlled attic fan was recorded
automatically. Indoor temperatures increased from 21 °C in winter to 27 °C in summer.
Windows were open only a few percent of the time in winter but more than half the time in
summer. About 4,600 hour-long average air exchange rates were calculated from the measured
tracer gas decay rates. The mean (±SD) rate was 0.65 (0.56) h"1. Tracer gas decay rates in
different rooms were very similar, ranging only from 0.62 to 0.67 h"1, suggesting that the air in
5-74
-------
the house was well mixed throughout the year. The strongest influence on air exchange rates
was opening windows, which could increase the rate to as much as 2 h"1 for extended periods,
and up to 3 h"1 for short periods of a few hours. The use of the attic fan also increased air
exchange rates by amounts up to 1 h"1. Use of the furnace fan had no effect on air exchange rates
(ducts were all inside the house). A clear effect of indoor-outdoor temperature difference, AT,
could be discerned. However, wind speed and direction were found to have very little influence
on air exchange rates at the house where the measurements were made.
The air exchange rate, a, is a critical parameter in determining the fraction of ambient PM
found indoors and the extent of build-up of indoor PM due to indoor sources. Wallace et al.
(2002) provided a brief review of the existing literature on the theory and measurement of air
exchange rates. Open windows and frequent opening of doors lead to higher values of a.
However, many homes are kept closed for heating in winter and air-conditioning in summer.
Windows may or may not be opened during moderate weather conditions. In some areas, on the
other hand, heating or air-conditioning may not be required and ventilation by open windows
may be more frequent. Thus, a may vary geographically with climate. However, wind speed
and direction were found to have very little influence on air exchange rates at the house where
measurements were made. The variation of a with AT, as shown in Figure 5-9, is given by
a (h'1) = 0.176 + 0.0162 AT (°C). Thus, an increase of 10 °C in AT would lead to an increase
in a of 0.164 h"1, or almost a doubling of the value of a for no indoor/outdoor temperature
difference and no open windows or forced ventilation.
The observation of a minimal effect of wind speed on a is an important finding. If a were
strongly dependent on wind speed, especially at low wind speeds, there might be a correlation
between the indoor-generated PM found indoors and the ambient concentration outdoors. Such a
correlation could lead to a confounding of the relationship of ambient PM with health outcomes
by indoor-generated PM. Wallace et al. (2002) suggested that "the generally tighter construction
of homes and the use of vapor barriers may have reduced the effect of wind speed and direction
on residential air change rates compared to earlier studies."
Wind speed might be expected to have a larger effect on a in a home with open windows.
Under conditions of large a, the ambient infiltrated indoors PM concentration will be a large
fraction of the outdoor PM and the two concentrations will be highly correlated. However, the
indoor-generated PM concentration will be kept low by the high a, making a significant
5-75
-------
^ 0.85
^
o
O) 0.75H
(0
I °-65H
0)
0)
0)
0.55-
0.45-
* 0.35 H
0.25
4 8 12 16 20 24 28 32 36
Absolute Temperature Difference (°C)
Figure 5-9. Regression of air exchange rate on absolute indoor-outdoor temperature
difference. Overnight values (midnight-8 a.m.) in winter (January-March
2000). a (h'1) = 0.176 (0.011 SE) + 0.0164 (0.0005) AJC0 (N= 183, R2 = 0.82).
correlation between ambient PM concentrations and indoor-generated PM concentrations
unlikely. The observed lack of a strong wind effect on a in closed houses (Wallace et al., 2002;
Howard-Reed et al., 2002) and the observed lack of correlation between ambient PM
concentrations and indoor-generated PM concentrations indicate that the possibility of
correlation of ambient and indoor-generated PM concentrations can be discounted.
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 and composition of particles also affect deposition rates. Surface properties of
5-76
-------
particles, such as their electrostatic properties, can have a significant influence on deposition
rates. In addition, thermophoresis can also affect k, but probably to a lesser degree in the indoor
microenvironment, because temperatures generally vary over a small range. Combined, these
effects can produce order of magnitude variations in k between particles of different size and, in
the case of electrophoresis and therm ophoresis, particles of the same size.
Particle penetration efficiency into indoor microenvironments depends on particle size and
air exchange rates. Penetration varies with particle size because of the size-dependent deposition
of particles caused by impaction, interception, and diffusion of particles onto surfaces as they
traverse through cracks and crevices. Penetration also is affected by air exchange rates. When
air exchange rates are high, P approaches unity because the majority of ambient particles have
less interaction with the building shell. In contrast, when air exchange rates are low, P is
governed by particle deposition as particles travel through cracks and crevices.
Significant advancements have been made in the past few years to better characterize
particle deposition rates and penetration factors. Several new studies, including two in which
semi-continuous measurements of size distributions were measured indoors and outdoors, have
produced new information on these quantities, which are key to understanding the contributions
of ambient PM to indoor PM concentrations (Equation 5-7).
Studies involving semi-continuous measurements of indoor and outdoor particle size
distributions have been used to estimate k and P as a function of particle size (Vette et al., 2001;
Long et al., 2001a; Abt et al., 2000b). These studies each demonstrated that the indoor/outdoor
concentration ratios (Cai/C in Equation 5-9) were highest for accumulation mode particles and
lowest for ultrafme and coarse-mode particles. Various approaches were used to estimate size-
specific values for k and P. Vette et al. (2001) and Abt et al. (2000b) estimated k by measuring
the decay of particles at times when indoor levels were significantly elevated. Vette et al. (2001)
estimated P using measured values ofk and indoor/outdoor particle measurements during
nonsource nighttime periods. Long et al. (2001a) used a physical-statistical model, based on
Equation 5-12, to estimate k and P during nonsource nighttime periods. The results for k
reported by Long et al. (200la) and Abt et al. (2000b) are compared with other studies in
Figure 5-10. Although not shown in Figure 5-10, the results for k obtained by Vette et al. (2001)
were similar to the values of k reported by Abt et al. (2000b) for particle sizes up to 1 jim.
Results for P by Long et al. (200la) showed that penetration was highest for accumulation-mode
5-77
-------
~
-
-
10-
__^ :
7
.G
0)
«- 1 -
O -
a;
a
0.1-
-
-
o Byrne ef al. (1992)
o Foghefa/. (1997)a
v Ligocki ef al. (1990)b
v Ozkaynak ef al. (1 996a)c
n Sinclair ef al. (1988, 1990)"
H Thatcher and Layton (1995)d
O Wallace ef a/. (1997)d
T Abt ef al. (2000b)d
• Long ef a/. (2001a)d'e
*
i
Tii
" * f i
1 |
0.01 0.1
r
1
* i
1
' ^
' i
1
f T 1
Srm* 1
; T i {if Ifl } T
j.1 I rt ! tf I
<
1
1 T i
f j I
} T 1
y
error bar error bar
includes 0 includes 0
""" 1 ' 1
2.5 10
Particle Size (pm)
aDecay Rates represent summary estimates from the four houses examined.
b Decay rates are based on sulfate and are presented as < 2.5 u:m. Estimates were computed using a surface-to-volume
ratio of 2 m'1 (Koutrakis era/., 1992).
cData represents PM25.
dParticle sizes are the midpoint of the ranges examined.
eDecay rates presented are estimates of k for nightly average data from all nine study homes.
f Decay 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 5-10. Comparison of deposition rates from Long et al. (2001a) with literature
values (from Abt et al., 2000b). Error bars represent standard deviations
for same-study estimates.
Source: Adapted from Abt et al. (2000b); Long et al. (2001a).
particles and decreased substantially for coarse-mode particles (Figure 5-11). The results for
P reported by Vette et al. (2001) show similar trends, but were lower than those reported by
Long et al. (2001a). This is likely because of the lower air exchange rates in the one Fresno, CA
test home (a « 0.5 h"1; Vette et al., 2001) than in the nine test homes in Boston, MA study
(a > 1 h"1; Long et al., 2001a). These data for P and k illustrate the role that the building shell
may provide in increasing the concentration of particles because of indoor sources and reducing
5-78
-------
C
03
'o
c
LU
C
_g
^
2
+•>
0)
c
0)
0.
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
CO
o
CM
O
p
d
CO
o
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
re
tt
c
_o
J^
'(0
o
a.
0)
Q
Size Interval (|jm)
Figure 5-11. Penetration efficiencies and deposition rates from models of nightly average
data. Error bars represent standard errors.
Source: Boston (1998).
the concentration of indoor particles from ambient sources, especially for homes with low air
exchange rates.
Several other studies have investigated particle loss as a function of particle size. The
penetration of particles across building envelopes has been modeled for several sizes of idealized
rectangular cracks (Liu and Nazaroff, 2001). Particles of 0.1 to 1.0 jim diameter had penetration
efficiencies near 1.0. Supermicron and ultrafme particles were removed to a greater degree by
gravitational settling and Brownian diffusion. Thatcher et al. (2002) conducted an experimental
study of the effects of room furnishings and air speed on particle deposition rates indoors. The
deposition loss rate (k) increased by as much as a factor of 2.6 in going from a bare room (35 m2
surface area) to a fully furnished room (12 m2 additional surface area) with the greatest increase
seen for the smallest particles. Air speed increases from < 5 to 19 cm/s enhanced the deposition
rates by factors of 1.3 to 2.4, with greater effects on large particles than small particles. The
5-79
-------
authors stated that, "The significant effect of particle size and room conditions on deposition loss
rates argues against using a single first-order loss-rate coefficient to represent deposition for
integrated mass measurements (PM25 or PM10)." Riley et al. (2002) have modeled the
infiltration of particles into two building types: offices and residences. They developed
representative values of P, &, and ventilation-system filter efficiencies for particle sizes from
0.001 to 10 jim. They then used a typical rural and urban outdoor size distribution and
calculated predicted indoor concentrations for number, surface area, and volume distributions.
Compositional Differences Between Indoor-Generated and Ambient Particulate Matter
Wilson et al. (2000) discussed the differences in composition between particles from
indoor and outdoor sources. They noted 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) suggested the following:
"Photochemistry is significantly reduced indoors; therefore, most secondary sulfate
[H2S04, NH4HS04, and (NH4)2S04] and nitrate (NH4N03) found indoors come from
ambient sources. Primary organic emissions from incomplete combustion may be
similar, regardless of the source. However, atmospheric reactions of polyaromatic
hydrocarbons and other organic compounds produce highly oxygenated and nitrated
products, so these species are also of ambient origin. Gasoline, diesel fuel, and vehicle
lubricating oil all contain naturally present metals or metal additives. Coal and heavy
fuel oil also contain more metals and nonmetals, such as selenium and arsenic, than do
materials such as wood or kerosene burned inside homes. Environmental tobacco
smoke (ETS), however, with its many toxic components, is primarily an indoor-
generated pollutant."
Particles generated indoors may have different chemical and physical properties than those
generated by anthropogenic ambient sources. Siegmann et al. (1999) have demonstrated that
elemental carbon in soot particles generated indoors has different properties than in those
generated outdoors by automotive or diesel engines. In the United States, combustion-product
PM in the ambient/outdoor air generally is produced by burning fossil fuels (e.g., coal, gasoline,
fuel oil) and wood, whereas combustion-product PM from indoor sources is mainly produced by
biomass burning (e.g., tobacco, wood, foods). However, some indoor sources of PM (such as
cigarette smoking, meat cooking, and coal burning) occur both indoors and outdoors and may
5-80
-------
constitute an identifiable portion of measured ambient PM (Cha et al., 1996; Kleeman and Cass,
1998).
Indoor Air Chemistry
Gas- and aerosol-phase chemical reactions in the indoor microenvironment are responsible
for secondary particle formation and modification of existing particles. This process could be
complex and may influence the interpretation of exposures to indoor-generated particles in
instances when particles are generated by outdoor gases reacting with gases generated or
released indoors to produce fresh particles. For example, homogeneous gas phase reactions
involving ozone and terpenes (specifically J-limonene, a-terpinene, and a-pinene) have been
identified as an important source of submicron particles (Weschler and Shields, 1999). Terpenes
are present in several commonly available household cleaning products, and J-limonene has
been identified in more than 50% of the buildings monitored in the Building Assessment and
Survey Evaluation (BASE) study (Hadwen et al., 1997). Long et al. (2000) found that when
PineSol (primary ingredient is a-pinene) was used indoors, indoor PM2 5 mass concentrations
increased from 11 to 32 |ig/m3 (indoor O3 concentrations unknown, but ambient O3 levels were
44 to 48 ppb). Similarly, a 10-fold increase in number counts of 0.1 to 0.2 |im particles was
observed in an experimental office containing supplemental J-limonene and normally
encountered indoor O3 concentrations (< 5 to 45 ppb), resulting in an average increase in particle
mass concentration of 2.5 to 5.5 |ig/m3 (Weschler and Shields, 1999). Ozone appears to be the
limiting reagent, because particle number concentration varied proportionally to O3
concentrations (Weschler and Shields, 1999). Other studies showed similar results (e.g., Jang
and Kamens, 1999; Wainman et al., 2000). Such particles, if toxic, would represent an increased
health risk due to ambient air pollution. However, the concentration would depend on the
ambient O3 concentration, the O3 infiltration factor, and the indoor generation rate of terpenes.
It seems unlikely that the concentration of the resulting particles would be correlated with
ambient PM on an individual or population basis. Therefore, it seems more appropriate to
consider indoor-reaction particles as part of nonambient exposure. However, since ambient O3
and ambient PM2 5 are correlated in the summertime, indoor-reaction particles due to O3-terpene
reactions might be correlated with outdoor PM2 5 mass.
5-81
-------
Indoor Sources of Particles
The major sources of indoor PM in nonsmoking residences and buildings include
suspension of PM from bulk material, cooking, cleaning, and the use of combustion devices such
as stoves and kerosene heaters. Human and pet activities also lead to PM detritus production
(from tracked-in soil, fabrics, skin and hair, home furnishings, etc.) that is found ubiquitously in
house dust deposited on floors and other interior surfaces. House dust and lint particles may be
resuspended indoors by agitation (cleaning) and turbulence (HVAC systems, human activities,
etc.). Ambient particles that have infiltrated into the indoor microenvironments also may be
resuspended after deposition to indoor surfaces. Typically, resuspension of particles from any
source involves coarse particles (> 1 jim); particles with smaller diameters are not resuspended
efficiently. On the other hand, cooking produces both fine- and coarse-mode particles, whereas
combustion sources typically produce ultrafine particles.
Environmental tobacco smoke (ETS) is also a major indoor source of PM. It is, however,
beyond the scope of this chapter to review the extensive literature on ETS. A number of articles
provide source-strength information for cigarette or cigar smoking (e.g., Daisey et al. [1998] and
Nelson etal. [1998]).
A study conducted on two homes in the Boston metropolitan area (Abt et al., 2000a)
showed that indoor PM sources predominate when air exchange rates were < 1 h"1, and outdoor
sources predominate when air exchange rates were > 2 h"1. The authors attributed this to the fact
that when air exchange rates were low (< 1 h"1), particles released from indoor sources tend to
accumulate, because particle deposition is the mechanism governing particle decay and not air
exchange. Particle deposition rates are generally < 1 h"1, especially for accumulation-mode
particles. When air exchange rates were higher (> 2 h"1), infiltration of ambient aerosols and
exfiltration of indoor-generated aerosols occur more rapidly, reducing the effect of indoor
sources on indoor particle levels. The study also confirmed previous findings that the major
indoor sources of PM are cooking, cleaning, and human activity. The authors discussed the size
characteristics of these ubiquitous sources as follows:
"The size of the particles generated by these activities reflected their formation
processes. Combustion processes (oven cooking, toasting, and barbecuing)
produced fine particles and mechanical processes (sauteing, frying, cleaning,
and movement of people) generated coarse particles. These activities increased
particle concentrations by many orders of magnitude higher than outdoor levels
and altered indoor size distributions." (Abt et al., 2000a; p. 43)
5-82
-------
They also noted that variability in indoor PM for all size fractions was greater (2 to 33 times
higher) than for outdoor PM, especially for short averaging times.
In a separate study conducted in nine nonsmoking homes in the Boston area, Long et al.
(2000) concluded that the predominant source of indoor fine particles was infiltration of outdoor
particles and that cooking activities were the only other significant source of fine particles.
Coarse particles, however, had several indoor sources such as cooking, cleaning, and various
indoor activities. This study also concluded that more than 50% of the particles (by volume)
generated during indoor events were ultrafme particles. Events that elevated indoor particle
levels were found to be brief, intermittent, and highly variable, thus requiring the use of
continuous instrumentation for their characterization. Because the concentration of ultrafme
particles is greater near the source (they grow in size into the accumulation mode as they age),
the personal cloud for ultrafme particles may be higher than for accumulation mode particles if
the person is nearer the source than the indoor monitor. Table 5-11 provides information on the
mean volume mean diameter (VMD) for various types of indoor particle sources. The
differences in mean VMD confirm the clear separation of source types and suggest that there
is very little resuspension of accumulation-mode PM. In addition, measurements of OC and EC
indicated that OC had significant indoor sources whereas EC was primarily of ambient origin.
Vette et al. (2001) found that resuspension was a significant indoor source of particles
> 1 |im. Concentrations of fine particles were not affected by resuspension. Figure 5-12 shows
the diurnal variability in the indoor/outdoor aerosol concentration ratio from an unoccupied
residence in Fresno, CA. The study was conducted in the absence of common indoor particle
sources such as cooking and cleaning. The data in Figure 5-12 show the mean indoor/outdoor
concentration ratio for particles > 1 |im increased dramatically during daytime hours. This
pattern was consistent with indoor human activity levels. In contrast, the mean indoor/outdoor
concentration ratio for particles < 1 |im (ultrafme and accumulation-mode particles) remain
fairly constant during both day and night.
Wallace and Howard-Reed (2002) used three instruments (SMPS, APS, and Climet) to
measure ultrafme, fine, and coarse particles in an inhabited residence for 18 months. They
confirmed the observations of Abt et al. (2000a) and Long et al. (2000) that indoor sources
primarily generate ultrafme and coarse particles. Wallace and Howard-Reed reported that
"Indoor sources affecting ultrafme particle concentrations were observed 22% of the time, and
5-83
-------
TABLE 5-11. VOLUME MEAN DIAMETER (VMD) AND MAXIMUM PM2 s
CONCENTRATIONS OF INDOOR PARTICLE SOURCESab
Size Statistics
PM,
Indoor Activity Backgrounda>e
Mean VMD Mean VMD
Particle Source
(urn)
(urn)
Maximum Concentration c'd
Mean SD
Oig/m3)
Oig/m3)
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
8
24
23
4
13
3
20
2
11
10
5
15
52
26
7
0.189f
0.107f
0.138f
0.114f
0.184f, 3.48g
0.135f
0.173f
0.159f
5.38B
3.86g
0.097f
3.96g
4.25B
4.28g
0.311f
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
14.8
101.2
54.9
29.3
65.6
37.2
40.5
14.8
22.6
6.5
11
12
8
4.8
28
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
18
Notes:
a All concentration data corrected for background particle levels.
b Includes only individual particle events that were unique for a given time period and could be detected above
background particle levels.
°PM concentrations in ug/m3.
dMaximum concentrations computed from 5-min data for each activity.
e Background data are for time periods immediately prior to the indoor event.
f Size statistics calculated for PV002.0 5 using SMPS data.
g Size statistics calculated for PV0 7_10 using APS data.
Source: Long et al. (2000).
5-84
-------
0.0
Figure 5-12. Mean hourly indoor/outdoor particle concentration ratio from an
unoccupied residence in Fresno, CA during spring 1999.
Source: Vette etal. (2001).
sources affecting fine and coarse particle concentrations were observed 12 and 15% of the time,
respectively . . . Indoor sources, such as cooking with natural gas, and simple physical activities,
such as walking, accounted for a majority (50-90%) of the ultrafine and coarse particle
concentrations, whereas outdoor sources were more important for accumulation-mode particles
between 0.1 and 1 |im in diameter."
5.3.3.2.3 Time-Activity Patterns
Total exposure to PM is the sum of various microenvironmental exposures that an
individual encounters during the day and will depend on the microenvironments occupied.
As discussed previously, PM exposure in each microenvironment is the sum of exposures from
ambient PM (in outdoor and indoor microenvironments), indoor-generated PM, and indoor-
reaction PM. In addition, there is exposure to PM generated by personal activities that is
determined by the specific personal activities that the individual conducts while in those
5-85
-------
microenvironments. As mentioned before, PM exposures and their components are variable
across the population and, thus, each are distributions rather than point estimates. A thorough
analysis of these distributions would require a comprehensive sensitivity and uncertainty
analysis.
Determining microenvironments and activities that contribute significantly to human
exposure begins with establishing human activity pattern information for the general population
as well as for subpopulations. Personal exposure and time-activity pattern studies have shown
that different populations have varying time-activity patterns and, accordingly, different personal
PM exposures. Both characteristics will vary greatly as a function of age, health status, ethnic
group, socioeconomic status, season, and region of the country. Collecting detailed time-activity
data can be very burdensome on participants but is clearly valuable in assessing human exposure
and microenvironments. For modeling purposes, human activity data frequently come from the
general databases that are discussed below.
The gathering of human activity information, often called "time-budget" data, started in the
1920s; however, their use for exposure assessment purposes only began to be emphasized in the
1980s. As noted earlier, many of the largest U.S. human activity databases have been
consolidated by the EPA's National Exposure Research Laboratory (NERL) into one
comprehensive database (CHAD) containing more than 22,000 person-days of 24-h activity
(Glen et al., 1997; McCurdy et al., 2000). The information in CHAD is accessible for
constructing population cohorts of people with diverse characteristics that are useful for analysis
and modeling (McCurdy, 2000). See Table 5-2 for a summary listing of human activity studies
in CHAD. Most of the databases in CHAD are also available elsewhere, including the National
Human Activity Pattern Survey (NHAPS), California's Air Resources Board (CARB), and the
University of Michigan's Institute for Survey Research data sets.
Although CHAD provides a very valuable resource for time and location data, there is little
information on PM-generating personal activities. In addition, very few of the time-activity
studies have collected longitudinal data within a season or over multiple seasons. Such
longitudinal data are important in understanding potential variability in activities and how they
affect correlations between PM exposure and ambient site measurements for both total PM and
PM of ambient origin.
5-86
-------
5.3.3.3 Effect of Ambient Sources on Exposures to Particulate Matter
Different sources may generate ambient PM with different aerodynamic and chemical
characteristics, which may, in turn, result in different health responses. Thus, to fully understand
the relationship between PM exposure and health outcomes, exposure from different sources
should be identified and quantified. Source apportionment techniques provide a method for
determining personal exposure to PM from specific sources. Daily contributions from sources
that have no indoor component can be used as tracers to generate exposure estimates for ambient
PM of similar aerodynamic size or used directly as exposure surrogates in epidemiological
analyses. The most recent EPA PM Research Needs Document (U. S. Environmental Protection
Agency, 1998) recommended using source apportionment techniques to determine daily time
series of source categories for use in community, time-series epidemiology.
A number of epidemiologic studies (discussed more fully in Chapter 8) have evaluated
relationships between health outcomes and sources of PM determined from measurements at a
community monitor. These studies suggest the importance of examining sources and
constituents of indoor, outdoor, and personal PM. For example, Ozkaynak and Thurston (1987)
evaluated the relationship between PM sources and mortality in 36 Standard Metropolitan
Statistical Areas (SMSAs). Particulate matter samples from the EPA's Inhalable Particle (IP)
Network were analyzed for SO42 and NO3 by automated colorimetry, and elemental
composition was determined with X-ray fluorescence (XRF). Mass concentrations from five PM
source categories were determined from multiple regression of absolute factor scores on the
mass concentration: (1) resuspended soil, (2) auto exhaust, (3) oil combustion, (4) metals, and
(5) coal combustion.
In another study, Mar et al. (2000) applied factor analysis to evaluate the relationship
between PM composition (and gaseous pollutants) in Phoenix, AZ. In addition to daily averages
of PM2 5 elements from XRF analysis, they included in their analyses OC and EC in PM2 5 and
gaseous species emitted by combustion sources (CO, NO2, and SO2). They identified five source
factors classified as (1) motor vehicles, (2) resuspended soil, (3) vegetative burning, (4) local
SO2, and (5) regional sulfate. Additionally, Laden et al. (2000) applied specific rotation factor
analysis to XRF PM composition data from six eastern cities (Ferris et al., 1979). Fine PM was
regressed on the recentered scores to determine the daily source contributions. Three main
5-87
-------
sources were identified: (1) resuspended soil (Si), (2) motor vehicle (Pb), and (3) coal
combustion (Se).
Source apportionment or receptor modeling has been applied to the personal exposure data
to understand the relationship between personal and ambient sources of PM. Application of
source apportionment to ambient, indoor, and personal PM composition data is especially useful
in sorting out the effects of particle size and composition. If a sufficient number of samples are
analyzed with sufficient compositional detail, it is possible to use statistical techniques to derive
source category signatures, to identify indoor and outdoor source categories, and to estimate
their contribution to indoor and personal PM.
Positive Matrix Factorization (PMF) has been applied to the PTEAM database by
Yakovleva et al. (1999). The authors utilized mass and XRF elemental composition data from
indoor and outdoor PM2 5 and personal, indoor, and outdoor PM10 samples. PMF is an advance
over ordinary factor analysis, because it allows measurements below the quantifiable limit to be
used by weighting them according to their uncertainty. This effectively increases the number of
species that can be used in the model. The factors used by the authors correspond to general
source categories of PM, such as outdoor soil, resuspended indoor soil, indoor soil, personal
activities, sea salt, motor vehicles, nonferrous metal smelters, and secondary sulfates. PMF, by
identifying the various source factors and apportioning them among the different monitor
locations (personal, indoor, and outdoor), was able to estimate the contribution of resuspended
indoor dust to the personal cloud (15% from indoor soil and 30% from resuspended indoor soil).
Factor scores for these items then were used in a regression analysis to estimate personal
exposures (Yakovleva et al., 1999).
The most important contributors to PM10 personal exposure were indoor soil, resuspended
indoor soil, and personal activities; these accounted for approximately 60% of the mass
(Yakovleva et al., 1999). Collectively, they include personal-cloud PM, smoking, cooking, and
vacuuming. For both PM2 5 and PM10, secondary sulfate and nonferrous metal operations
accounted for another 25% of PM mass. Motor vehicle exhausts, especially from vehicles
started inside attached garages, accounted for another 10% of PM mass. The PTEAM study was
conducted in Riverside, CA in the fall of 1990. Yakovleva et al. (1999) cautioned that their
results may not apply to other geographic areas, seasons of the year, or weather conditions.
-------
Simultaneous measurement of personal (PM10) and outdoor measurements (PM2 5
and PM10) were evaluated as a three-way problem with PMF, which allowed for differentiation
of source categories based on their variation in time and type of sample, as well as their variation
in composition. By using this technique, three sources of coarse-mode, soil-type PM were
identified. One was associated with ambient soil, one with indoor soil dispersed throughout the
house, and one with soil resulting from the personal activity of the subject.
Two other source apportionment models have been applied to ambient measurement data
and can be used for the personal exposure studies. The effective variance weighted Chemical
Mass Balance (CMB) receptor model (Watson et al., 1984, 1990, 1991) solves a set of linear
equations that incorporate the uncertainty in the sample and source composition. CMB requires
the composition of each potential source of PM and the uncertainty for the sources and ambient
measurements. Source apportionment with CMB can be conducted on individual samples;
however, composition of each of the sources of PM must be known. An additional source
apportionment model, UNMIX (Henry et al., 1994) is a multivariate source apportionment
model. UNMIX is similar to PMF, but does not explicitly use the measurement uncertainties.
Because measurement uncertainties are not used, only species above the detection limit are
evaluated in the model. UNMIX provides the number of sources and source contributions and
requires a similar number of observations as PMF.
The Yakovleva et al. (1999) study demonstrated that source apportionment techniques also
could be very useful in determining parameters needed for exposure models and for determining
exposure to ambient PM. Exposure information, similar to that obtained in the PTEAM study,
but including other PM components useful for definition of other source categories (e.g., OC,
EC; organic tracers for EC from diesel vehicle exhaust, gasoline vehicle exhaust, and wood
combustion; nitrate; Na, Mg, and other metal tracers; and gas-phase pollutants) would be useful
as demonstrated in the use of EC/OC and gas-phase pollutants by Mar et al. (2000).
5.3.3.4 Correlations of Particulate Matter with Other Pollutants
Correlations between ambient concentrations and between ambient concentrations and
personal exposures for PM and other pollutants are of importance in understanding possible
confounding in epidemiological studies and are discussed more fully in Chapter 9. Available
information from exposure studies is presented in this section. Several epidemiologic studies
5-89
-------
have included the gaseous pollutants CO, NO2, SO2, and O3 along with PM10 or PM2 5 in the
analysis of the statistical association of health responses with pollutants. In a recent study,
the personal exposure to O3 and NO2 was determined, as well as that to PM2 5 and PM10_2 5 for
a cohort of 15 elderly subjects in Baltimore, MD, although measured personal exposures
to O3, NO2, and SO2 were below their respective level of detection (LOD) for 70% of the
samples. Spearman correlations for 14 subjects in summer and 14 subjects in winter are given
in Table 5-12 for relationships between personal PM25 and ambient concentrations of PM25,
PM10_2 5, O3, and NO2. In contrast to ambient concentrations, neither personal exposure to
total PM2 5 nor to ambient PM2 5 was correlated significantly with personal exposures to the
co-pollutants, nonambient PM2 5 and ambient PM10_2 5, O3, NO2, and SO2. Personal-ambient
associations for PM10_2 5, O3, NO2, and SO2 were similarly weak and insignificant. Based on
these results, Sarnat et al. (2000) concluded that the potential for confounding of PM2 5
by O3, NO2, or PM10_2 5 appears to be limited, because, despite significant correlations observed
among ambient pollutant concentrations, the correlations among personal exposures were low.
Sarnat et al. (2001) further evaluated the role of gaseous pollutants in PM epidemiology by
extending the measurements taken on the earlier adult cohort of 20 individuals in Baltimore, MD
through inclusion of additional PM and gaseous pollutant measurements that were collected
during the same 1998 to 1999 period from 15 individuals with COPD and from 21 children.
Twenty-four-hour average personal exposures for PM2 5, O3, SO2 and NO2, and corresponding
ambient concentrations for PM25, O3, SO2, NO2and CO for all 56 subjects were collected over
12 consecutive days. Results from correlation and regression analysis of the personal and
ambient data showed that personal PM2 5 and personal gaseous pollutant exposures were
generally not correlated. The analysis also showed that ambient PM2 5 concentrations had
significant associations with personal PM2 5 exposures in both seasons. On the other hand,
ambient gaseous pollutant concentrations were not correlated with their corresponding personal
exposure concentrations. However, ambient gaseous concentrations were found to be strongly
associated with personal PM2 5 exposures, suggesting that ambient gaseous concentrations
for O3, NO2, and SO2 are acting as surrogates, as opposed to confounders, of PM25 in the
estimation of PM health effects based on multipollutant models. This study did not measure
personal CO and did not find a significant association between summertime ambient CO and
personal PM2 5 (a significant wintertime association, however, was found).
5-90
-------
TABLE 5-12. CORRELATIONS BETWEEN PERSONAL PM25 AND AMBIENT
POLLUTANT CONCENTRATIONS1
Personal PM2 s vs. Ambient
Subject
SUMMER SA1
SA2
SA5
SB1
SB2
SB3
SB4
SB5
SB6
SCI
SC2
SC3
SC4
SC5
WINTER WAI
WA2
WA4
WA5
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.S10
-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 PM2 s of Ambient
Origin vs. Ambient
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
PM10_2.5
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 r values; italicized values indicate significance at the p < 0.05 level.
Source: Sarnat et al. (2000).
5-91
-------
Personal EC and SO42 were also measured during the winter for the cohort of COPD
patients only. The analysis of this subset of the data showed that personal SO42 was
significantly and negatively associated with ambient O3 and SO2, and that personal EC was
significantly associated with ambient O3, NO2, and CO. The authors interpret these findings as
suggesting that O3 is primarily a surrogate for secondary particle exposures, whereas ambient
CO and NO2 are primarily surrogates for particles from traffic. Sarnat et al. (2001) cautioned
that these findings were from only one location and various physical and personal factors, such
as ventilation, time spent outdoors, and household characteristics, could affect the strength of the
reported associations for certain individuals and cohorts even though the qualitative results
found were unlikely to change.
A newly developed Roll-Around System (RAS) was used to evaluate the hourly
relationship between gaseous pollutants (CO, O3, NO2, SO2, and VOCs) and PM (Chang et al.,
2000). Exposures were characterized over a 15-day period for the summer and winter in
Baltimore, MD based on scripted activities to simulate activities performed by older adults
(65+ years of age). Spearman rank correlations were reported for PM2 5, O3, CO, and toluene for
both the summer and winter. The correlations are given for each microenvironment in
Table 5-13: indoor residence, indoor other, outdoor near roadway, outdoor away from road,
and in vehicle. No significant relationships (p < 0.05) were found between hourly PM2 5 and O3.
Significant relationships were found between hourly PM2 5 and CO in: indoor residence, winter;
indoor other, summer and winter; and outdoor away from roadway, summer. Significant
relationships also were found between hourly PM2 5 and toluene in: indoor residence, winter;
indoor other, winter; and in vehicle, winter. The significant relationships between CO and PM2 5
in the winter may be caused by reduced air exchange rates that could allow them to accumulate
(Chang et al., 2000). Note that, although no significant correlation was found between in-vehicle
PM25 and CO, another significant component of vehicle exhaust, toluene (Conner et al., 1995),
was significantly correlated to PM2 5 in the winter.
Carrer et al. (1998) presented data on the correlations among microenvironmental and
personal PM10 exposures and concentrations and selected environmental chemicals monitored
simultaneously (using methods not described). These chemicals were nitrogen oxides (NOX),
carbon monoxide (CO), and total volatile organic compounds (TVOC), benzene, toluene, xylene,
5-92
-------
TABLE 5-13. CORRELATIONS BETWEEN HOURLY PERSONAL PM25 AND
GASEOUS POLLUTANTS
Indoor Outdoor Near Outdoor Away
Residence Indoor Other Roadway from road
PM2Svs. O3
Summer
Winter
PM25vs.
CO
Summer
Winter
N
35
56
41
59
rs N rs N rs N
0.29 16 -0.14 10 0.05 12
0.05 37 -0.06 11 -0.28 7
0.25 19 0.59a 13 0.14 12
0.43a 39 0.62a 13 0.37 8
rs
0.45
0.04
0.62
0.41
In Vehicle
N rs
37 0.21
34 -0.10
46 0.23
37 0.1
PM25vs. Toluene
Summer
Winter
46
66
0.23 21 -0.14 14 0.26 14
0.38a 47 0.44a 17 0.4 8
0.02
0.48
48 0.12
42 0.43a
"Correlations represent Spearman's rvalues; italicized values indicate significance at the p < 0.05 level.
Source: Chang et al. (2000).
and formaldehyde. The Kendall T correlation coefficient was used; only results significant at
p < 0.05 are mentioned here. Significant associations were found only between the following
pairs of substances (T shown in parentheses): personal PM10 (24 h) and NOX (0.34), CO (0.34),
TVOC (0.18), toluene (0.19), and xylene (0.26); office PM10 and NOX (0.31); home PM10
and NOX (0.24), CO (0.24), toulene (0.17), and xylene (0.25). Even though most of these
chemicals are associated with motor vehicular emissions, there was no significant correlation
between commuting PM10 and any of the substances.
5-93
-------
5.4 SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA
5.4.1 Introduction
Atmospheric PM contains a number of chemical constituents that may be of significance to
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 PAHs in ambient, personal, and microenvironmental PM samples.
However, there have not been extensive analyses of the constituents of PM in personal or
microenvironmental samples. Data from relevant studies are summarized in this section. The
summary does not address bacteria, bioaerosols, viruses, or fungi (e.g., Owen et al., 1992; Ren
etal., 1999).
5.4.2 Monitoring Studies That Address Particulate Matter Constituents
Relevant studies published in recent years that have measured the constituents of PM in
personal or microenvironmental samples are summarized in Tables 5-11 and 5-12 for personal
exposure measurements of PM and microenvironmental samples, respectively. Studies that
measured both personal and microenvironmental samples are included in Table 5-11.
The largest database on personal, microenvironmental, and outdoor measurements of PM
elemental concentrations is the PTEAM study (Ozkaynak et al., 1996b). The results are
highlighted in the table and discussed below. The table shows that a number of studies have
measured concentrations of elements (by XRF), organic carbon (OC), various indicators of
elemental carbon (EC), aerosol acidity, sulfate, ammonia, and nitrate. Additionally, a number of
studies have measured PAHs, both indoors and outdoors. Other than the PAHs, there are few
data on the organic constituents of PM.
5.4.3 Key Findings
5.4.3.1 Correlations of Personal and Indoor Concentrations with Ambient Concentrations
of Particulate Matter Constituents
The elemental composition of PM10 in personal samples was measured in the PTEAM
study, the first probability-based study of personal exposure to particles. A number of important
observations made from the PTEAM data collected in Riverside, CA are summarized by
5-94
-------
Ozkaynak et al. (1996b). Population-weighted daytime personal exposures averaged
150 ± 9 |ig/m3 compared to concurrent indoor and outdoor concentrations of 95 ± 6 |ig/m3.
The personal exposure measurements suggested that there was a "personal cloud" of particles
associated with personal activities. Daytime personal exposures to 14 of the 15 elements
measured in the samples were considerably greater than concurrent indoor or outdoor
concentrations; sulfur was the only exception.
The PTEAM data also showed good agreement between the concentrations of the elements
measured outdoors in the backyard of the residences with the concentrations measured at the
central site in the community. The agreement was excellent for sulfur. Although the particle
and element mass concentrations were higher in personal samples than for indoor or outdoor
samples, a nonlinear mass-balance method showed that the penetration factor was nearly 1 for all
particles and elements.
Measurements of element concentrations in NHEXAS are similar to the PTEAM results in
that they show elevated concentrations of arsenic (As) and lead (Pb) in personal samples relative
to indoor and outdoor samples (Clayton et al., 1999b). The elevated concentrations of As and Pb
were consistent with elevated levels of PM50 in personal samples (median particle exposure of
101 |ig/m3) as compared to indoor concentrations (34.4 |ig/m3). There was a strong association
between personal and indoor concentrations and between indoor and outdoor concentrations for
both As and Pb. However, there were no central site ambient measurements for comparison to
the outdoor or indoor measurements at the residences.
Manganese (Mn) concentrations were measured in PM2 5 samples collected in Toronto
(Crump, 2000). The mean PM25 Mn concentrations were higher outdoors than indoors.
However, the outdoor concentrations measured at the participants' homes were lower than those
measured at two fixed locations. Crump (2000) suggested that the difference in the
concentrations may have been because the fixed locations were likely closer to high-traffic areas
than were the participants' homes.
Studies of acidic aerosols and gases typically measure strong acidity (H+), SO42 , NH4+,
and NO"3. The relationship between the concentrations of these ions and the relationship
between indoor and outdoor concentrations have been addressed in a number of studies during
which personal, microenvironmental, and outdoor samples have been collected, as shown in
Tables 5-14 and 5-15. Key findings from these studies include the following:
5-95
-------
vo
TABLE 5-14. STUDIES THAT HAVE MEASURED PARTICULATE MATTER CONSTITUENTS IN
PERSONAL EXPOSURE SAMPLES
PM Constituent
Elements
As and Pb
Mn
Acid Aerosol
Constituents
Study Name/Reference
PTEAM/Ozkaynak et al.
(1996b)
NHEXAS/Clayton et al.
(1999b)
Pellizzari et al. (1998,
1999); Clayton etal.
(1999a); Crump (2000)
Samat et al. (2000)
Study Location
Riverside, CA
EPA Region 5
Toronto
Baltimore, MD
Population Size/No, of Samples
178 adults
167 samples
925 personal samples
20 adults
Summary of Results
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 vs. outdoor at the residences were 0.98 for S and
0.5 to 0.9 for other elements (except Cu).
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.
Braueretal. (1989)
Suhetal. (1992)
Suhetal. (1993a,b)
Suh etal. (1994)
Boston, MA
Uniontown, PA
24 children for 2 days
State College, PA 47 children
Personal exposures to aerosol strong acidity slightly lower than
concentrations measured at stationary site.
Personal exposures to H+ and SO42~ lower than outdoor levels, but
higher than indoor microenvironmental levels; personal NH4+ and
NO3~ higher than indoor or outdoor levels.
Results similar to Uniontown, PA study.
Results indicate strong neutralization of acidity indoors.
PAHs
Individual
Particle
Analyses by
CCSEM
Trace Elemental
analyses by
XRF
Elemental
analysis by
HR-ICP-MS
Waldman and Liang
(1993); Waldman etal.
(1990)
Zmirou et al. (2000)
Conner etal. (2001)
Landis etal. (2001)
Kinney et al. (2002)
Georgia and
New Jersey
Grenoble, France
Baltimore, MD
Same study
locations for both
New York City,
NY
Hospital, day cares
38 adults
3 sets of indoor-outdoor-personal
filters, > 2000 particles/filter
1 9 day s with P, I, A for
10 elderly retirees
46 student volunteers, 1 week in
summer and winter, PM2 5
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.
CCSEM was used to identify individual particles and classify them
by most likely sources.
P and A highly correlated for PM2 5 and sulfates. P and A not
significantly correlated for soil and trace element oxides.
29 PM constituents were measured.
Most were similar for P, I, and O, suggesting outdoor sources
dominant.
-------
TABLE 5-15. STUDIES THAT HAVE MEASURED PARTICULATE MATTER CONSTITUENTS IN
MICROENVIRONMENTAL SAMPLES
Study
PM Constituent Name/Reference
Study Location
Population Size/No, of Samples Summary of Results
vo
Acid Aerosol
Constituents
PAHs
PAHs and
phthalates
Jones et al. (2000)
Birmingham,
England
Patterson and Eatough Lindon, UT
(2000)
Leaderer et al. (1999) Virginia and
Connecticut
Brauer et al. (1990) Boston, MA
Chuang et al. (1999) Durham, NC
Dubowsky et al.
(1999)
Sheldon et al.
(1993a,b)
Boston, MA
Placerville and
Roseville, CA
PTEAM/Ozkaynak Riverside, CA
etal. (1996b),
Sheldon etal. (1993c)
12 residences
1 school
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 PM2 5
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 SO42~ exceeded indoor
levels in winter and summer. I/O ratios of H+ lower than I/O
ratios of SO42~ 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.
Abbreviations used:
I/O = Indoor/Outdoor.
A = ambient.
I = indoor.
O = outdoor.
P = personal.
-------
• Acid aerosol concentrations measured at the residences in the Uniontown, PA study were
significantly different from those measured at a fixed ambient site located 16 km from the
community. However, Leaderer et al. (1999) reported that the regional ambient air
monitoring site in Vinton, VA provided a reasonable estimate of indoor and outdoor
sulfate measurements during the summer at homes without tobacco combustion.
• Approximately 75% of the fine aerosol indoors during the summer was associated with
outdoor sources based on I/O sulfate ratios measured in the Leaderer et al. (1999) study.
• Personal exposures to strong acidity (H+) were lower than corresponding outdoor levels
measured in studies by Brauer et al. (1989, 1990) and Suh et al. (1992). However, the
personal exposure levels measured by Suh et al. (1992) were higher than the indoor
microenvironmental levels.
• Personal exposures to NH4+ and NO3 were reported by Suh et al. (1992) to be higher than
either indoor or outdoor levels.
Personal exposures to SO42 were also lower than corresponding outdoor levels, but higher
than the indoor microenvironmental levels (Suh et al., 1992; 1993a,b) as shown in Table 5-16.
The fact that the personal and indoor H+ concentrations were substantially lower than
outdoor concentrations suggests that a large fraction of aerosol strong acidity is neutralized by
ammonia. Ammonia is emitted in relatively high concentrations in exhaled breath and sweat.
The difference between indoor and outdoor H+ concentrations in the Suh et al. (1992, 1993a,b)
studies was also much higher than the difference for indoor and outdoor SO42 , indicative of
neutralization of the H+. Results of the Suh et al. (1992, 1993a,b) studies also showed
substantial interpersonal variability of H+ concentrations that could not be explained by variation
in outdoor concentrations.
Similar results for ammonia were reported by Waldman and Liang (1993). They reported
that levels of ammonia in monitored institutional settings were 10- to 50-times higher than
outdoors and that acid aerosols were largely neutralized. Leaderer et al. (1999) reported that
ammonia concentrations during both winter and summer in residences were an order of
magnitude higher indoors than outdoors, consistent with results of other studies and the presence
of sources of ammonia indoors.
Sulfate aerosols appear to penetrate indoors effectively. Waldman et al. (1990) reported
I/O ratios of 0.7 to 0.9 in two nursing care facilities and a day-care center. Sulfate I/O ratios
were measured for three particle size fractions in 12 residences in Birmingham, England by
Jones et al. (2000). The sulfate I/O ratios were 0.7 to 0.9 for PM< 1.1 |im, 0.6 to 0.8 for PM
5-98
-------
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
so42-
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
Number of
Samples
(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 ± GSD"
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) Personal (12 h)
GM±GSD" GM±GSD"
1.4 ±2.1 —
1.4 ±2.1 —
109.4 ±2.4 —
109.4 ±2.4 —
91.0 ±2.5 71.5 ±2.4
104.4 ±2.3 —
82.5 ±2.6
82.5 ±2.6
72.4 ±2.9 18.4 ±3.0
124.9 ±1.9 110.3 ±1.8
139.4 ±2.1 167.0 ±2.0
76.6 ±2.7 42.8 ±2.2
a!n/Out = Indoor samples/outdoor samples.
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).
1.1 to 2.1 |im, and 0.7 to 0.8 for PM 2.1 to 10 |im. Suhetal. (1993b) reported that personal and
outdoor sulfate concentrations were highly correlated, as depicted in Figure 5-13.
Patterson and Eatough (2000) reported indoor/outdoor relationships for a number of PM25
components and related species in Lindon, UT during January and February of 1997. Outdoor
samples were collected at the Utah State Air Quality monitoring site. Indoor samples were
collected in the adjacent Lindon Elementary School. The infiltration factors, Cai/C, given by the
5-99
-------
600
100 200 300 400 500
Outdoor Sulfate (nmoles/m3)
600
Figure 5-13. Personal versus outdoor SO42 in State College, PA. Open circles represent
children living in air-conditioned homes; the solid line is the 1:1 line.
Source: Suhetal. (1993b).
slope of the regression lines (Table 5-17), were low (0.27 for sulfate and 0.12 for PM25) possibly
because of removal of particles in the air heating and ventilation system. The authors concluded
that indoor PM2 5 mass may not always be a good indicator of exposure to ambient combustion
material due to the influence of indoor particle sources. Presumably this occurs because the
concentrations of indoor-generated particles are not well correlated with the concentrations of
ambient combustion particles. However, ambient sulfate, SO2, nitrate, soot, and total particle
number displayed strong correlations with indoor exposure, presumably because these species
have few indoor sources in the absence of indoor combustion. Ambient PM2 5 mass was not a
good indicator of indoor PM25 mass exposure, presumably due to uncorrelated indoor sources
of PM25 mass.
5-100
-------
TABLE 5-17. REGRESSION ANALYSIS OF INDOOR VERSUS OUTDOOR
CONCENTRATIONS (units are nmol/m3, except for soot and metals, which are ug/m3
and absorption units/m3, 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 = FJNP
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 = C;g
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
0.7
0.71
0.7
0.54
0.88
0.44
0.43
0.69
0
0.01
Average
Outdoors, C0
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).
Oglesby et al. (2000) conducted a study to evaluate the validity of fixed-site fine particle
concentration measurements as exposure surrogates for air pollution epidemiology. Using 48-h
EXPOLIS data from Basel, Switzerland, they investigated the personal exposure/outdoor
concentration relationships for four indicator groups: (1) PM25 mass, (2) sulfur and potassium
for regional air pollution, (3) lead and bromine for traffic-related particles, and (4) calcium for
crustal particles. The authors reported that personal exposures to PM2 5 mass were not correlated
to corresponding home outdoor levels (n = 44, r = 0.07). In the study group reporting neither
relevant indoor sources nor relevant activities, personal exposures and home outdoor levels of
sulfur were highly correlated (n = 40, r = 0.85). These results are consistent with spatially
homogeneous regional pollution and higher spatial variability of traffic and crustal materials.
5-101
-------
Indoor, outdoor, and personal concentrations of PM25 mass and a variety of PM
constituents were measured at an 18-story retirement facility in Towson, MD for 28 24-hour
monitoring periods during July and August 1998 (Landis et al., 2001). Indoor and outdoor
measurements were made with a Versatile Air Pollutant Sampler (VAPS). Indoor, outdoor, and
personal samples were made with a Personal Exposure Monitor (PEM). The VAPS
(a dichotomous sampler) collected PM25 (15 L/min) and PM10_25 (2 L/min) while the PEM
collected PM25 (2 L/min). A comparison of the VAPS and the PEM indicated that the indoor
PEM collected much higher mass and more soil components than the indoor VAPS although the
differences between outdoor results were smaller and not significant. These differences were
attributed to the presence of a larger coarse particle concentration inside (and perhaps larger
diameter particles) and either more particle bounce or a higher 50% cut point for the PEM.
In their analysis, Landis et al. (2001) compared indoor and outdoor VAPS data, as well as
outdoor and personal PEM data.
As shown in Table 5-18 (PEM) and Table 5-19 (VAPS), higher correlations were found for
fine-particle components of PM25 and lower correlations for coarse-particle components. Like
Patterson and Eatough (2000), Landis et al. (2001) found low infiltration factors for nitrate along
with a reasonable correlation suggesting that fine-mode ammonium nitrate may be evaporating
after it penetrates indoors. Neither sulfate nor nitrate had indoor sources.
Indoor and outdoor PM2 5 and PM10 mass and chemical composition were measured in
13 homes (2 to 4 days for each home) in the Coachella Valley, a unique desert area in southern
California during the winter and spring of 2000 (Geller et al., 2002). Maximum infiltration of
ambient PM would be expected during this period because the mild climate minimizes the use of
heating or air conditioning. Regression analysis was used to estimate FINF and Cig. Results are
shown in Table 5-20. The Coachella PM is generally considered to be rich in coarse PM and
epidemiological studies have associated PM10 (Ostro et al., 1999) and estimated PM10_25 (Ostro
et al., 2000) with mortality. However, the results of Geller et al. (2002) indicate that even during
periods of high air exchange rates, indoor exposures would be dominated by PM25. Geller et al.
(2002) also reported results for some chemical components of PM25 and PM10_25 as well as EC
T^f"1 9 S
and OC in PM2 5. For EC, FINF = 0.74, in good agreement with FINF = 0.74. Indoor concentrations
of OC were much higher than outdoor concentrations; the average I/O ratio was 1.77.
5-102
-------
TABLE 5-18. MIXED MODEL ANALYSIS OF PERSONAL VERSUS
OUTDOOR CONCENTRATIONS3
PM25 (ug/m3)
Sulfateb (ug/m3)
Soil0 (ng/m3)
TEOd (ng/m3)
w
0.82
0.95
0.03
0.39
a
0.46 ± 0.04
0.40 ± 0.02
0.05 ±0.12
0.43 ±0.11
C
21
10.2
458
165
A
9.66
4.08
93
71
N
3.57 ±0.80
0.1 ±0.04
544 ± 95
99 ±23
T
12.8
4.5
553
170
A+N
13.2
4.2
637
170
a a (slope) and N (intercept) taken from mixed model results. A calculated as a • C; T is measured value.
Note good agreement between measured T and T estimated from A + N except for soil.
b as (NH4)2SO4.
0 Soil = sum of oxides Si, Ca, Fe, and Ti (due to coarse PM in PM25).
d TEO = trace element oxides (mostly fine PM components).
Source: Landis et al. (2001).
TABLE 5-19. REGRESSION ANALYSIS OF INDOOR VERSUS
OUTDOOR CONCENTRATIONS3
PM2 5 (ug/m3)
Sulfateb (ug/m3)
OC« 1.4 (ug/m3)
ECC (ug/m3)
NaCl (ng/m3)
Soild (ng/m3)
TEOe (ng/m3)
Nitrate (ng/m3)
R2
0.74
0.98
0.44
0.58
0.32
0.3
0.38
0.83
FINF
0.35
0.41
0.3
0.32
0.06
0.07
0.35
0.09
C
18.9
10.4
5.4
0.5
231
363
94
372
cal
6.6
4.4
1.6
0.16
14
25
33
33
Clg
0.32
-0.24
8
0.09
37
51
7
12
c,
6.7
4
9.7
0.4
48
74
39
68
cal+clg
6.9
4.2
9.6
0.25
53
76
40
45
a F^p (slope) and Clg (intercept) taken from regression equation. Cm calculated as C • FINF; Q is measured value.
Note reasonable agreement between measured Q and Q estimated from C^ + Clg except for EC and nitrate.
b as (NH4)2S04.
0 revised regression with outliers omitted.
d Soil = sum of oxides Si, Ca, Fe, and Ti (due to coarse PM in PM25).
e TEO = trace element oxides (mostly fine PM components).
Source: Landis et al. (2001).
5-103
-------
TABLE 5-20. REGRESSION ANALYSIS OF INDOOR VERSUS
OUTDOOR CONCENTRATIONS3
Fine
Coarse
R2
0.37
0.35
FINF
0.74
0.3
C
15
8.6
cal
11.1
2.6
c,g
4.3
3
c,
15.4
5.6
a F^p (slope) and Clg (intercept) taken from regression equation. C^ calculatedjis C • FINF. Q is measured value.
Note excellent agreement between measured Q and Q estimated from C^ + Clg.
Source: Geller et al. (2002).
PAHs have been measured in studies by EPA and the California Air Resources Board.
PAH results from a probability sample of 125 homes in Riverside are discussed in reports by
Sheldon et al. (1992a,b) and Ozkaynak et al. (1996b). Data for two sequential 12-h samples
were reported for PAHs by ring size (3 to 7) and for individual phthalates. The results can be
summarized as follows.
• The particulate-phase 5- to 7-ring species had lower relative concentrations than the more
volatile 3- to 4-ring species.
• The 12-h I/O ratios for the 5- to 7-ring species ranged from 1.1 to 1.4 during the day and
from 0.64 to 0.85 during the night (Sheldon et al., 1993a).
An indoor air model used to calculate indoor "source strengths" for the PAHs showed that
smoking had the strongest effect on indoor concentrations.
• An
Results from a larger PAH probability study in 280 homes in Placerville and Roseville, CA
(Sheldon et al., 1993a,b) were similar to the 125-home study. The higher-ring, particle-bound
PAHs had lower indoor and outdoor concentrations than the lower-ring species. For most PAHs,
the I/O ratio was greater than 1 for smoking and smoking/fireplace homes and less than 1 for
fireplace-only, wood stove, wood stove/gas heat, gas heat, and "no source" homes.
A study of PAHs in indoor and outdoor air was conducted in 14 inner-city and 10 rural
low-income homes near Durham, NC in two seasons (winter and summer) in 1995 (Chuang
et al., 1999). Fine-particle-bound PAH concentrations measured with a real-time monitor were
usually higher indoors than outdoors (2.47 ± 1.90 versus 0.53 ± 0.58 |ig/m3). Higher indoor
5-104
-------
levels were seen in smokers' homes compared with nonsmokers' homes, and higher outdoor and
indoor PAH levels were seen in urban areas compared with rural areas.
In a study reported by Dubowsky et al. (1999), the weekday indoor PAH concentrations
attributable to traffic (indoor source contributions were removed) were 39 ± 25 ng/m3 in a
dormitory that had a high air exchange rate because of open windows and doors; 26 ± 25 ng/m3
in an apartment; and 9 ± 6 ng/m3 in a suburban home. The study showed that both indoor and
outdoor sources (especially motor vehicular traffic) contributed to indoor PAH concentrations.
BaP concentrations were measured in the THEES study (Waldman et al., 1991). A compre-
hensive analysis of the data showed considerable seasonal variability of indoor and outdoor
sources and resultant changes in personal exposures to BaP.
The indoor and outdoor concentrations of 30 PAHs were measured in 55 nonsmoking
residences in Los Angeles, CA; Houston, TX; and Elizabeth, NJ (Naumova et al., 2002).
A comparison of I/O ratios of low molecular weight PAHs (3 to 4 rings) and higher molecular
weight PAHs (5 to 7 rings) indicated that indoor sources had a significant effect on indoor
concentrations of 3-ring PAHs and a smaller effect on 4-ring PAHs, while outdoor sources
dominated the indoor concentrations of 5- to 7-ring PAHs.
5.4.4 Factors Affecting Correlations Between Ambient Measurements
and Personal or Microenvironmental Measurements of
Particulate Matter Constituents
The primary factors affecting correlations between personal exposure and ambient air PM
measurements are discussed in Section 4.3.2. These include air exchange rates, particle
penetration factors, decay rates, removal mechanisms, indoor air chemistry, indoor sources, and
freshly-generated particles indoors. The importance of these factors varies for different PM
constituents. For acid aerosols, indoor air chemistry is particularly important as indicated by the
discussion of the neutralization of the acidity by ammonia which is present at higher
concentrations indoors because of the presence of indoor sources. For SVOCs, including PAHs
and phthalates, the presence of indoor sources will substantially affect the correlation between
indoor and ambient concentrations (Ozkaynak et al., 1996b; Sheldon et al., 1993b). Penetration
factors for PM will affect correlations between indoors and outdoors for most elements, except
Pb, which may have significant indoor sources in older homes. Indoor air chemistry, decay
5-105
-------
rates, and removal mechanisms may affect soot and OC. Furthermore, reactions between indoor
or outdoor gases and particles may produce freshly generated aerosols indoors. These factors
must be fully evaluated when attempting to correlate concentrations of PM constituents in
ambient, personal, or indoor samplers.
5.4.5 Limitations of Available Data
The previous discussion demonstrates that there is limited data available that can be used
to compare personal, microenvironmental, and ambient air concentrations of PM constituents.
Because of resource limitations, PM constituents have not been measured in many studies of PM
exposure. There are little data on freshly generated indoor aerosols. Although there are some
data on acid aerosols, the comparisons between the personal and indoor data generally have been
with outdoor measurements at the participants' residences, not with community ambient air
measurement sites. The relationship between personal exposure and indoor levels of acid
aerosols is not clear because of the limited database, although there is some evidence that acidity
will be neutralized by indoor ammonia (Suh et al., 1994). The exception is total sulfate, for
which there appears to be a strong correlation between indoor and ambient concentrations.
With the exception of PAHs, there are practically no data available with which to relate
personal or indoor concentrations to outdoor or ambient site concentrations of SVOCs that may
be generated from a variety of combustion and industrial sources. The relationship between
exposure and ambient concentrations of particles from specific sources, such as diesel engines,
has not been determined.
Although there is an increasing amount of research being performed to measure PM
constituents in different PM size fractions, the current data are inadequate to adequately assess
the relationship between indoor and ambient concentrations of most PM constituents. Additional
information is also needed on PM exposures that result from outdoor gases reacting with indoor
gases. This is a source that could also vary with outdoor PM, e.g., when the outdoor gas is O3.
5-106
-------
5.5 IMPLICATIONS OF USING AMBIENT PM CONCENTRATIONS
IN TOXICOLOGICAL AND EPIDEMIOLOGICAL STUDIES OF
PM HEALTH EFFECTS
5.5.1 Toxicology
Most studies of PM toxicity have used either pure chemicals or ambient PM. Indoor-
generated PM differs somewhat from ambient PM in terms of sources, size, and composition.
It is possible, therefore, that indoor-generated PM could have different toxicological properties.
Environmental tobacco smoke (ETS) has been studied extensively; however, little toxicological
information exists on PM from other indoor sources. Long et al. (2001b) have assessed the in
vitro toxicity of 14 paired indoor and outdoor PM2 5 samples collected in nine Boston-area
homes. Bioasseys using rat alveolar macrophages (AMs) and measuring tumor necrosis factor
(TNF) were used to assess particle-induced proinflammatory responses. TNF production was
found to be significantly higher in AMs exposed to indoor PM than to those exposed to outdoor
PM. This result held even after normalization for endotoxin concentrations, which were higher
in indoor samples. The authors concluded that their results "suggest that indoor-generated
particles may be more bioactive than ambient particles." PM, in its various forms, produces
many types of biological effects. It seems possible that indoor PM could be more active than
ambient PM for some effects and less for others.
5.5.2 Potential Sources of Error Resulting from Using Ambient Particulate
Matter Concentrations in Epidemiological Analyses
In this section, the exposure issues that relate to the interpretation of the findings from
epidemiologic studies of PM health effects are examined. This section examines the errors that
may be associated with using ambient PM concentrations in epidemiologic analyses of PM
health effects. First, implications of associations found between personal exposure and ambient
PM concentrations are reviewed. This is discussed separately in the context of either community
time-series studies or long-term, cross-sectional studies of chronic effects. Next, the role of
compositional and spatial differences in PM concentrations are discussed and how these may
influence the interpretation of findings from PM epidemiology. Finally, using statistical
methods, an evaluation of the influence of exposure measurement errors on PM epidemiologic
studies is presented.
5-107
-------
Measurement studies of personal exposures to PM are still few in number and limited in
spatial, temporal, and demographic coverage. Consequently, with the exception of a few
longitudinal panel studies, most epidemiologic studies of PM health effects rely on ambient
community monitoring data giving 24-h average PM concentration measurements. Moreover,
because of limited sampling for PM2 5, many of these epidemiologic studies used available PM10
or in some instances relied on historic data for other PM measures or indicators, such as
TSP, SO42 , TP15 (inhalable particles with an upper cut of 15 jim), RSP (respirable suspended
particles), COH (coefficient of haze), etc. A critical question often raised in the interpretation of
results from acute or chronic epidemiological community-based studies of PM is whether the use
of ambient stationary site PM concentration data influenced or biased the findings from those
studies.
If it is assumed that total personal PM exposure is responsible for observed effects, use of
ambient concentrations could lead to misclassification of individual exposures and to errors in
the epidemiologic analysis of pollution and health data depending on the pollutant and on the
mobility and lifestyles of the population studied. Ambient monitoring stations can be some
distance away from the individuals and may represent only a fraction of all likely outdoor
microenvironments that individuals come in contact with during the course of their daily lives.
Furthermore, most individuals are quite mobile and move through multiple microenvironments
(e.g., home, school, office, commuting, shopping) and engage in diverse personal activities at
home (e.g., cooking, gardening, cleaning, smoking). Some of these microenvironments and
activities may have different sources of PM and result in distinctly different concentrations of
PM than that monitored by the fixed-site ambient monitors. Consequently, exposures of some
individuals could be classified incorrectly if only ambient monitoring data are used to estimate
individual-level total personal PM exposures. Thus, improper assessment of exposures using
data routinely collected by the neighborhood monitoring stations could conceivably lead to a
bias or increase in the standard error in epidemiologic analyses. Except for extremely unlikely
situations, however, the bias would be expected to reduce the estimated health risk coefficient.
Many individuals are typically exposed to particles in a multitude of indoor and outdoor
microenvironments during the course of a day. Thus, concern about possible error introduced in
the estimation of PM risk coefficients using ambient, as opposed to personal, PM measurements
has received considerable attention recently from exposure analysts, epidemiologists, and
5-108
-------
biostatisticians. Some exposure analysts contend that, for community time-series epidemiology
to yield information on the statistical association of a pollutant with a health response, there must
be an association between personal exposure to a pollutant and the ambient concentration of that
pollutant because people tend to spend around 90% time indoors and are exposed to both indoor-
generated and ambient-infiltrated PM (cf Wallace, 2000b; Brown and Paxton, 1998; Ebelt et al.,
2000). Consequently, numerous findings reported in the epidemiologic literature on significant
associations between ambient PM concentrations and various morbidity and mortality health
indices, in spite of the low correlations between ambient PM and concentrations and measures of
personal exposure, have been described by some exposure analysts as an exposure paradox
(Lachenmyer and Hidy, 2000; Wilson et al., 2000).
To resolve that exposure paradox, two types of analyses must considered. The first type of
analysis is to examine the correlations between ambient PM concentrations and personal
exposures that are relevant to most of the existing PM epidemiologic studies using either pooled,
daily-average, or longitudinal exposure data. The second approach is to study the degree of
correlations between the two key components of personal PM exposures (i.e., exposures caused
by ambient PM and exposures caused by nonambient PM) with ambient or outdoor PM
concentrations, for each of the three types of exposure study designs. Yet, even with these two
approaches, it may still be difficult to examine complex synergisms which, in some situations,
may preclude simple decoupling of indoor and outdoor particles either in terms of exposure or
total dose delivered to the lung. In addition, several factors that influence either the exposure or
health response characterization of the subjects have to be addressed. Such factors include:
• spatial variability of PM components,
• health or sensitivity status of subjects,
• variations of PM with other co-pollutants,
• co-generation of fine and ultrafine particles from outdoor air and indoor gaseous pollutants,
• formal evaluation of exposure errors in the analysis of health data, and
• how the results may depend on the variations in epidemiological study design.
To facilitate the discussion of these topics, a brief review of concepts pertinent to exposure
analysis issues in epidemiology is presented.
5-109
-------
5.5.2.1 Associations Between Personal Exposures and Ambient Particulate
Matter Concentrations
As defined in Sections 5.3 and 5.4, personal exposures to PM result from an individual's
exposures to PM in many different types of microenvironments (e.g., outdoors near home,
outdoors away from home, indoors at home, indoors at office or school, commuting, restaurants,
malls, other public places). Total personal exposures that occur in these indoor and outdoor
microenvironments can be classified as those resulting from ambient PM (ambient PM exposure
includes exposure to ambient PM while outdoors and exposure while indoors to ambient PM that
has infiltrated indoors) and those primarily generated by indoor sources, indoor reaction, and
personal activities (nonambient PM exposure). The associations between personal exposures
and ambient PM concentrations that have been reported from various personal exposure
monitoring studies under three broad categories of study design, (1) longitudinal, (2) daily-
average, or (3) pooled exposure studies, and are summarized below.
In Sections 5.4.3.1.2 and 5.4.3.1.3, recent studies conducted mainly in the United States
and involving children, the elderly, and subjects with COPD were reviewed, and they indicated
that both intra- and interindividual variability in the relationships between personal exposures
and ambient PM concentrations were observed. A variety of different physical, chemical, and
personal or behavioral factors were identified by the original investigators that seem to influence
the magnitude and the strength of the associations reported.
For cohort studies in which individual daily health responses are obtained, individual
longitudinal PM personal exposure data (including ambient and nonambient components) may
provide the appropriate indicators. In this case, health responses of each individual can be
associated with the total personal exposure, the ambient exposure, or the nonambient exposure of
each individual. Additionally, the relationships of personal exposure indicators with ambient
concentration can be investigated. In the case of community time-series epidemiology, however,
it is not feasible to obtain experimental measurements of personal exposure for the millions of
people over the periods of years that are needed to investigate the relationship between air
pollution and infrequent health responses such as deaths or even hospital admissions. The
epidemiologist must work with the aggregate number of health responses occurring each day and
a measure of the ambient concentration that is presumed to be representative of the entire
community. The relationship of PM exposures of the potentially susceptible groups to
monitored ambient PM concentrations depends on their activity pattern and level, residential
5-110
-------
building and HVAC factors (which influence the infiltration factor), status of exposure to ETS,
amount of cooking or cleaning indoors, and seasonal factors, among others. For these special
subgroups, average personal exposures to ambient PM correlate well with ambient PM
concentrations regardless of individual variation in the absence of major microenvironmental
sources.
Even though both ambient and nonambient PM exposure contribute to daily baseline PM
dose received by the lung, there seem to be clear differences in the relationships of ambient and
nonambient PM exposure with ambient PM concentration. Various researchers have shown that
nonambient PM exposure is independent of ambient PM concentration, but that ambient PM
exposure is a function of ambient PM concentration. Wilson et al. (2000) explained the
difference based on different temporal patterns that affect PM concentrations, "Concentrations of
ambient PM are driven by meteorology and by changes in the emission rates and locations of
emission sources, while concentrations of nonambient PM are driven by the daily activities of
people." Still, although nonambient PM exposure may not correlate with ambient PM
concentration or ambient PM exposure, it will nevertheless add to the daily baseline dose
received by the lung. An important concern, for which there is little information, is the relative
biological activity of ambient and indoor-generated PM both in terms of the type of toxic effect
and the relative potency for a given effect.
Ott et al. (2000) also discussed the reasons for assuming that nonambient PM exposure is
independent of ambient PM exposure and ambient PM concentration. They showed that the
nonambient component of total personal exposure is uncorrelated with the outdoor concentration
data. Ott et al. (2000) demonstrated that the ambient PM exposure was similar for three
population-based exposure studies: two large probability-based studies (the PTEAM study
conducted in Riverside [Clayton et al., 1993; Thomas et al., 1993; Ozkaynak et al., 1996a,b] and
a study in Toronto [Pelizzarri et al., 1999; Clayton et al., 1999a]), and a nonprobability-based
study conducted in Phillipsburg (Lioy et al., 1990). Based on these three studies, they concluded
that the average nonambient PM exposure and the distribution of individual, daily values of
nonambient PM exposure can be treated as constant from city to city.
Dominici et al. (2000) examined a larger database consisting of five different PM exposure
studies and also concluded that nonambient PM exposure can be treated as relatively constant
from city to city, although their data show greater variability than the data reported by Ott
5-111
-------
(2000). However, the constancy of nonambient PM exposure remains an open question. If daily
values of nonambient PM exposure were constant, this would imply a zero correlation with
ambient PM concentration. However, this hypothesis of constant individual, daily nonambient
PM exposure has not been fully established, since only a few studies have produced the data
needed to estimate daily, individual values of nonambient PM exposure. Although nonambient
PM exposure is independent of ambient PM concentration, it may not be independent of the
attenuation factor (ambient PM exposure/ambient PM concentration). Sarnat et al. (2000)
showed that nonambient PM exposure increases as the ventilation rate (and attenuation factor)
decreases. By comparing winter and summer regression equations Lachenmeyer and Hidy
(2000) also showed that as the slope, which gives the attenuation factor, decreases, the intercept,
which gives the nonambient PM exposure, increases. If ambient and nonambient PM2 5 are not
correlated, as seems likely, statistical analyses using ambient concentrations will not be able to
demonstrate health effects, or the lack of health effects, due to indoor-generated PM. However,
indoor-generated PM could raise the overall exposure level, contribute to the observed health
effects, and make it more difficult to observe a threshold level for ambient PM.
Mage et al. (1999) assumed that the PM10 concentration component from indoor sources
(e.g., smoking, cooking, cleaning, burning candles) is not correlated with the outdoor
concentration. They indicated that this lack of correlation is to be expected because people are
unaware of ambient concentrations and do not necessarily change their smoking or cooking
activities as outdoor PM10 concentrations vary, an assumption supported by other empirical
analyses of personal exposure data. For the PTEAM data set, Mage et al. (1999) have shown
that individual, daily exposures to indoor-generated PM and daily ambient PM concentrations
have a correlation coefficient near zero (R2 = 0.005). Wilson et al. (2000) have shown that
individual, daily values of concentration of ambient PM which has infiltrated indoors and
indoor-generated PM concentrations also have a near zero correlation (R2 = 0.03). Figure 5-14
shows the relationship of estimated values of nonambient PM exposure with ambient PM
concentrations (calculated by EPA, daily individual values from PTEAM and daily average
values for the cohort from THEES).
Based on these results, it is reasonable to assume that nonambient PM exposure will
ordinarily have little correlation with ambient PM concentration. A possible exception could be
caused by indoor-reaction PM, PM formed when an ambient gas infiltrates indoors and reacts
5-112
-------
(0
o
Q_ c
+- ^
C
0)
C Q_
O ^
200
150-
100-
N = 39.6+0.0860*
r= 0.051
R2 = 0.0026
n=147
3 3
•O (/)
II
C LU
(0
O
-50
0
50
100
150 200
250
(0
c
O
(2 —
/i\ ^
0) ^
E §
0 Q.
> X
< UJ
(0
D
60
c 3.
-------
with an indoor-generated gas, e.g., the reaction of O3 with terpenes from air cleaners. However,
ambient O3 does not appear to be highly correlated with ambient or personal PM2 5 (see
Table 5-12, Sarnat et al., [2000]). Additionally, not every home will use air fresheners or have
the same level of terpene emissions. Hence, indoor-reaction PM concentrations would not be
expected to correlate with ambient PM concentrations on a community basis. Therefore, in
linear nonthreshold models of PM health effects, nonambient PM exposure is not expected to
contribute to the relative risk determined in a regression of health responses on ambient PM
concentration. Furthermore, in time-series analysis of pooled or daily health data, it is expected
that ambient PM exposure, rather than total personal PM exposure, will have the stronger
association with ambient PM concentration.
5.5.2.2 Role of Compositional Differences in Exposure Characterization for Epidemiology
The majority of the available data on PM exposures and relationships with ambient PM
have come from a few large-scale studies, such as PTEAM, or longitudinal studies on selected
populations. Consequently, for most analyses, exposure scientists and statisticians had to rely
on PM10 or PM2 5 mass data, instead of elemental or chemical compositional information on
individual or microenvironmental samples. In a few cases, researchers have examined the
factors influencing I/O ratios or penetration and deposition coefficients using elemental mass
data on personal, indoor, and outdoor PM data (e.g., Ozkaynak et al., 1996a,b; Yakovleva et al.,
1999). These results have been informative in terms of understanding relative infiltration of
different classes of particle sizes and sources into residences (e.g., fossil fuel combustion, mobile
source emissions, soil-derived). Clearly, accumulation-mode particles, associated with
stationary or mobile combustion sources, have greater potential for penetration into homes and
other microenvironments than crustal material. There will be variability in the chemical
composition of these broad categories of source classes as well as probable variations in relative
toxicity. Moreover, when particles and reactive gases are present indoors in the presence of
other pollutants or household chemicals, they may react to form additional or different
compounds and particles with yet unknown physical, chemical, and toxic compositions
(Wainman et al., 2000). Thus, if indoor-generated and ambient PM were responsible for
different types of health effects or had significantly different toxicities on a per unit mass basis,
it would then be important that ambient and nonambient exposure should be separated and
5-114
-------
treated as different components, much like the current separation of PM10 into PM25 and PM10_25.
These complexities in personal exposure profiles may introduce nonlinearities and other
statistical challenges in the selection and fitting of concentration-response models.
Unfortunately, PM health effects models have not yet been able to meaningfully consider such
complexities.
It is important to also note that individuals spend time in places other than their homes and
outdoors. Many of the interpretations reported in the published literature on factors influencing
personal PM10 exposures, as well as in this chapter, come from the PTEAM study. The PTEAM
study was conducted 10 years ago in one geographic location in California, during one season,
and most residences had very high and relatively uniform air exchange rates. Non-home indoor
microenvironments were not monitored directly during the PTEAM study. Commuting
exposures from traffic or exposures in a variety of different public places or office buildings
could not be assessed directly. Nonresidential buildings may have lower or higher ambient
infiltration rates depending on the use and type of the mechanical ventilation systems employed.
Because the source and chemical composition of PM affecting personal exposures in different
microenvironments vary by season, day-of-the-week, and time of day, it is conceivable that
some degree of misclassification of exposures to toxic PM agents of concern could be introduced
when health-effects models use only daily-average mass measures such as PM10or PM25.
However, because of the paucity of currently available data on many of these factors, it is not
now possible to ascertain the potential magnitude or severity of any such complex exposure
missclassification problems or their potential implications for interpretation of results from PM
epidemiology.
5.5.2.3 Role of Spatial Variability in Exposure Characterization for Epidemiology
Chapter 3, Section 3.2.3 and Chapter 5, Section 5.3 present information on the spatial
variability of PM mass and chemical components at fixed-site ambient monitors; for purposes of
this chapter, this spatial variability is called an "ambient gradient." Any gradient that may exist
between a fixed-site monitor and the outdoor microenvironments near where people live, work,
and play will obviously affect the concentration profile actually experienced by people as they
go about their daily lives.
5-115
-------
However, the evidence so far indicates that PM concentrations, especially fine PM (mass
and sulfate), generally are uniformly distributed in most metropolitan areas. This reduces the
potential for exposure misclassification because of outdoor spatial gradients when a limited
number of ambient PM monitors are used to represent population average ambient exposures in
time-series or cross-sectional epidemiological studies of PM. This topic is further discussed in
Section 5.6.5. However, as discussed earlier, the same assumption is not necessarily true for
different components of PM such as PM10_25, because source-specific and other spatially
nonuniform pollutant emissions could alter the spatial profile of individual PM components in a
community. For example, particulate and gaseous pollutants emitted from motor vehicles tend
to be higher near roadways and inside cars. Likewise, acidic and organic PM species may be
location- and time-dependent. Furthermore, human activities are complex. If outdoor PM
constituent concentration profiles are either spatially or temporally variable, it is likely that
exposure misclassification errors could be introduced in the analysis of PM air pollution and
health data.
5.5.3 Analysis of Exposure Measurement Error Issues in Particulate Matter
Epidemiology
The effects of exposure misclassification on the relative risk estimates of disease using the
classical 2x2 contingency design (i.e., exposed/nonexposed versus diseased/nondiseased) have
been studied extensively in the epidemiologic literature. It has been shown that the magnitude of
the exposure-disease association (e.g., relative risk), because of either misclassification of
exposure or disease alone (i.e., nondifferential misclassification), biases the effect results toward
the null; and differential misclassification (i.e., different magnitudes of disease misclassification
in exposed and nonexposed populations) can bias the effect measure toward or away from the
null value relative to the true measure of association (Shy et al., 1978; Gladen and Rogan, 1979;
Copeland et al., 1977; Ozkaynak et al., 1986). However, the extension of these results from
contingency analysis design to multivariate models (e.g., log-linear regression, Poissson, logit)
typically used in recent PM epidemiology has been more complicated.
5-116
-------
5.5.3.1 Time-Series Analyses
Researchers have investigated the appropriateness of using ambient PM concentration as
an exposure metric and have developed a framework for analyzing measurement errors typically
encountered in the analysis of time-series mortality and morbidity effects from exposures to
ambient PM (cf. Zeger et al., 2000; Dominici et al., 2000; Samet et al., 2000). Use of this
framework, discussed more extensively in Chapters 8 and 9, leads to the following conclusions:
the deviation of an individual's personal exposure from the risk-weighted average exposure due
to variations in ambient concentrations, infiltration rates, and indoor-generated PM
concentrations is a Berkson error and will not bias the estimated regression coefficient (P, the
increase in risk per unit increase in PM) in a time-series analysis of mortality as a function of
ambient PM concentrations. However, the difference between the average personal exposure
and the true ambient concentration will bias p. If the daily, individual values of ambient PM
exposure and nonambient PM exposure are poorly correlated and the attenuation coefficient, a,
equal to the ambient PM concentration/ambient PM exposure were constant, the bias would be
given by pc = apA where pc is calculated using the ambient PM concentration, C, and PA is
calculated using the ambient PM exposure, A; i.e., the risk determined from an analysis using the
daily ambient PM concentrations would be lower than the risk obtained using ambient PM
exposure by the factor a. However, pc provides the correct information on the change in health
risks that would be produced by a change in ambient concentrations. However, if the daily,
individual values of ambient PM exposure and nonambient PM exposure are highly correlated
and the nonambient PM is toxic for the effect being studied, nonambient PM exposure will act as
a confounder and could introduce substantial bias into Pc. Only one study has reported the
association between the daily, individual values of ambient PM exposure and nonambient PM
exposure. Wilson et al. (2000), in a further analysis of the PTEAM data set, found a coefficient
of determination, R2 = 0.03, suggesting that the daily, individual values of ambient PM exposure
and nonambient PM exposure are independent, meaning that nonambient PM exposure will not
confound Pc.
5.5.3.2 Studies of Chronic Effects
The Six Cities (Dockery et al., 1993) and American Cancer Society (ACS) (Pope et al.,
1995) Studies have played an important role in assessing the health effects from long-term
5-117
-------
exposures to particulate pollution. Even though these studies often have been considered as
chronic epidemiologic studies, it is not easy to differentiate the role of historic exposures from
those of recent exposures on chronic disease mortality. In the Six Cities study, fine particles and
sulfates were measured at the community level, and the final analysis of the database used six
city-wide average ambient concentration measurements. This limitation also applies to the ACS
study but has less impact because of the larger number of cities considered in that study. In a
HEI-sponsored reanalysis of the Six Cities and the ACS data sets, Krewski et al. (2000)
examined some of the exposure misclassification issues either analytically or through sensitivity
analysis of the aerometric and health data. The HEI reanalysis project also addressed exposure
measurement error issues related to the Six Cities study. For example, the inability to account
for exposures prior to the enrollment of the cohort hampered accurate interpretation of the
relative risk estimates in terms of acute versus chronic causes. Although the results seem to
suggest that past exposures are more strongly associated with mortality than recent exposures,
the measurement error for long-term averages could be higher, influencing these interpretations.
As another example, Krewski et al. (2000), using the individual mobility data available for the
Six Cities cohort, analyzed the mover and nonmover groups separately. The relative risk of fine
particle effects on all-cause mortality was shown to be higher for the nonmover group than for
the mover group, suggesting the possibility of higher exposure misclassification biases for the
movers. The issue of using selected ambient monitors in the epidemiological analyses was also
investigated by the ACS and Six Cities studies reanalysis team. Krewski et al. (2000) described
the sensitivity of results to choices made in selecting stationary or mobile-source-oriented
monitors. For the ACS study, reanalysis of the sulfate data using only those monitors designated
as residential or urban and excluding sites designated as industrial, agricultural, or mobile did
not appreciably change the risk estimates. On the other hand, application of spatial analytic
methods designed to control confounding at larger geographic scales (i.e., between cities) caused
changes in the particle and sulfate risk coefficients. Spatial adjustment may account for
differences in pollution mix or PM composition, but many other cohort-dependent risk factors
will vary across regions or cities in the United States. Therefore, it is difficult to interpret these
findings solely in terms of spatial differences in pollution composition or relative PM toxicity
until further research is concluded.
5-118
-------
The influence of measurement errors in air pollution exposure and health effects
assessments has also been examined by Navidi et al. (1999). This study developed techniques to
incorporate exposure measurement errors encountered in long-term air pollution health-effects
studies and tested them on the data from the University of Southern California Children's Health
Study conducted in 12 communities in California. These investigators developed separate error
analysis models for direct (i.e., personal sampling) and indirect (i.e., microenvironmental)
personal exposure assessment methods. These models were generic to most air pollutants,
but a specific application was performed using a simulated data set for studying ozone health
effects on lung function decline in children. Because the assumptions made in their
microenvironmental simulation modeling framework were similar to those made in estimating
personal PM exposures, it is useful to consider the conclusions from Navidi et al. (1999).
According to Navidi et al. (1999), neither the microenvironmental nor the personal sampler
method produces reliable estimates of the exposure-response slope for O3 when measurement
error is uncorrected. Because of nondifferential measurement error, the bias was toward zero
under the assumptions made in Navidi et al. (1999) but could be away from zero if the
measurement error was correlated with the health response. A simulation analysis indicated that
the standard error of the estimate of a health effect increases as the errors in exposure assessment
increase (Navidi et al., 1999). According to Navidi et al. (1999), when a fraction of the ambient
level in a microenvironment is estimated with a standard error of 30%, the standard error of the
estimate is 50% higher than it would be if the true exposures were known. It appears that errors
in estimating ambient PM indoor/ambient PM outdoor ratios have much more influence on the
accuracy of the microenvironmental approach than do errors in estimating time spent in these
microenvironments.
Epidemiologic studies of chronic effects that use long-term average ambient PM
concentrations as the exposure metric generally do not address the nonambient component of
personal exposure (N). However, if N contributes to the health effect being studied and the
average N is different in different cities, the correlation between the average ambient PM
concentration and the health effect could be reduced. In an analysis of the effect of nonambient
exposure on time-series epidemiology, Dominici et al. (2000) examined nonambient exposure
data from several cities and concluded that the average nonambient PM exposure varies little
5-119
-------
among cities in developed countries. Also, Ott et al. (2000) examined estimated average and
daily, individual values of nonambient exposure from three studies and concluded that both the
average and the distribution of daily, individual values were similar for the three studies.
5.5.4 Conclusions from Analysis of Exposure Measurement Errors on
Particulate Matter Epidemiology
Personal exposures to PM are influenced by a number of factors and sources of PM located
in both indoor and outdoor microenvironments. However, PM resulting from ambient sources
penetrate into indoor environments such as residences, offices, public buildings, etc., in which
individuals spend a large portion of their daily lives. The correlations between total personal
exposures and ambient or outdoor PM concentrations can vary depending on the relative
contributions of indoor PM sources to total personal exposures. Panel studies of both adult and
young subjects have shown that, in fact, individual correlations of personal exposures with
ambient PM concentrations could vary from person to person, and even day to day, depending
on the specific activities of each person. Separation of PM exposures into two components,
ambient PM and nonambient PM, would reduce uncertainties in the analysis and interpretation
of PM health effects data. Nevertheless, because ambient PM is an integral component of total
personal exposures to PM, statistical analyses of cohort-average exposures are strongly
correlated with ambient PM concentrations when a large underlying population is studied.
Using the PTEAM study data, analysis of exposure measurement errors, in the context of time-
series epidemiology, has also shown that errors or uncertainties introduced by using surrogate
exposure variables, such as ambient PM concentrations, could lead to biases in the estimation of
health risk coefficients. These then would need to be corrected by suitable calibration of the PM
health risk coefficients. Correlations between the PM exposure variables and other covariates
(e.g., gaseous co-pollutants, weather variables) also could influence the degree of bias in the
estimated PM regression coefficients. However, most time-series regression models employ
seasonal or temporal detrending of the variables, thus reducing the magnitude of this cross-
correlation problem (Ozkaynak and Spengler, 1996).
Ordinarily, exposure measurement errors are not expected to influence the interpretation of
findings from either the chronic or time-series epidemiologic studies that have used ambient
concentration data if they include sufficient adjustments for seasonality and key confounders.
5-120
-------
There is no question that better estimates of exposures to components of PM of health concern
are beneficial. Composition of PM may vary in different geographic locations and different
exposure microenvironments. Compositional and spatial variations could lead to further errors
in using ambient PM measures as surrogates for exposures to PM. Even though the spatial
variability of PM (PM25 in particular) mass concentrations in urban environments seems to be
small, the same conclusions drawn above regarding the influence of measurement errors may not
necessarily hold for all of the toxic PM components. Again, the expectation based on statistical
modeling considerations is that these exposure measurement errors or uncertainties will most
likely reduce the statistical power of the PM health effects analysis, making it difficult to detect
a true underlying association between the correct exposure metric and the health outcome
studied. However, until more data on exposures to specific toxic components of PM become
available, existing studies on PM exposure measurement errors must be relied on; thus, at this
time, the use of ambient PM concentrations as a surrogate for exposures is not expected to
change the principal conclusions from PM epidemiologic studies that use community average
health and pollution data.
5.6 SUMMARY OF OBSERVATIONS AND LIMITATIONS
Exposure Definitions and Components
• Personal exposure to PM mass or its constituents results when individuals come in contact
with particulate pollutant concentrations in locations or microenvironments that they
frequent during a specific period of time. Various PM exposure metrics can be defined
according to its source (i.e., ambient, nonambient) and the microenvironment where
exposure occurs.
• Personal exposure to PM results from an individual's exposure to PM in many different
types of microenvironments (e.g., outdoors near home, outdoors away from home, indoors
at home, indoors at office or school, commuting, restaurants, malls, other public places).
Thus, total daily exposure to PM for a single individual can be expressed as the sum of
various microenvironmental exposures that the person encounters during the course of
a day.
• In a given microenvironment, particles may originate from a wide variety of sources. In an
indoor microenvironment, PM may originate from PM-generating activities (e.g., cooking,
cleaning, smoking, resuspending PM from PM resulting from both indoor and outdoor
sources that had settled out), from outside (outdoor PM entering through cracks and
5-121
-------
openings in the structure), and from the chemical interaction of pollutants from outdoor
air with indoor-generated pollutants.
The total daily exposure to PM for a single individual can also be expressed as the sum of
contributions of ambient and nonambient PM. Nonambient PM exposure is due to PM
generated by indoor sources, personal activities, and chemical reactions in indoor air.
Ambient PM exposure includes exposure to ambient PM while outdoors, and ambient PM
that has infiltrated indoors while indoors. However, within a large population, there will be
distributions of total personal exposure and its components due to variations in human
activities and microenvironmental concentrations and sources each individual encounters.
Exposure models are useful tools for examining the importance of sources, micro-
environments, and physical and behavioral factors that influence personal exposures to PM.
However, the development and evaluation of population exposure models for PM and its
components have been limited; and improved modeling methodologies and new model input
data are needed.
Factors Affecting Concentrations and Exposures to Particulate Matter
• Concentrations of PM indoors are affected by several factors and mechanisms:
ambient concentrations outdoors, air exchange rates, particle penetration factors,
particle production from indoor sources and indoor air chemistry, and indoor particle
decay rates and removal mechanisms caused by physical processes or resulting from
mechanical filtration, ventilation or air-conditioning devices.
• Average personal exposures to PM mass and its constituents are influenced by
microenvironmental PM concentrations and by how much time each individual spends
in these various indoor and outdoor microenvironments. Nationwide, individuals, on
average, spend nearly 90% of their time indoors (at home and in other indoor locations)
and about 6% of their time outdoors.
• Personal exposures are associated with both indoor as well as outdoor sources.
The personal exposure/outdoor concentration ratios present substantial intra- and
interpersonal variability. This variability is due to both the presence of personal
and microenvironmental sources and the varying effect of the outdoor particles on
indoor environments.
• Home characteristics may be the most important factor that affects the relationship
between the average population exposures and ambient concentrations. Air exchange rate
seems to be an important home characteristic surrogate that can explain a large fraction
of the observe inter- and intrapersonal variability. One reason why longitudinal studies
(many repeated measurements per person) provide stronger correlations between personal
exposure and outdoor concentrations than pooled studies (few repeated measurements per
individual) may be because home characteristics remain the same.
5-122
-------
Because home characteristics constitute the most important factor affecting personal
exposures, one would expect that correlations between average population exposures
and outdoor concentrations will vary by season and geography.
The relative size of personal exposure to ambient PM relative to nonambient PM depends
on the ambient concentration, the infiltration rate of outdoor PM into indoor
microenvironments, the amount of PM generated indoors (e.g., ETS, cooking and cleaning
emissions), and the amount of PM generated by personal activity sources. Infiltration
rates primarily depend on air exchange rate, size-dependent particle penetration across the
building membrane, and size-dependent removal rates. All of these factors vary over time
and across subjects and building types.
The relationship between PM exposure, dose, and health outcome could depend on the
concentration, composition, and toxicity of PM or PM components originating from
different sources. Application of source apportionment techniques to indoor and
outdoor PM2 5 and personal, indoor, and outdoor PM10 composition data have identified
the following general source categories: outside soil, resuspended indoor soil, indoor
soil, personal activities, sea salt, motor vehicles, nonferrous metal smelters, and
secondary sulfates.
Only a limited number of studies have measured the physical and chemical constituents of
PM in personal or microenvironmental samples. Available data on PM constituents
indicate the following:
- personal and indoor sulfate measurements often are correlated highly with outdoor and
ambient sulfate concentration measurements;
- for acid aerosols, indoor air chemistry is particularly important because of the
neutralization of the acidity by ammonia, which is present at higher concentrations
indoors because of the presence of indoor sources of ammonia;
- for SVOCs, including PAHs and phthalates, the presence of indoor sources will
substantially affect the relation between indoor and ambient concentrations;
- penetration and decay rates are functions of size and will cause variations in the
attenuation factors as a function of particle size (infiltration rates will be higher for
PMj and PM2 5 than for PM10, PM10_2 5 or ultrafme particles); and
- indoor air chemistry may increase indoor concentrations of organic PM.
Even though there is an increasing amount of research being performed to measure PM
constituents in different PM size fractions, with few exceptions (i.e., sulfur or sulfates) the
current data are inadequate to adequately assess the relationship between personal, indoor,
and ambient concentrations of most PM constituents.
5-123
-------
Correlations Between Personal Exposures, Indoor, Outdoor, and Ambient Measurements
• Most of the available personal data on PM measurements and information on the
relationships between personal and ambient PM come from a few large-scale studies (e.g.,
the PTEAM study) or the longitudinal panel studies that have been conducted on selected
populations, such as the elderly.
• Panel and cohort studies that have measured PM exposures and concentrations typically
have reported their results in terms of three types of correlations: (1) longitudinal,
(2) pooled, and (3) daily-average correlations between personal and ambient or outdoor PM.
• The type of correlation analysis performed can have a substantial effect on the resulting
correlation coefficient. Low correlations with ambient concentrations could result when
people with very different nonambient exposures are pooled even though their individual
personal exposures may temporally be highly correlated with ambient concentrations.
• Longitudinal and pooled correlations between personal exposure and ambient or outdoor
PM concentrations reported by various investigators varied considerably among the
different studies and in each study between the study subjects. Most studies report
longitudinal correlation coefficients that range from close to 0 to near 1.0, indicating that
individual's activities and residence type may have a significant effect on total personal
exposures to PM.
• Longitudinal studies that measured sulfate found high correlations between personal and
ambient sulfate.
• In general, probability-based population studies tend to show low pooled correlations
because of the high differences in levels of nonambient PM-generating activities from one
subject to another. In contrast, the absence of indoor sources for the populations in several
of the longitudinal panel studies resulted in high correlations between personal exposure
and ambient PM within subjects over time for these populations. However, even for these
studies, correlations varied by individual depending on their activities and the
microenvironments that they occupied.
Potential Sources of Error Resulting from Using Ambient Particulate Matter
Concentrations in Epidemiological Analyses
As yet, there is no clear consensus among exposure analysts as to how well community
monitor measurements of ambient air PM concentrations represent a surrogate for personal
exposure to total PM or to ambient PM.
• Measurement studies of personal exposures to PM are still few in number and limited in
spatial, temporal, and demographic coverage. Consequently, with the exception of a few
longitudinal panel studies, most epidemiological studies on PM health effects have relied
on daily-average PM concentration measurements obtained from ambient community
monitoring data as a surrogate for the exposure variable.
5-124
-------
Because individuals are exposed to particles in a multitude of indoor and outdoor
microenvironments during the course of a day, error is introduced in the estimation of
PM risk coefficients using ambient, as opposed to personal, PM measurements.
Total personal exposures to PM could vary from person to person, and even day to day,
depending on the specific activities of each person. Separation of PM exposures into two
components, ambient and nonambient PM, would reduce potential uncertainties in the
analysis and interpretation of PM health effects data.
Available data indicate that PM mass concentrations, especially fine PM, typically are
distributed relatively uniformly in most metropolitan areas, thus reducing the potential for
exposure misclassification because of spatial variability when a limited number of ambient
PM monitors are used to represent population average ambient exposures in community
time-series or long-term, cross-sectional epidemiological studies of PM.
Even though the spatial variability of PM (in particular, PM25) mass concentrations in urban
environments seems to be small, the same conclusions drawn above regarding the influence
of measurement errors may not necessarily hold for all PM components.
There are important differences in the relationship of ambient PM concentrations with
ambient PM exposures and with nonambient PM exposures. Various researchers have
shown that ambient PM exposure is a function of ambient PM concentration and that
concentrations of ambient PM are driven by meteorology, by changes in source emission
rates, and in locations of emission sources relative to the measurement site. However,
nonambient PM exposure is independent of ambient PM concentration because
concentrations of nonambient PM are driven by the daily activities of people.
Because personal exposures also include a contribution from ambient concentrations, the
correlation between daily-average personal exposure and the daily-average ambient
concentration increases as the number of subjects measured daily increases. An application
of a Random Component Superposition model has shown that the contributions of ambient
PM10 and indoor-generated PM10 to community mean exposure can be decoupled in
modeling urban population exposure distributions.
If linear nonthreshold models are assumed in time-series analysis of daily-average ambient
PM concentrations and community health data, nonambient PM exposure is not expected to
contribute to the relative risk estimates determined by regression of health responses on
ambient PM concentration.
Using the PTEAM study data, analysis of exposure measurement errors in the context of
time-series epidemiology has shown that the error introduced by using ambient PM
concentrations as a surrogate for ambient PM exposure biases the estimation of health risk
coefficients low by the ratio of ambient PM exposure to ambient PM concentration (called
the attenuation factor). However, the health risk coefficient determined using ambient PM
concentrations provides the correct information on the change in health risks that would be
produced by a change in ambient concentrations.
5-125
-------
Because sources and chemical composition of PM affecting personal exposures in different
microenvironments vary by season, day of the week, and time of day, it is likely that some
degree of misclassification of exposures to PM toxic agents of concern will be introduced
when health-effects models use only daily-average mass measures such as PM10or PM25.
Because of the paucity of currently available data on many of these factors, at this point it is
impossible to ascertain the significance of these more complex exposure misclassification
problems in the interpretation of results from PM epidemiology.
Exposure measurement errors may depend on particle size and composition. PM2 5 better
reflects personal exposure to PM of outdoor origin than PM10. Various indicators of
ultrafme particle concentrations or other components of PM may also be useful exposure
indicators for epidemiological studies.
Seasonal or temporal variations in the measurement errors and their correlations between
different PM concentration measures and co-pollutants (e.g., SO2, CO, NO2, O3) could
influence the error analysis results but are not likely to change the interpretation of current
findings.
Multipollutant personal exposure studies have suggested that ambient concentrations of
gaseous co-pollutants serve as surrogates of personal exposures to particles rather than as
confounders. The implications for epidemiology are discussed in Chapters 8 and 9.
Ordinarily, PM exposure measurement errors are not expected to influence the
interpretation of findings from either the community time-series or long-term epidemiologic
studies that have used ambient concentration data if they include sufficient adjustments for
seasonality and key personal and geographic confounders.
In the context of long-term epidemiological studies, it appears that the errors introduced in
estimating ambient PM indoor/ambient PM outdoor ratios have much more influence on the
accuracy of the microenvironmental exposure estimation approach than do errors in
estimating time spent in these microenvironments.
To reduce exposure misclassification errors in PM epidemiology, conducting new cohort
studies of sensitive populations with better real-time techniques for exposure monitoring
and further speciation of indoor-generated, ambient, and personal PM mass are essential.
Based on statistical modeling considerations, it is expected that existing PM exposure
measurement errors or uncertainties most likely will reduce the statistical power of the PM
health effects analysis, thus making it difficult to detect a true underlying association
between the correct exposure metric and the health outcome studied.
Although exposure measurement errors for fine particles are not expected to influence the
interpretation of findings from either the community time-series or the long-term, cross-
sectional epidemiological studies that have used ambient concentration data, they may
underestimate the strength of the effect. Sufficient data are not available to evaluate the
effect of exposure measurement error for other PM species or size fractions.
5-126
-------
Key Findings
• Most people spend most of their time indoors where they are exposed to indoor-generated
PM and ambient PM that has infiltrated indoors.
• Indoor-generated and ambient PM differ in sources, sizes, chemical composition, and
toxicity.
• The ambient PM concentration and the indoor PM concentration can be measured by
outdoor and indoor monitors. The total personal exposure can be measured by a personal
exposure monitor carried by the person. However, the concentrations of indoor-generated
PM and ambient PM that has infiltrated indoors and the related values of ambient and
nonambient PM exposures must be estimated.
• From a regression of individual, daily values of total personal exposure on daily PM
concentrations, the intercept gives the average nonambient PM exposure and the slope gives
the average attenuation factor (the ratio of ambient PM exposure to ambient PM
concentration).
• Similarly, from a regression of individual, daily values of indoor PM concentration on daily
ambient concentrations, the intercept gives the average concentration of indoor-generated
PM and the slope gives the average infiltration factor (concentrations of ambient PM that
has infiltrated indoors/ambient PM concentration).
• The attenuation factor and the infiltration factor depend on the penetration coefficient, the
fraction of ambient PM that penetrates through the walls, doors, windows, etc.; the
deposition or removal rate, a measure of how rapidly PM within the indoor
microenvironment is removed by deposition to surfaces or by filtration in a heating/cooling
system; and the air exchange rate, a measure of how rapidly indoor air is replaced by
outdoor air. The attenuation factor also depends on the fraction of time spent outdoors.
• The air exchange rate is an important variable for determining the concentration of ambient
PM found indoors. It can be measured by release and measurement of an inert tracer gas
indoors. The air exchange rate increases with the opening of windows or doors or the
operation of window or attic fans. It also increases as the indoor/outdoor temperature
difference increases. For closed homes, i.e., no open windows or doors, the air exchange
rate does not appear to be a function of wind speed or direction.
• The penetration coefficient and the deposition rate can be estimated from measurements of
outdoor and indoor concentrations under conditions when there are no indoor sources
(nighttime or unoccupied home). These parameters are functions of particle size. The
penetration coefficient is high and the deposition rate is low for accumulation mode particles
(0.1 to 1.0 |im). The penetration coefficient is lower and the deposition rate is higher for
ultrafine particles (< 0.1 jim) and coarse mode particles (> 1.0 jim). The attenuation factor
and the infiltration factor are higher for particles in the accumulation mode than for ultrafine
or coarse particles.
5-127
-------
The attenuation factor and the infiltration factor will vary with the air exchange rate and,
therefore, will vary with season and housing characteristics. These factors will increase with
increased opening of windows and doors. For closed homes, these factors will increase with
an increase of the indoor/outdoor temperature difference but they do not appear to be
affected by wind speed or direction.
The regression technique is useful for finding average values of the attenuation factor and
the nonambient exposure and possibly for estimating the distribution of individual, daily
values of the nonambient PM exposure.
Individual, daily values of the ambient PM exposure, the nonambient PM exposure, and the
attenuation factor may be determined from individual, daily values of the total PM personal
exposure and daily ambient PM concentrations by several techniques:
- Mass balance technique. Direct measurement of the air exchange rate, measurement of
the fraction of time spent outdoors by a diary of the subject's activity pattern, and use
of the equilibrium mass balance equation for the attenuation factor with estimated
values of the penetration coefficient and the deposition rate.
- Sulfate ratio technique. Individual, daily values of the attenuation factor (for PM2 5)
will be given by individual, daily values of personal exposure to sulfate / the daily
ambient sulfate concentration, provided that there are no indoor sources of sulfate and
sulfate and PM2 5 have similar particle size distributions.
- Recursive technique. Indoor-generated emissions, which tend to be episodic, can be
removed from a continuous record of indoor PM concentration, allowing separation of
indoor-generated PM and ambient PM that has infiltrated indoors.
In pooled studies (different subjects measured on different days), individual, daily values
of the total PM exposure are usually not well correlated with the daily ambient PM
concentrations. In longitudinal studies (each subject measured for multiple days),
individual, daily values of the total PM personal exposure and the daily ambient PM
concentrations are found to be highly correlated for some, but not all, subjects.
Only one study has reported estimated individual, daily values of ambient and nonambient
PM exposure. Individual, daily values of the total PM personal exposure and the daily
ambient PM concentrations were poorly correlated. However, individual, daily values of
ambient PM exposure and the daily ambient PM concentrations were highly correlated.
Individual daily values of ambient and nonambient PM exposure were not well correlated.
Individual daily values of nonambient PM exposure and daily ambient PM concentrations
were also not well correlated.
5-128
-------
REFERENCES
Abraham, J. L.; Siwinski, G.; Hunt, A. (2002) Ultrafme paniculate exposures in indoor, outdoor, personal and
mobile environments: effects of diesel, traffic, pottery kiln, cooking and HEP A filtration on
micro-environmental particle number concentration. Ann. Occup. Hyg. 46(suppl. 1): 406-411.
Abt, E.; Suh, H. H.; Allen, G.; Koutrakis, P. (2000a) Characterization of indoor particle sources: a study conducted
in the metropolitan Boston area. Environ. Health Perspect. 108: 35-44.
Abt, E.; Suh, H. H.; Catalano, P.; Koutrakis, P. (2000b) The relative contribution of outdoor and indoor particle
sources to indoor concentrations. Environ. Sci. Technol. 34: 3579-3587.
Adams, H. S.; Nieuwenhuijsen, M. J.; Colvile, R. N.; McMullen, M. A. S.; Khandelwal, P. (2001) Fine particle
(PM2 5) personal exposure levels in transport microenvironments, London, UK. Sci. Total Environ.
279: 29-44.
Adgate, J. L.; Ramachandran, G.; Pratt, G. C.; Waller, L. A.; Sexton, K. (2002) Spatial and temporal variability in
outdoor, indoor, and personal PM25 exposure. Atmos. Environ. 36: 3255-3265.
Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
Technol. 19: 911-918.
Al-Raheem, M. Y. A.; El-Desouky, M.; Mage, D. T.; Kollander, M.; Wallace, L. A.; Spengler, J. D. (2000) Personal
exposures to PM and gaseous air pollutants at the end of the 1991 Kuwait oil fires. Presented at: PM2000:
paniculate matter and health—the scientific basis for regulatory decision-making, specialty conference &
exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
Aim, S.; Jantunen, M. J.; Vartiainen, M. (1999) Urban commuter exposure to particle matter and carbon monoxide
inside an automobile. J. Exposure Anal. Environ. Epidemiol. 9: 237-244.
Ames, R. B.; Hand, J. L.; Kreidenweis, S. M. (2000) Optical measurements of aerosol size distributions in Great
Smoky Mountains National Park: dry aerosol characterization. J. Air Waste Manage. Assoc. 50: 665-676.
Anuszewski, J.; Larson, T. V.; Koenig, J. Q. (1998) Simultaneous indoor and outdoor particle light-scattering
measurements at nine homes using a portable nephelometer. J. Exposure Anal. Environ. Epidemiol.
8: 483-493.
Babich, P.; Davey, M.; Allen, G.; Koutrakis, P. (2000) Method comparisons for paniculate nitrate, elemental carbon,
and PM2 5 mass in seven U.S. cities. J. Air Waste Manage. Assoc. 50: 1095-1105.
Baek, S.-O.; Kim, Y.-S.; Perry, R. (1997) Indoor air quality in homes, offices and restaurants in Korean urban
areas-indoor/outdoor relationships. Atmos. Environ. 31: 529-544.
Bahadori, T. (1998) Human paniculate exposure assessment: relationship between outdoor, indoor, and personal
measurements (air pollution) [dissertation]. Boston, MA: Harvard University.
Bohadana, A. B.; Massin, N.; Wild, P.; Toamain, J.-P.; Engel, S.; Goutet, P. (2000) Symptoms, airway
responsiveness, and exposure to dust in beech and oak wood workers. Occup. Environ. Med. 57: 268-273.
Brauer, M. (1995) Assessment of indoor aerosols with an integating nephelometer. J. Exposure Anal. Environ.
Epidemiol. 5: 45-56.
Brauer, M.; Mannetje, A. (1998) Restaurant smoking restrictions and environmental tobacco smoke exposure. Am.
J. Public Health 88: 1834-1836.
Brauer, M.; Koutrakis, P.; Spengler, J. D. (1989) Personal exposures to acidic aerosols and gases. Environ. Sci.
Technol. 23: 1408-1412.
Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1990) Indoor and outdoor concentrations of acidic aerosols
and gases. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and climate,
volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
International Conference on Indoor Air Quality and Climate, Inc.; pp. 447-452.
Brauer, M.; Bartlett, K.; Regalado-Pineda, J.; Perez-Padilla, R. (1996) Assessment of paniculate concentrations from
domestic biomass combustion in rural Mexico. Environ. Sci. Technol. 30: 104-109.
Brauer, M.; Hirtle, R. D.; Hall, A. C.; Yip, T. R. (1999) Monitoring personal fine particle exposure with a particle
counter. J. Exposure Anal. Environ. Epidemiol. 9: 228-236.
Brauer, M.; Hruba, F.; Mihalikova, E.; Fabianova, E.; Miskovic, P.; Plzikova, A.; Lendacka, M.; Vanderberg, J.;
Cullen, A. (2000) Personal exposure to particles in Banska Bystrica, Slovakia. J. Exposure Anal. Environ.
Epidemiol. 10: 478-487.
5-129
-------
Brown, K. G.; Paxton, M. B. (1998) Predictability of personal exposure to airborne paniculate matter from ambient
concentrations in four communities. In: Vostal, J. I, ed. Health effects of paniculate matter in ambient air.
Proceedings of an international conference; 1997; Prague, Czech Republic. Pittsburgh, PA: Air& Waste
Management Association; pp. 333-350. (A&WMA publication VTP-80).
Burke, J.; Zufall, M.; Ozkaynak, H. (2001) A population exposure model for paniculate matter: case study results
forPM25 in Philadelphia, PA. J. Exposure Anal. Environ. Epidemiol. 11: 470-489.
Burton, R. M.; Suh, H. H.; Koutrakis, P. (1996) Spatial variation in paniculate concentrations within metropolitan
Philadelphia. Environ. Sci. Technol. 30: 400-407.
Buzorius, G.; Hameri, K.; Pekkanen, J.; Kulmala, M. (1999) Spatial variation of aerosol number concentration in
Helsinki city. Atmos. Environ. 33: 553-565.
Byrne, M. A.; Lange, C.; Goddard, A. J. H.; Roed, J. (1992) Indoor aerosol deposition studies using
neutron-activatable tracers. J. Aerosol Sci. 23(suppl. 1): S543-S546.
Calder, K. L. (1957) A numerical analysis of the protection afforded by buildings against BW aerosol attack.
Fort Detrick, MD: Office of the Deputy Commander for Scientific Activities; BWL tech. study no. 2.
Carrer, P.; Alcini, D.; Cavallo, D. (1997) Daily personal exposure to air pollutants of office workers in Milano. In:
Woods, et al. (eds.), Proceedings of Healthy Buildings '97, v.2, pp. 249-254; Washington, DC.
Carrer, P.; Alcini, D.; Cavallo, D.; Bollini, D.; Ghalandar, R.; Lovato, L.; Ploia, E.; Vercelli, F.; Visigalli, F.;
Maroni, M. (1998) Personal exposure to PM10 of subjects living in Milan. In: Vostal, J. J., ed. Health effects
of paniculate matter in ambient air. Proceedings of an international conference; 1997; Prague, Czech
Republic. Pittsburgh, PA: Air & Waste Management Association; pp. 359-365. (A&WMA publication
VIP-80).
Cha, Q.; Chen, Y.-Z.; Du, Y.-X. (1996) Influence of indoor air pollution on quality of outdoor air in city of
Guangzhou, China. In: Indoor air '96: the 7th international conference on indoor air quality, volume I; July;
Nagoya, Japan. Tokyo, Japan: Institute of Public Health; pp. 1131-1135.
Chan, A. T. (2002) Indoor-outdoor relationships of paniculate matter and nitrogen oxides under different outdoor
meteorological conditions. Atmos. Environ. 36: 1543-1551.
Chan, L. Y.; Lau, W. L.; Lee, S. C.; Chan, C. Y. (2002) Commuter exposure to paniculate matter in public
transportation modes in Hong Kong. Atmos. Environ. 36: 3363-3373.
Chang, L.-T.; Koutrakis, P.; Catalano, P. J.; Suh, H. H. (2000) Hourly personal exposures to fine particles and
gaseous pollutants-results from Baltimore, Maryland. J. Air Waste Manage. Assoc. 50: 1223-1235.
Chang, L.-T.; Suh, H. H.; Wolfson, J. M.; Misra, K.; Allen, G. A.; Catalano, P. J.; Koutrakis, P. (2001) Laboratory
and field evaluation of measurement methods for one-hour exposures to O3, PM2 5, and CO. J. Air Waste
Manage. Assoc. 51: 1414-1422.
Chao, C. Y. H.; Tung, T. C. (2001) An empirical model for outdoor contaminant transmission into residential
buildings and experimental verification. Atmos. Environ. 35: 1585-1596.
Chuang, J. C.; Callahan, P. J.; Lyu, C. W.; Wilson, N. K. (1999) Fob/cyclic aromatic hydrocarbon exposures of
children in low-income families. J. Exposure Anal. Environ. Epidemiol. 2: 85-98.
Clayton, C. A.; Perritt, R. L.; Pellizzari, E. D.; Thomas, K. W.; Whitmore, R. W.; Wallace, L. A.; Ozkaynak, H.;
Spengler, J. D. (1993) Particle total exposure assessment methodology (PTEAM) study: distributions of
aerosol and elemental concentrations in personal, indoor, and outdoor air samples in a southern California
community. J. Exposure Anal. Environ. Epidemiol. 3: 227-250.
Clayton, C. A.; Pellizzari, E. D.; Rodes, C. E.; Mason, R. E.; Piper, L. L. (1999a) Estimating distributions of
long-term paniculate matter and manganese exposures for residents of Toronto, Canada. Atmos. Environ.
33:2515-2526.
Clayton, C. A.; Pellizzari, E. D.; Whitmore, R. W.; Perritt, R. L., Quackenboss, J. J. (1999b) National human
exposure assessment survey (NHEXAS): distributions and associations of lead, arsenic, and volatile organic
compounds in EPA Region 5. J. Exposure Anal. Environ. Epidemiol. 9: 381-392.
Conner, T. L.; Lonneman, W. A.; Seila, R. L. (1995) Transportation-related volatile hydrocarbon source profiles
measured in Atlanta. J. Air Waste Manage. Assoc. 45: 383-394.
Conner, T. L.; Norris, G. A.; Landis, M. S.; Williams, R. W. (2001) Chemical and physical characterization of
indoor, outdoor, and personal samples collected in and around a retirement facility. In: Measurement of Toxic
and Related Air Pollutants: proceedings of a specialty conference; September, 2000; Research Triangle Park,
NC. Pittsburgh, PA: (A&WMA publication VTP-100/CD; pp. 678-700).
Copeland, K. T.; Checkoway, H.; McMichael, A. J.; Holbrook, R. H. (1977) Bias due to misclassification in the
estimation of relative risk. Am. J. Epidemiol. 105: 488-495.
5-130
-------
Crump, K. S. (2000) Manganese exposures in Toronto during use of the gasoline additive, methylcyclopentadienyl
manganese tricarbonyl. J. Exposure Anal. Environ. Epidemiol. 10: 227-239.
Daisey, J. M; Mahanama, K. R. R.; Hodgson, A. T. (1998) Toxic volatile organic compounds in simulated
enviromental tobacco smoke: emission factors for exposure assessment. J. Exposure Anal. Environ.
Epidemiol. 8:313-334.
Demokritou, P.; Kavouras, I. G.; Harrison, D.; Koutrakis, P. (2001) Development and evaluation of an impactor for
a PM25 speciation sampler. J. Air Waste Manage. Assoc. 51: 514-523.
Dockery, D. W.; Spengler, J. D. (1981) Indoor-outdoor relationships of respirable sulfates and particles.
Atmos. Environ. 15: 335-343.
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.
(1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329: 1753-1759.
Dominici, F.; Zeger, S. L.; Samet, J. (2000) A measurement error model for time-series studies of air pollution and
mortality. Biostatistics 1: 157-175.
Donham, K. J.; Cumro, D.; Reynolds, S. J.; Merchant, J. A. (2000) Dose-response relationships between
occupational aerosol exposures and cross-shift declines of lung function in poultry workers: recommendations
for exposure limits. J. Occup. Environ. Med. 42: 260-269.
Douce,D. S.; Clench, M. R.; Frost, B. (2001) Variations in the estimation of the contribution of environmental
tobacco smoke (ETS) to respirable (< 5|_im) indoor air particulates obtained by the use of different analytical
methods. J. Environ. Monit. 3: 295-301.
Duan, N. (1982) Models for human exposure to air pollution. Environ. Int. 8: 305-309.
Duan, N. (1989) Estimation of microenvironment concentration distribution using integrated exposure
measurements. In: Starks, T. H., ed. Proceedings of the Research Planning Conference on Human Activity
Patterns. Las Vegas, NV: U.S. Environmental Protection Agency, Environmental Monitoring Systems
Laboratory; pp. 15-1 - 15-12; report no. EPA/600/4-89-004. Available from: NTIS, Springfield, VA;
PB89-166599.
Duan, N. (1991) Stochastic microenvironment models for air pollution exposure. J. Exposure Anal. Environ.
Epidemiol. 1:235-257.
Dubowsky, S. D.; Wallace, L. A.; Buckley, T. J. (1999) The contribution of traffic to indoor concentrations of
polycyclic aromatic hydrocarbons. J. Exposure Anal. Environ. Epidemiol. 9: 312-321.
Ebelt, S. T.; Petkau, A. J.; Vedal, S.; Fisher, T. V.; Brauer, M. (2000) Exposure of chronic obstructive pulmonary
disease patients to paniculate matter: relationships between personal and ambient air concentrations. J. Air
Waste Manage. Assoc. 50: 1081-1094.
Esmen, N. A.; Hall, T. A. (2000) Theoretical investigation of the interrelationship between stationary and personal
sampling in exposure estimation. Appl. Occup. Environ. Hyg. 15: 114-119.
Evans, G. F.; Highsmith, R. V; Sheldon, L. S.; Suggs, J. C.; Williams, R. W.; Zweidinger, R. B.; Creason, J. P.;
Walsh, D.; Rodes, C. E.; Lawless, P. A. (2000) The 1999 Fresno paniculate matter exposure studies:
comparison of community, outdoor, and residential PM mass measurements. J. Air Waste Manage. Assoc.
50: 1887-1896.
Ezzati, M.; Kammen, D. M. (2001) Quantifying the effects of exposure to indoor air pollution from biomass
combustion on acute respiratory infections in developing countries. Environ. Health Perspect. 109: 481-488.
Federal Register. (1986) Guidelines for estimating exposures. F. R. (September 24) 51: 34,042-34,054.
Ferris, B. G., Jr.; Speizer, F. E.; Spengler, J. D.; Dockery, D.; Bishop, Y. M. M.; Wolfson, M.; Humble, C. (1979)
Effects of sulfur oxides and respirable particles on human health: methodology and demography of
populations in study. Am. Rev. Respir. Dis. 120: 767-779.
Fischer, P. H.; Hoek, G.; VanReeuwijk, H.; Briggs, D. J.; Lebret, E.; Van Wijnen, J. H.; Kingham, S.; Elliott, P. E.
(2000) Traffic-related differences in outdoor and indoor concentrations of particles and volatile organic
compounds in Amsterdam. Atmos. Environ. 34: 3713-3722.
Fogh, C. L.; Byrne, M. A.; Roed, J.; Goddard, A. J. H. (1997) Size specific indoor aerosol deposition measurements
and derived I/O concentrations ratios. Atmos. Environ. 31: 2193-2203.
Freijer, J. I.; Bloemen, H. J. T. (2000) Modeling relationships between indoor and outdoor air quality. J. Air Waste
Manage. Assoc. 50: 292-300.
Frey, H. C.; Rhodes, D. S. (1996) Characterizing, simulating and analyzing variability and uncertainty:
an illustration of methods using an air toxic emissions example. Hum. Ecol. Risk Assess. 2: 762-797'.
Geller, M. D.; Chang, M.; Sioutas, C.; Ostro, B. D.; Lipsett, M. J. (2002) Indoor/outdoor relationship and chemical
composition of fine and coarse particles in the southern California deserts. Atmos. Environ. 36: 1099-1110.
5-131
-------
Georgopoulos, P. G.; Lioy, P. J. (1994) Conceptual and theoretical aspects of human exposure and dose assessment.
J. Exposure Anal. Environ. Epidemiol. 4: 253-285.
Gladen, B.; Rogan, W. J. (1979) Misclassification and the design of environmental studies. Am. J. Epidemiol.
109: 607-616.
Glen, G.; Lakkadi, Y.; Tippett, J. A.; del Valle-Torres, M. (1997) Development of NERL/CHAD: the National
Exposure Research Laboratory consolidated human activity database. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development; contract no. 68-D5-0049.
Goldstein, B. R.; Tardiff, G.; Hoffnagle, G.; Kester, R. (1992) Valdez health study: summary report. Anchorage,
AK: Alyeska Pipeline Service Company.
Goswami, E.; Larson, T.; Lumley, T.; Sally, Liu, L.-J. (2002) Spatial characteristics of fine paniculate matter:
identifying representative monitoring locations in Seattle, Washington. J. Air Waste Manage. Assoc.
52: 324-333.
Gotschi, T.; Oglesby, L.; Mathys, P.; Monn, C.; Manalis, N.; Koistinen, K.; Jantunen, M.; Hanninen, O.;
Polanska, L.; Kunzli, N. (2002) Comparison of black smoke and PM2 5 levels in indoor and outdoor
environments of four European cities. Environ. Sci. Technol. 36: 1191-1197.
Hadwen, G. E.; McCarthy, J. F.; Womble, S. E.; Girman, J. R.; Brightman, H. S. (1997) Volatile organic compound
concentrations in 41 office buildings in the continental United States. In: Woods, J. E.; Grimsrud, D. T.;
Boschi, N., eds. Proceedings of healthy buildings/indoor air quality international conference; September;
Bethesda, MD. Washington, DC: Healthy Buildings/IAQ. [Healthy Buildings - International Conference; v. 2,
pp. 465-470.]
Harrison, R. M.; Thornton, C. A.; Lawrence, R. G.; Mark, D.; Kinnersley, R. P.; Ayres, J. G. (2002) Personal
exposure monitoring of paniculate matter, nitrogen dioxide, and carbon monoxide, including susceptible
groups. Occup. Environ. Med. 59: 671-679.
Hartwell, T. D.; Clayton, C. A.; Michie, R. M., Jr.; Whitmore, R. W.; Zelon, H. S.; Whitehurst, D. A.; Akland, G. G.
(1984) Study of carbon monoxide exposures of residents of Washington, D.C. Research Triangle Park, NC:
U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
EPA-600/D-84-177. Available from: NTIS, Springfield, VA; PB84-230051.
Hayes, S. R.; Marshall, J. R. (1999) Designing optimal strategies to attain the new U.S. paniculate matter standards:
some initial concepts. J. Air Waste Manage. Assoc. 49: PM192-PM198.
Henry, R. C.; Lewis, C. W.; Collins, J. F. (1994) Vehicle-related source compositions from ambient data:
the GRACE/SAFER method. Environ. Sci. Technol. 28: 823-832.
Hoek, G.; Fischer, P.; Van Den Brandt, P.; Goldbohm, S.; Brunekreef, B. (2001) Estimation of long-term average
exposure to outdoor air pollution for a cohort study on mortality. J. Exposure Anal. Environ. Epidemiol.
11:459-469.
Hoek, G.; Brunekreef, B.; Goldbohm, S.; Fischer, P.; Van den Brandt, P. A. (2002) Association between mortality
and indicators of traffic-related air pollution in the Netherlands: a cohort study. Lancet 360: 1203-1209.
Houseman, E. A.; Ryan, L.; Levy, J.; Richardson, D.; Spenger, J. D. (2002) Autocorrelation in real-time continuous
monitoring of microenvironments. J. Appl. Stat. 29: 855-872.
Howard-Reed, C.; Rea, A. W.; Zufall, M. J.; Burke, J. M.; Williams, R. W.; Suggs, J. C.; Sheldon, L. S.; Walsh, D.;
Kwock, R. (2000) Use of a continuous nephelometer to measure personal exposure to particles during the
U.S. Environmental Protection Agency Baltimore and Fresno panel studies. J. Air Waste Manage. Assoc.
50: 1125-1132.
Howard-Reed, C.; Wallace, L. A.; Ott, W. R. (2002) The effect of opening windows on air change rates in two
homes. J. Air Waste Manage Assoc. 52: 147-159.
Institute for Social Research. (1997) Child development supplement to the panel study of income dynamics.
Ann Arbor, MI: University of Michigan. Available: www.isr.umich.edu/src/child-development/home.html
[2001, January 31].
Isukapalli, S. S.; Georgopoulos, P. G. (2000) Modeling of outdoor/indoor relationships of gas phase pollutants and
paniculate matter. Presented at: PM2000: paniculate matter and health—the scientific basis for regulatory
decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste
Management Association.
Jang, M.; Kamens, R. M. (1999) Newly characterized products and composition of secondary aerosols from the
reaction of a-pinene with ozone. Atmos. Environ. 33: 459-474.
Janssen, N. (1998) Personal exposure to airborne particles: validity of outdoor concentrations as a measure of
exposure in time series studies [dissertation]. Wageningen, Netherlands: Landbouwuniversiteit Wageningen.
5-132
-------
Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1995) A relation between personal and ambient.
Epidemiology 6(suppl.): S45.
Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1997) Childhood exposure to PM10: relation between
personal, classroom, and outdoor concentrations. Occup. Environ. Med. 54: 888-894.
Janssen, N. A. H.; Hoek, G.; Brunekreef, B.; Harssema, H.; Mensink, I.; Zuidhof, A. (1998a) Personal sampling of
particles in adults: relation among personal, indoor, and outdoor air concentrations. Am. J. Epidemiol.
147: 537-547.
Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1998b) Personal sampling of airborne particles:
method performance and data quality. J. Exposure Anal. Environ. Epidemiol. 8: 37-49.
Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1999a) Personal exposure to fine particles in children
correlates closely with ambient fine particles. Arch. Environ. Health 54: 95-101.
Janssen, N. A. H.; Hoek, G.; Brunekreef, B.; Harssema, H. (1999b) Mass concentration and elemental composition
of PM10 in classrooms. Occup. Environ. Med. 56: 482-487.
Janssen, L. H. J. M.; Buringh, E.; van der Meulen, A.; van den Hout, K. D. (1999c) A method to estimate the
distribution of various fractions of PM10 in ambient air in the Netherlands. Atmos. Environ. 33: 3325-3334.
Janssen, N. A. H.; De Hartog, J. J.; Hoek, G.; Brunekreef, B.; Lanki, T.; Timonen, K. L.; Pekkanen, J. (2000)
Personal exposure to fine paniculate matter in elderly subjects: relation between personal, indoor, and outdoor
concentrations. J. Air Waste Manage. Assoc. 50: 1133-1143.
Jantunen, M. J.; Hanninen, O.; Katsouyanni, K.; Knoppel, H.; Kuenzli, N.; Lebret, E.; Maroni, M.; Saarela, K.;
Sram, R.; Zmirou, D. (1998) Air pollution exposure in European cities: the EXPOLIS study. J. Exposure
Anal. Environ. Epidemiol. 8: 495-518.
Jantunen, M. J.; Kousa, A.; Hanninen, O.; Koistinen, K.; Sram, R.; Lanki, T.; Lioy, P. J. (2000) Long term
population PM exposures correlate well with outdoor air levels - short term individual exposures do not.
Presented at: PM2000: paniculate matter and health—the scientific basis for regulatory decision-making,
specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management
Association.
Jenkins, R. A.; Palausky, M. A.; Counts, R. W.; Guerin, M. R.; Dindal, A. B.; Bayne, C. K. (1996a) Determination
of personal exposure of non-smokers to environmental tobacco smoke in the United States. Lung Cancer
14(suppl. 1): S195-S213.
Jenkins, R. A.; Palausky, A.; Counts, R. W.; Bayne, C. K.; Dindal, A. B.; Guerin, M. R. (1996b) Exposure to
environmental tobacco smoke in sixteen cities in the United States as determined by personal breathing zone
air sampling. J. Exposure Anal. Environ. Epidemiol. 6: 473-502.
Jinsart, W.; Tamura, K.; Loetkamonwit, S.; Thepanondh, S.; Karita, K.; Yano, E. (2002) Roadside paniculate air
pollution in Bangkok. J. Air Waste Manage. Assoc. 52: 1102-1110.
Johnson, T. (1984) A study of personal exposure to carbon monoxide in Denver, Colorado. Research Triangle Park,
NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
EPA-600/4-84-014.
Johnson, T. (1989) Human activity patterns in Cincinnati, Ohio [final report]. Palo Alto, CA: Electric Power
Research Institute; project no. RP940-06; report no. EN-6204.
Johnson, T. R. (1995) Recent advances in the estimation of population exposure to mobile source pollutants.
J. Exposure Anal. Environ. Epidemiol. 5: 551-571.
Johnson, T.; Long, T.; Ollison, W. (2000) Prediction of hourly microenvironmental concentrations of fine particles
based on measurements obtained from the Baltimore scripted activity study. J. Exposure Anal. Environ.
Epidemiol. 10:403-411.
Jones, N. C.; Thornton, C. A.; Mark, D.; Harrison, R. M. (2000) Indoor/outdoor relationships of paniculate matter in
domestic homes with roadside, urban and rural locations. Atmos. Environ. 34: 2603-2612.
Kamens, R.; Lee, C.-T.; Wiener, R.; Leith, D. (1991) A study to characterize indoor particles in three non-smoking
homes. Atmos. Environ. Part A 25: 929-948.
Keeler, G. J.; Dvonch, J. T.; Yip, F. Y.; Parker, E. A.; Israel, B. A.; Marsik, F. J.; Morishita, M.; Barres, J. A.;
Robins, T. G.; Brakefield-Caldwell, W.; Sam, M. (2002) Assessment of personal and community-Level
exposures to paniculate matter among children with asthma in Detroit, Michigan, as part of Community
Action Against Asthma (CAAA). Environ. Health Perspect. 110(suppl. 2): 173-181
Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
segregated mass concentration data and ultrafine particle number concentrations in urban areas.
Atmos. Environ. 33: 2907-2913.
5-133
-------
Kingham, S.; Briggs, D.; Elliott, P.; Fischer, P.; Lebret, E. (2000) Spatial variations in the concentrations of
traffic-related pollutants in indoor and outdoor air in Huddersfield, England. Atmos. Environ. 34: 905-916.
Kinney, P. L.; Chillrud, S. N.; Ramstrom, S.; Ross, I; Spengler, J. D. (2002) Exposures to multiple air toxics in
New York City. Environ. Health Perspect. 110(suppl. 4): 539-546.
Kleeman, M. J.; Cass, G. R. (1998) Source contributions to the size and composition distribution of urban paniculate
air pollution. Atmos. Environ. 32: 2803-2816.
Klepeis, N. E. (1999) An introduction to the indirect exposure assessment approach: modeling human exposure
using microenvironmental measurements and the recent national human activity pattern survey. Environ.
Health Perspect. Suppl. 107(2): 365-374.
Klepeis, N. E.; Ott, W. R.; Switzer, P. (1994) A Total Human Exposure Model (THEM) for Respirable Suspended
Particles (RSP). Presented at: 87th annual meeting & exhibition of the Air & Waste Management Association;
June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no. 94-WA75A.03.
Available from: NTIS, Springfield, VA; PB94-197415.
Klepeis, N. E.; Ott, W. R.; Switzer, P. (1995) Modeling the time series of respirable suspended particles and carbon
monoxide from multiple smokers: validation in two public smoking lounges. Presented at: 88th annual
meeting and exhibition of the Air & Waste Management Association; June; San Antonio, TX. Pittsburgh, PA:
Air & Waste Management Association; paper no. 95-MPH.03.
Klepeis, N. E.; Ott, W. R.; Switzer, P. (1996) A multiple-smoker model for predicting indoor air quality in public
lounges. Environ. Sci. Technol. 30: 2813-2820.
Klepeis, N. E.; Switzer, P.; Ott, W. R. (2000) Estimating exposure to environmental tobacco smoke particles using
24-hour time-activity recall diaries. Presented at: PM2000: paniculate matter and health—the scientific basis
for regulatory decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA:
Air & Waste Management Association.
Koistinen, K. J.; Kousa, A.; Tenhola, V.; Hanninen, O.; Jantunen, M. J.; Oglesby, L.; Kuenzli, N.; Georgoulis, L.
(1999) Fine particle (PM2 5) measurement methodology, quality assurance procedures and pilot results of the
EXPOLIS study. J. Air Waste Manage. Assoc. 49: 1212-1220.
Koontz, M. D.; Rector, H. E. (1995) Estimation of distributions for residential air exchange rates: final report.
Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics; contract
nos. 68-D9-0166 and 68-D3-0013.
Koontz, M. D.; Niang, L. L. (1998) California population indoor exposure model (CPIEM): version 1.4F. Final
report A933-157: user's guide. Sacramento, CA: California Air Resources Board; GEOMET report no.
IE-2632.
Korenstein, S.; Piazza, B. (2002) An exposure assessment of PM10 from a major highway interchange: are children
in nearby schools at risk? J. Environ. Health 65: 9-17.
Kousa, A.; Oglesby, L.; Koistinen, K.; Ktinzli, N. Jantunen, M. (2002) Exposure chain of urban air
PM2 ^-associations between ambient fixed site, residential outdoor, indoor, workplace and personal exposures
in four European cities in the EXPOLIS-study. Atmos. Environ. 36: 3031-3039.
Koutrakis, P.; Briggs, S. L. K.; Leaderer, B. P. (1992) Source apportionment of indoor aerosols in Suffolk and
Onondaga Counties, New York. Environ. Sci. Technol. 26: 521-527.
Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J.; Jerrett, M.; Abrahamowicz, M.;
White, W. H. (2000) Reanalysis of the Harvard Six Cities study and the American Cancer Society study of
paniculate air pollution and mortality: a special report of the Institute's Particle Epidemiology Reanalysis
Project. Cambridge, MA: Health Effects Institute.
Lachenmyer, C.; Hidy, G. M. (2000) Urban measurements of outdoor-indoor PM2 5 concentrations and personal
exposure in the deep south. Part I. Pilot study of mass concentrations for nonsmoking subjects. Aerosol Sci.
Technol. 32: 34-51.
Laden, F.; Neas, L. M.; Dockery, D. W.; Schwartz, J. (2000) Association of fine paniculate matter from different
sources with daily mortality in six U.S. cities. Environ. Health Perspect. 108: 941-947.
Landis, M. S.; Norris, G. A.; Williams, R. W.; Weinstein, J. P. (2001) Personal exposures to PM2 5 mass and trace
elements in Baltimore, MD, USA. Atmos. Environ. 35: 6511-6524.
Lanki, T.; Aim, S.; Ruuskanen, J.; Janssen, N. A. H.; Jantunen, M.; Pekkanen, J. (2002) Photometrically measured
continuous personal PM25 exposure: levels and correlation to a gravimetric method. J. Exposure Anal.
Environ. Epidemiol. 12: 172-178.
5-134
-------
Leaderer, B. P.; Naeher, L.; Jankun, T.; Balenger, K.; Holford, T. R.; Toth, C.; Sullivan, I; Wolfson, J. M.;
Koutrakis, P. (1999) Indoor, outdoor, and regional summer and winter concentrations of PM10, PM2 5, SO42~,
H+, NH4+, NO3", NH3, and nitrous acid in homes with and without kerosene space heaters. Environ. Health
Perspect. 107:223-231.
Lee, S. C.; Chang, M. (1999) Indoor air quality investigations at five classrooms. Indoor Air 9: 134-138.
Lena, T. S.; Ochieng, V.; Carter, M.; Holguin-Veras, J.; Kinney, P. L. (2002) Elemental carbon and PM25 levels in
an urban community heavily impacted by truck traffic. Environ. Health Perspect. 110: 1009-1015.
Lessler, J. T.; Kalsbeek, W. D. (1992) Nonsampling error in surveys. New York, NY: John Wiley & Sons, Inc.;
p,efy, J. I.; Dumyahn, T.; Spengler, J. D. (2002) Paniculate matter and polycyclic aromatic hydrocarbon
concentrations in indoor and outdoor microenvironments in Boston, Massachusetts. J. Exposure Anal.
Environ. Epidemiol. 12: 104-114.
Liao, D.; Creason, J.; Shy, C.; Williams, R.; Watts, R.; Zweidinger, R. (1999) Daily variation of paniculate air
pollution and poor cardiac autonomic control in the elderly. Environ. Health Perspect. 107: 521-525.
Ligocki, M. P.; Liu, H. I. H.; Cass, G. R.; John, W. (1990) Measurements of particle deposition rates inside Southern
California museums. Aerosol Sci. Technol. 13: 85-101.
Lillquist, D. R.; Lee, J. S.; Ramsay, J. R.; Boucher, K. M.; Walton, Z. L.; Lyon, J. L. (1998) A comparison of
indoor/outdoor PM10 concentrations measured at three hospitals and a centrally located monitor in Utah. Appl.
Occup. Environ. Hyg. 13: 409-415.
Linn, W. S.; Gong, H., Jr.; Clark, K. W.; Anderson, K. R. (1999) Day-to-day paniculate exposures and health
changes in Los Angeles area residents with severe lung disease. J. Air Waste Manage. Assoc.
49:PM108-PM115.
Lioy, P. J. (1990) Assessing total human exposure to contaminants: a multidisciplinary approach. Environ. Sci.
Technol. 24: 938-945.
Lioy, P. J.; Waldman, J. M.; Buckley, T.; Butler, J.; Pietarinen, C. (1990) The personal, indoor and outdoor
concentrations of PM-10 measured in an industrial community during the winter. Atmos. Environ. Part B
24: 57-66.
Lioy, P. J.; Wainman, T.; Zhang, J.; Goldsmith, S. (1999) Typical household vacuum cleaners: the collection
efficiency and emissions characteristics for fine particles. J. Air Waste Manage. Assoc. 49: 200-206.
Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report no.
95.
Liu, D.-L.; Nazaroff, W. W. (2001) Modeling pollutant penetration across building envelopes. Atmos. Environ.
35: 4451-4462.
Liu, L.-J. S.; Slaughter, J. C.; Larson, T. V. (2002) Comparison of light scattering devices and impactors for
paniculate measurements in indoor, outdoor, and personal environments. Environ. Sci. Technol.
36: 2977-2986.
Long, C. M.; Suh, H. H.; Koutrakis, P. (2000) Characterization of indoor particle sources using continuous mass and
size monitors. J. Air Waste Manage. Assoc. 50: 1236-1250.
Long, C. M.; Suh, H. H.; Catalano, P. J.; Koutrakis, P. (2001a) Using time- and size-resolved paniculate data to
quantify indoor penetration and deposition behavior. Environ. Sci. Technol. 35: 2089-2099.
Long, C. M.; Suh, H. H.; Kobzik, L.; Catalano, P. J.; Ning, Y. Y.; Koutrakis, P. (200Ib) A pilot investigation of the
relative toxicity of indoor and outdoor fine particles: in-vitro effects of endotoxin and other paniculate
properties. Environ. Health Perspect. 109: 1019-1026.
Lurmann, F. W.; Korc, M. E. (1994) User's guide to the regional human exposure (REHEX) model. San Francisco,
CA: Bay Area Air Quality Management District; draft report no. STI-93150-1414-DR.
Macintosh, D. L.; Xue, J.; Ozkaynak, H.; Spengler, J. D.; Ryan, P. B. (1995) A population-based exposure model for
benzene. J. Exposure Anal. Environ. Epidemiol. 5: 375-403.
Magari, S. R.; Schwartz, J.; Williams, P. L.; Hauser, R.; Smith, T. J.; Christiani, D. C. (2002) The association
between personal measurements of environmental exposure to particulates and heart rate variability.
Epidemiology 13: 305-310.
Mage, D. T. (1985) Concepts of human exposure assessment for airborne paniculate matter. Environ. Int.
11:407-412.
Mage, D. T. (1998) How much PM of ambient origin are people exposed to? In: Measurement of toxic and related
air pollutants. Volume II: proceedings of a specialty conference; September; Gary, NC. Pittsburgh, PA: Air&
Waste Management Association; pp. 666-681. (A&WMA publication VTP-85).
5-135
-------
Mage, D. T.; Ott, W. R. (1996) Accounting for nonuniform mixing and human exposure in indoor environments.
In: Tichenor, B. A., ed. Characterizing sources of indoor air pollution and related sink effects. Philadelphia,
PA: American Society for Testing and Materials; pp. 263-278.
Mage, D.; Wilson, W.; Hasselblad, V.; Grant, L. (1999) Assessment of human exposure to ambient paniculate
matter. J. Air Waste Manage. Assoc. 49: 174-185.
Mar, T. F.; Norris, G. A.; Koenig, J. Q.; Larson, T. V. (2000) Associations between air pollution and mortality in
Phoenix, 1995-1997. Environ. Health Perspect. 108: 347-353.
Marple, V. A.; Rubow, K. L.; Turner, W.; Spengler, J. D. (1987) Low flow rate sharp cut impactors for indoor air
sampling: design and calibration. JAPCA37: 1303-1307.
McBride, S. J.; Ferro, A. R.; Ott, W. R; Switzer, P.; Hildemann, L. M. (1999) Investigations of the proximity effect
for pollutants in the indoor environment. J. Exposure Anal. Environ. Epidemiol. 9: 602-621.
McCurdy, T. (1995) Estimating human exposure to selected motor vehicle pollutants using the NEM series of
models: lessons to be learned. J. Exposure Anal. Environ. Epidemiol. 5: 533-550.
McCurdy, T. (1997) Modeling the dose profile in human exposure assessments: ozone as an example. Rev. Toxicol.
1: 3-23.
McCurdy, T. (2000) Conceptual basis for multi-route intake dose modeling using an energy expenditure approach.
J. Exposure Anal. Environ. Epidemiol. 10: 86-97.
McCurdy, T.; Glen, G.; Smith, L.; Lakkadi, Y. (2000) The National Exposure Research Laboratory's Consolidated
Human Activity Database. J. Exposure Anal. Environ. Epidemiol. 10: 566-578.
Mihlan, G. J.; Todd, L. A.; Truong, K. N. (2000) Assessment of occupational exposure patterns by
frequency-domain analysis of time series data. Appl. Occup. Environ. Hyg. 15: 120-130.
Miller, S. L.; Branoff, S.; Nazaroff, W. W. (1998) Exposure to toxic air contaminants in environmental tobacco
smoke: an assessment for California based on personal monitoring data. J. Exposure Anal. Environ.
Epidemiol. 8:287-311.
Monn, C. (2001) Exposure assessment of air pollutants: a review on spatial heterogeneity and
indoor/outdoor/personal exposure to suspended paniculate matter, nitrogen dioxide and ozone. Atmos.
Environ. 35: 1-32.
Monn, Ch.; Carabias, V.; Junker, M.; Waeber, R.; Karrer, M.; Wanner, H. U. (1997) Small-scale spatial variability
of paniculate matter <10 |_im (PM10) and nitrogen dioxide. Atmos. Environ. 31: 2243-2247.
Morawska, L.; He, C.; Hitchins, J.; Gilbert, D.; Parappukkaran, S. (2001) The relationship between indoor and
outdoor airborne particles in the residential environment. Atmos. Environ. 35: 3463-3473.
Muraleedharan, T. R.; Radojevic, M. (2000) Personal particle exposure monitoring using nephelometry during haze
in Brunei. Atmos. Environ. 34: 2733-2738.
Murray, D. M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
parametric distributions by season and climatic region. Risk Anal. 15: 459-465.
National Air Pollution Control Administration. (1969) Air quality criteria for paniculate matter. Washington, DC:
U.S. Department of Health, Education, and Welfare, Public Health Service; NAPCA publication no. AP-49.
Available from: NTIS, Springfield, VA; PB-190251/BA.
National Research Council. (1991) Human exposure assessment for airborne pollutants: advances and opportunities.
Washington, DC: National Academy of Sciences.
National Research Council. (1998) Research priorities for airborne paniculate matter. I. Immediate priorities and a
long-range research portfolio. Washington, DC: National Academy Press.
National Research Council. (1999) Research priorities for airborne paniculate matter. II. Evaluating research
progress and updating the portfolio. Washington, DC: National Academy Press.
Naumova, Y. Y.; Eisenreich, S. J.; Turpin, B. J.; Weisel, C. P.; Morandi, M. T.; Colome, S. D.; Totten, L. A.; Stock,
T. H.; Winer, A. M.; Alimokhtari, S.; Kwon, J.; Shendell, D.; Jones, J.; Maberti, S.; Wall, S. J. (2002)
Polycyclic aromatic hydrocarbons in the indoor and outdoor air of three cities in the U.S. Environ. Sci.
Technol. 36: 2552-2559.
Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
Nelson, P. R.; Conrad, F. W.; Kelly, S. P.; Maiolo, K. C.; Richardson, J. D.; Ogden, M. W. (1998) Composition of
environmental tobacco smoke (ETS) from international cigarettes part II: nine country follow-up. Environ.
Int. 24:251-257.
Nieuwenhuijsen, M. J.; Noderer, K. S.; Schenker, M. B.; Vallyathan, V.; Olenchock, S. (1999) Personal exposure to
dust, endotoxin and crystalline silica in California agriculture. Ann. Occup. Hyg. 43: 35-42.
5-136
-------
Oglesby, L.; Ktinzli, N.; Roosli, M; Braun-Fahrlander, C.; Mathys, P.; Stern, W.; Jantunen, M.; Kousa, A. (2000)
Validity of ambient levels of fine particles as surrogate for personal exposure to outdoor air pollution—results
of the European EXPOLIS-EAS study (Swiss Center Basel). J. Air Waste Manage. Assoc. 50: 1251-1261.
Ostro, B. D.; Hurley, S.; Lipsett, M. J. (1999) Air pollution and daily mortality in the Coachella Valley, California:
a study of PM10 dominated by coarse particles. Environ. Res. 81: 231-238.
Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2000) Coarse and fine particles and daily mortality in the Coachella
Valley, California: a follow-up study. J. Exposure Anal. Environ. Epidemiol. 10: 412-419.
Ott, W. R. (1984) Exposure estimates based on computer generated activity patterns. J. Toxicol. Clin. Toxicol.
21: 97-128.
Ott, W. R. (1999) Mathematical models for predicting indoor air quality from smoking activity. Environ. Health
Perspect. 107(suppl. 2): 375-381.
Ott, W.; Thomas, J.; Mage, D.; Wallace, L. (1988) Validation of the simulation of human activity and pollutant
exposure (SHAPE) model using paired days from the Denver, CO, carbon monoxide field study. Atmos.
Environ. 22:2101-2113.
Ott, W. R.; Mage, D. T.; Thomas, J. (1992) Comparison of microenvironmental CO concentrations in two cities for
human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 2: 249-267.
Ott, W. R.; Klepeis, N. E.; Switzer, P. (1995) Modeling environmental tobacco smoke in the home using transfer
functions. Presented at: 88th annual meeting and exhibition of the Air & Waste Management Association;
June; San Antonio, TX; paper no. 95-WP84B.03. Research Triangle Park, NC: U.S. Environmental Protection
Agency, Atmospheric Research And Exposure Assessment Laboratory; report no. EPA/600/A-95/079.
Available from: NTIS, Springfield, VA; PB95-225488.
Ott, W.; Switzer, P.; Robinson, J. (1996) Particle concentrations inside a tavern before and after prohibition of
smoking: evaluating the performance of an indoor air quality model. J. Air Waste Manage. Assoc.
46: 1120-1134.
Ott, W.; Wallace, L.; Mage, D. (2000) Predicting paniculate (PM10) personal exposure distributions using a random
component superposition (RCS) statistical model. J. Air Waste Manage. Assoc. 50: 1390-1406.
Owen, M. K.; Ensor, D. S.; Sparks, L. E. (1992) Airborne particle sizes and sources found in indoor air.
Atmos. Environ. Part A 26: 2149-2162.
Ozkaynak, H.; Spengler, J. (1996) The role of outdoor paniculate matter in assessing total human exposure.
In: Wilson, R.; Spengler, J., eds. Particles in our air: concentrations and health effects. Cambridge, MA:
Harvard University Press.
Ozkaynak, H.; Thurston, G. D. (1987) Associations between 1980 U.S. mortality rates and alternative measures of
airborne particle concentration. Risk Anal. 7: 449-461.
Ozkaynak, H.; Ryan, P. B.; Spengler, J. D.; Laird, N. M. (1986) Bias due to misclassification of personal exposures
in epidemiologic studies of indoor and outdoor air pollution. In: Berglund, B.; Berglund, U.; Lindvall, T.;
Spengler, J.; Sundell, J., eds. Indoor air quality: papers from the third international conference on indoor air
quality and climate; August 1984; Stockholm, Sweden. Environ. Int. 12: 389-393.
Ozkaynak, H.; Spengler, J. D.; Ludwig, J. F.; Butler, D. A.; Clayton, C. A.; Pellizzari, E.; Wiener, R. W. (1990)
Personal exposure to paniculate matter: findings from the Particle Total Exposure Assessment Methodology
(PTEAM) prepilot study. In: Indoor air '90: precedings of the 5th international conference on indoor air
quality and climate, volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON,
Canada: International Conference on Indoor Air Quality and Climate, Inc.; pp. 571-576.
Ozkaynak, H.; Spengler, J. D.; Xue, J.; Koutrakis, P.; Pellizzari, E. D.; Wallace, L. (1993) Sources and factors
influencing personal and indoor exposures to particles, elements and nicotine: findings from the particle
TEAM pilot study. In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H.,
eds. Indoor air '93: proceedings of the 6th international conference on indoor air quality and climate, v. 3,
combustion products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93;
pp. 457-462.
Ozkaynak, H.; Xue, J.; Spengler, J.; Wallace, L.; Pellizzari, E.; Jenkins, P. (1996a) Personal exposures to airborne
particles and metals: results from the particle TEAM study in Riverside, California. J. Exposure Anal.
Environ. Epidemiol. 6: 57-78.
Ozkaynak, H.; Xue, J.; Weker, R.; Butler, D.; Koutrakis, P.; Spengler, J. (1996b) The particle TEAM (PTEAM)
study: analysis of the data: final report, volume III. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Atmospheric Research and Exposure Assessment Laboratory; report no.
EPA/600/R-95/098. Available from: NTIS, Springfield, VA; PB97-102495.
5-137
-------
Patterson, E.; Eatough, D. J. (2000) Indoor/outdoor relationships for ambient PM25 and associated pollutants:
epidemiological implications inLindon, Utah. J. Air Waste Manage. Assoc. 50: 103-110.
Pellizzari, E. D., Mason, R. E.; Clayton, C. A.; Thomas, K. W.; Cooper, S.; Piper. L.; Rodes, C.; Goldberg, M.;
Roberds, J.; Michael, L. (1998) Manganese exposure study (Toronto). Final report. Research Triangle Park,
NC: Research Triangle Institute, Analytical and Chemical Sciences; report no. RTI/6312/02-01 F.
Pellizzari, E. D.; Clayton, C. A.; Rodes, C. E.; Mason, R. E.; Piper, L. L.; Fort, B.; Pfeifer, G.; Lynam, D. (1999)
Paniculate matter and manganese exposures in Toronto, Canada. Atmos. Environ. 33: 721-734.
Pellizzari, E. D.; Clayton, C. A.; Rodes, C. E.; Mason, R. E.; Piper, L. L.; Fort, B.; Pfeifer, G.; Lynam, D. (2001)
Paniculate matter and manganese exposures in Indianapolis, Indiana. J. Exposure Anal. Environ. Epidemiol.
11:423-440.
Phillips, K.; Howard, D. A.; Browne, D.; Lewsley, J. M. (1994) Assessment of personal exposures to environmental
tobacco smoke in British nonsmokers. Environ. Int. 20: 693-712.
Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G. (1996) Assessment of air quality in Stockholm by personal
monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Scand. J.
Work Environ. Health 22(suppl. 1): 1-24.
Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G.; Huici, A. (1997a) Assessment of air quality in Barcelona by
personal monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke.
Environ. Int. 23: 173-196.
Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1997b) Assessment of air quality in Turin by personal
monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Environ. Int.
23: 851-871.
Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G. (1998a) Assessment of air quality in Paris by personal
monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Environ. Int.
24: 405-425.
Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1998b) Assessment of environmental tobacco smoke and
respirable suspended particle exposures for nonsmokers in Lisbon by personal monitoring. Environ. Int.
24: 301-324.
Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1999) Assessment of environmental tobacco smoke and
respirable suspended particle exposures for nonsmokers in Basel by personal monitoring. Atmos. Environ.
33: 1889-1904.
Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
(1995) Paniculate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J. Respir.
Crit. Care Med. 151: 669-674.
Pope, C. A., Ill; Hill, R. W.; Villegas, G. M. (1999) Paniculate air pollution and daily mortality on Utah's Wasatch
Front. Environ. Health Perspect: 107: 567-573.
Praml, G.; Schierl, R. (2000) Dust exposure in Munich public transportation: a comprehensive 4-year survey in
buses and trams. Int. Arch. Occup. Environ. Health 73: 209-214.
Quintana, P. J. E.; Samimi, B. S.; Kleinman, M. T.; Liu, L.-J.; Soto, K.; Warner, G. Y.; Bufalino, C.; Valencia, J.;
Francis, D.; Hovell, M. H.; Delfino, R. J. (2000) Evaluation of a real-time passive personal particle monitor in
fixed site residential indoor and ambient measurements. J. Exposure Anal. Environ. Epidemiol. 10: 437-445.
Quintana, P. J. E.; Valenzia, J. R.; Delfino, R. J.; Liu, L.-J. S. (2001) Monitoring of 1-min personal paniculate
matter exposures in relation to voice-recorded time-activity data. Environ. Res. 87: 199-213.
Ramachandran, G.; Vincent, J. H. (1999) A Bayesian approach to retrospective exposure assessment. Appl. Occup.
Environ. Hyg. 14: 547-557.
Rea, A. W.; Zufall, M. J.; Williams, R. W.; Sheldon, L. S.; Howard-Reed, C. (2001) The influence of human activity
patterns on personal PM exposure: a comparative analysis of filter-based and continuous particle
measurements. J. Air Waste Manage. Assoc. 51: 1271-1279.
Ren, P.; Jankun, T. M.; Leaderer, B. P. (1999) Comparisons of seasonal fungal prevalence in indoor and outdoor air
and in house dusts of dwellings in one northeast American county. J. Exposure Anal. Environ. Epidemiol.
9: 560-568.
Riediker, MM.; Keller, S.; Wuthrich, B.; Koller, T.; Monn, C. (2000) Personal pollen exposure compared to
stationary measurements. J. Invest. Allergol. Clin. Immunol. 10: 200-203.
Riley, W. J.; McKone, T. E.; Lai, A. C. K.; Nazaroff, W. W. (2002) Indoor paniculate matter of outdoor origin:
importance of size-dependent removal mechanisms. Environ. Sci. Technol. 36: 200-207.
5-138
-------
Robinson, J. P.; Wiley, J. A.; Piazza, T.; Garrett, K.; Cirksena, K. (1991) Activity patterns of California residents
and their implications for potential exposure to pollution (final report). Sacramento, CA: California Air
Resources Board report no. CARB-A6-177-33.
Robinson, J. P.; Switzer, P.; Ott, W. (1994) Microenvironmental factors related to Californians' potential exposure to
environmental tobacco smoke (ETS). Research Triangle Park, NC: U.S. Environmental Protection Agency,
Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/R-94/116. Available from:
NTIS, Springfield, VA; PB94-214814.
Rodes, C.; Sheldon, L.; Whitaker, D.; Clayton, A.; Fitzgerald, K.; Flanagan, J.; DiGenova, F.; Hering, S.; Frazier, C.
(1998) Measuring concentrations of selected air pollutants inside California vehicles [final report].
Sacramento, CA: California Environmental Protection Agency, Air Resources Board; contract no. 95-339.
Rodes, C. E.; Lawless, P. A.; Evans, G. F.; Sheldon, L. S.; Williams, R. W.; Vette, A. F.; Creason, J. P.; Walsh, D.
(2001) The relationships between personal PM exposures for elderly populations and indoor and outdoor
concentrations for three retirement center scenarios. J. Exposure Anal. Environ. Epidemiol. 11: 103-115.
Rojas-Bracho, L.; Suh, H. H.; Koutrakis, P. (2000) Relationships among personal, indoor, and outdoor fine and
coarse particle concentrations for individuals with COPD. J. Exposure Anal. Environ. Epidemiol.
10: 294-306.
Rojas-Bracho, L.; Suh, H. H.; Oyola, P.; Koutrakis, P. (2002) Measurements of children's exposures to particles and
nitrogen dioxide in Santiago, Chile. Sci. Total Environ. 287: 249-264.
Roorda-Knape, M. C.; Janssen, N. A. H.; De Hartog, J. J.; Van Vliet, P. H. N.; Harssema, H.; Brunekreef, B. (1998)
Air pollution from traffic in city districts near major motorways. Atmos. Environ. 32: 1921-1930.
Rotko, T.; Oglesby, L.; Ktinzli, N.; Jantunen, M. J. (2000a) Population sampling in European air pollution exposure
study, EXPOLIS: comparisons between the cities and representativeness of the samples. J. Exposure Anal.
Environ. Epidemiol. 10: 355-364.
Rotko, T.; Koistinen, K.; Hanninen, O.; Jantunen, M. (2000b) Sociodemographic descriptors of personal exposure to
fine particles (PM25) in EXPOLIS Helsinki. J. Exposure Anal. Environ. Epidemiol. 10: 385-393.
Ryan, P. B. (1991) An overview of human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 1: 453-474.
Samet, J. M.; Dominici, F.; Zeger, S. L.; Schwartz, J.; Dockery, D. W. (2000) National morbidity, mortality, and air
pollution study. Part I: methods and methodologic issues. Cambridge, MA: Health Effects Institute; research
report no. 94.
Santos-Burgoa, C.; Rojas-Bracho, L.; Rosas-Perez, I.; Ramirez-Sanchez, A.; Sanchez-Rico, G.; Mejia-Hernandez, S.
(1998) Modelaje de exposicion a particulas en poblacion general y riesgo de enfermedad respiratoria [Particle
exposure modeling in the general population and risk of respiratory disease]. Gac. Med. Mex. 134: 407-417.
Sarnat, J. A.; Koutrakis, P.; Suh, H. H. (2000) Assessing the relationship between personal paniculate and gaseous
exposures of senior citizens living in Baltimore, MD. J. Air Waste Manage. Assoc. 50: 1184-1198.
Sarnat, J. A.; Schwartz, J.; Catalano, P. J.; Suh, H. H. (2001) Gaseous pollutants in paniculate matter epidemiology:
confounders or surrogates? Environ. Health Perspect. 109: 1053-1061.
Sarnat, J. A.; Long, C. M.; Koutrakis, P.; Coull, B. A.; Schwartz, J.; Suh, H. H. (2002) Using sulfur as a tracer of
outdoor fine paniculate matter. Environ. Sci. Technol. 36: 5305-5314.
Scheff, P. A.; Wadden, R. (1979) Comparison of three methods of paniculate measurement in Chicago air. Atmos.
Environ. 13:639-643.
Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles? J. Air
Waste Manage. Assoc. 46: 927-939.
Sexton, K.; Ryan, P. B. (1988) Assessment of human exposure to air pollution: methods, measurements, and models.
In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile, and public health.
Washington, DC: National Academy Press; pp. 207-238.
Sexton, K.; Spengler, J. D.; Treitman, R. D. (1984) Personal exposure to respirable particles: a case study in
Waterbury, Vermont. Atmos. Environ. 18: 1385-1398.
Sheldon, L.; Clayton, A.; Jones, B.; Keever, J.; Perritt, R.; Smith, D.; Whitaker, D.; Whitmore, R. (1992a) Indoor
pollutant concentrations and exposures. Sacramento, CA: California Air Resources Board; report no.
ARB/R-92/486.
Sheldon, L.; Clayton, A.; Keever, J. Perritt, R.; Whitaker, D. (1992b) PTEAM: monitoring of phthalates and PAHs
in indoor and outdoor air samples in Riverside, California, volume II [final report]. Sacramento, CA:
California Environmental Protection Agency, State Air Resources Board, Research Division; contract no.
A933-144. Available from: NTIS, Springfield, VA; PB93-205649/XAB.
5-139
-------
Sheldon, L.; Whitaker, D.; Keever, I; Clayton, A.; Perritt, R. (1993a) Phthalates and PAHs in indoor and outdoor air
in a southern California community. In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.;
Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings of the 6th international conference on indoor air
quality and climate. Volume 3: combustion products, risk assessment, policies; July; Helsinki, Finland.
Helsinki, Finland: Indoor Air '93; pp. 109-114.
Sheldon, L.; Clayton, A.; Keever, I; Perritt, R.; Whitaker, D. (1993b) Indoor concentrations of polycyclic aromatic
hydrocarbons in California residences. Sacramento, CA: Air Resources Board; ARE contract no. A033-132.
Sheldon, L.; Clayton, A.; Perritt, R.; Whitaker, D. A.; Keever, J. (1993c) Indoor concentrations of polycyclic
aromatic hydrocarbons in California residences and their relationship to combustion source use. In: Jantunen,
M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings
of the 6th international conference on indoor air quality and climate. Volume 3: combustion products, risk
assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 29-34.
Shy, C. M.; Kleinbaum, D. G.; Morgenstern, H. (1978) The effect of misclassification of exposure status in
epidemiological studies of air pollution health effects. Bull. N. Y. Acad. Med. 54: 1155-1165.
Siegmann, K.; Schemer, L.; Siegmann, H. C. (1999) Physical and chemical properties of airborne nanoscale particles
and how to measure the impact on human health. J. Mol. Struct. (Theochem) 458: 191-201.
Sinclair, J. D.; Psota-Kelty, L. A.; Weschler, C. J. (1988) Indoor/outdoor ratios and indoor surface accumulations of
ionic substances at Newark, New Jersey. Atmos. Environ. 22: 461-469.
Sinclair, J. D.; Psota-Kelty, L. A.; Weschler, C. J.; Shields, H. C. (1990) Measurement and modeling of airborne
concentrations and indoor surface accumulation rates of ionic substances at Neenah, Wisconsin. Atmos.
Environ. Part A 24: 627-638.
Sioutas, C.; Wang, P. Y.; Ferguson, S. T.; Koutrakis, P.; Mulik, J. D. (1996) Laboratory and field evaluation of an
improved glass honeycomb denuder/filter pack sampler. Atmos. Environ. 30: 885-895.
Sioutas, C.; Kim, S.; Chang, M.; Terrell, L. L.; Gong, H., Jr. (2000) Field evaluation of a modified DataRAM MIE
scattering monitor for real-time PM2 5 mass concentration measurements. Atmos. Environ. 34: 4829-4838.
Serensen, M.; Daneshvar, B.; Hansen, M.; Dragsted, L. O.; Hertel, O.; Knudsen, L.; Loft, S. (2003) Personal PM25
exposure and markers of oxidative stress in blood. Environ. Health Perspect. Ill: 161-165.
Spengler, J. D.; Treitman, R. D.; Tosteson, T. D.; Mage, D. T.; Soczek, M. L. (1985) Personal exposures to
respirable particulates and implications for air pollution epidemiology. Environ. Sci. Technol. 19: 700-707.
Spengler, J. D.; Ozkaynak, H.; Ludwig, J.; Allen, G.; Pellizzari, E. D.; Wiener, R. (1989) Personal exposures to
paniculate matter: instruments and methodologies P-TEAM. In: Measurement of toxic and related air
pollutants: proceedings of the 1989 U.S. EPA/A&WMA international symposium; May; Raleigh, NC.
Pittsburgh, PA: Air & Waste Management Association; pp. 449-463. (A&WMA publication VTP-13).
Spier, C. E.; Little, D. E.; Trim, S. C.; Johnson, T. R.; Linn, W. S.; Hackney, J. D. (1992) Activity patterns in
elementary and high school students exposed to oxidant pollution. J. Exposure Anal. Environ. Epidemiol.
2: 277-293.
Suggs, J. C.; Burton, R. M. (1983) Spatial characteristics of inhalable particles in the Philadelphia metropolitan area.
J. AirPollut. Control Assoc. 33: 688-691.
Suh, H. H.; Spengler, J. D.; Koutrakis, P. (1992) Personal exposures to acid aerosols and ammonia. Environ. Sci.
Technol. 26:2507-2517.
Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993a) Indoor and outdoor acid aerosol and gas concentrations.
In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air '93:
proceedings of the 6th international conference on indoor air quality and climate. Volume 3: combustion
products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 257-262.
Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993b) Validation of personal exposure models for sulfate and aerosol
strong acidity. J. Air Waste Manage. Assoc. 43: 845-850.
Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1994) The relationship between airborne acidity and ammonia in indoor
environments. J. Exposure Anal. Environ. Epidemiol. 4: 1-23.
Suh, H. H.; Allen, G. A.; Koutrakis, P.; Burton, R. M. (1995) Spatial variation in acidic sulfate and ammonia
concentrations within metropolitan Philadelphia. J. Air Waste Manage. Assoc. 45: 442-452.
Tamura, K.; Ando, M.; Sagai, M.; Matsumoto, Y. (1996a) Estimation of levels of personal exposure to suspended
paniculate matter and nitrogen dioxide in Tokyo. Environ. Sci. (Tokyo) 4: 37-51.
Tamura, K.; Ando, M.; Miyazaki, T.; Funasaka, K.; Kuroda, K. (1996b) Indoor concentrations and personal
exposure to paniculate matter at roadside houses. In: Indoor air '96: the 7th international conference on indoor
air quality, volume I; July; Nagoya, Japan. Tokyo, Japan: Institute of Public Health; pp. 599-602.
5-140
-------
Teschke, K.; Marion, S. A.; Vaughan, T. L.; Morgan, M. S.; Camp, J. (1999) Exposures to wood dust in U.S.
industries and occupations, 1979 to 1997. Am. J. Ind. Med. 35: 581-589.
Thatcher, T. L.; Layton, D. W. (1995) Deposition, resuspension, and penetration of particles within a residence.
Atmos. Environ. 29: 1487-1497.
Thatcher, T. L.; Lai, A. C. K.; Moreno-Jackson, R.; Sextro, R. G.; Nazaroff, W. W. (2002) Effects of room
furnishings and air speed on particle deposition rates indoors. Atmos. Environ. 36: 1811-1819.
Thomas, K. W.; Pellizzari, E. D.; Clayton, C. A.; Whitaker, D. A.; Shores, R. C.; Spengler, J.; Ozkaynak, H.;
Froehlich, S. E.; Wallace, L. A. (1993) Particle total exposure assessment methodology (PTEAM) 1990
study: method performance and data quality for personal, indoor, and outdoor monitoring. J. Exposure Anal.
Environ. Epidemiol. 3: 203-226.
Tsang, A. M.; Klepeis, N. E. (1996) Descriptive statistics tables from a detailed analysis of the national human
activity pattern survey (NHAPS) data. Las Vegas, NV: U.S. Environmental Protection Agency, Office of
Research and Development; report no. EPA/600/R-96/148.
U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
U.S. Environmental Protection Agency. (1992) Guidelines for exposure assessment. Washington, DC: Risk
Assessment Forum, USEPA 600Z-92/001.
U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
U.S. Environmental Protection Agency. (1998) Particulate matter research needs for human health risk assessment to
support future reviews of the national ambient air quality standards for particulate matter. Research Triangle
Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
U.S. Environmental Protection Agency. (2000) Aerometric Information Retrieval System (AIRS). Research Triangle
Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/ [2000, October 6].
Vakeva, M.; Ha'meri, K.; Kulmala, M.; Lahdes, R.; Ruuskanen, J.; Laitinen, T. (1999) Street level versus rooftop
concentrations of submicron aerosol particles and gaseous pollutants in an urban street canyon.
Atmos. Environ. 33: 1385-1397.
Van Vliet, P.; Knape, M.; De Hartog, J.; Janssen, N.; Harssema, H.; Brunekreef, B. (1997) Motor vehicle exhaust
and chronic respiratory symptoms in children living near freeways. Environ. Res. 74: 122-132.
Vette, A. F.; Rea, A. W.; Lawless, P. A.; Rodes, C. E.; Evans, G.; Highsmith, V R.; Sheldon, L. (2001)
Characterization of indoor-outdoor aerosol concentration relationships during the Fresno PM exposure
studies. Aerosol Sci. Technol. 34: 118-126.
Waggoner, A. P.; Weiss, R. E. (1980) Comparison of fine particle mass concentration and light scattering extinction
in ambient aerosol. Atmos. Environ. 14: 623-626.
Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of submicron
particle exposure. Environ. Health Perspect. 108: 1139-1145.
Waldman, J. M.; Liang, C. S. K. (1993) Penetration and neutralization of acid aerosols at institutional settings.
In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air '93:
proceedings of the 6th international conference on indoor air quality and climate. Volume 3: combustion
products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 263-268.
Waldman, J. M.; Liang, C. S. K.; Menon, P. (1990) Indoor exposures to acidic aerosols at child and elderly care
facilities. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and climate,
volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
International Conference on Indoor Air Quality and Climate, Inc.; pp. 607-612.
Waldman, J. M.; Lioy, P. J.; Greenberg, A.; Butler, J. P. (1991) Analysis of human exposure to benzo(a)pyrene via
inhalation and food ingestion in the Total Human Environmental Exposure Study (THEES). J. Exposure Anal.
Environ. Epidemiol. 1: 193-225.
Wallace, L. (1996) Indoor particles: a review. J. Air Waste Manage. Assoc. 46: 98-126.
Wallace, L. (2000a) Correlations of personal exposure to particles with outdoor air measurements: a review of recent
studies. Aerosol Sci. Technol. 32: 15-25.
Wallace, L. (2000b) Real-time monitoring of particles, PAH, and CO in an occupied townhouse. Appl. Occup.
Environ. Hyg. 15: 39-47.
Wallace, L.; Howard-Reed, C. (2002) Continuous monitoring of ultrafine, fine, and coarse particles in a residence
for 18 months in 1999-2000. J. Air Waste Manage. Assoc. 52: 828-844.
5-141
-------
Wallace, L.; Quackenboss, I; Rodes, C. (1997) Continuous measurements of particles, PAH, and CO in an occupied
townhouse inReston, VA. In: Measurement of toxic and related air pollutants. Volume II: proceedings of a
specialty conference; April-May; Research Triangle Park, NC. Pittsburgh, PA: Air & Waste Management
Association, pp. 860-871. (A&WMA publication VIP-74).
Wallace, L. A.; Emmerich, S. I; Howard-Reed, C. (2002) Continuous measurements of air change rates in an
occupied house for 1 year: the effect of temperature, wind, fans, and windows. J. Exposure Anal. Environ.
Epidemiol. 12: 296-306.
Watson, J. G.; Cooper, J. A.; Huntzicker, J. J. (1984) The effective variance weighting for least squares calculations
applied to the mass balance receptor model. Atmos. Environ. 18: 1347-1355.
Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
(1990) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
Watson, J. G.; Chow, J. C.; Pace, T. G. (1991) Chemical mass balance. In: Hopke, P. K., ed. Receptor modeling for
air quality management. Amsterdam, The Netherlands: Elsevier Science Publishers B.V.; pp. 83-116.
(Vandeginste, B. G. M.; Kvalheim, O. M., eds. Data handling in science and technology, v. 7).
Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
Environ. 33:2301-2312.
Wheeler, A. J.; Williams, L; Beaumont, R. A.; Hamilton, R. S. (2000) Characterisation of paniculate matter sampled
during a study of children's personal exposure to airborne paniculate matter in a UK urban environment.
Environ. Monit. Assess. 65: 69-11.
Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
White, W. H. (1998) Statistical considerations in the interpretation of size-resolved paniculate mass data. J. Air
Waste Manage. Assoc. 48: 454-458.
Wiener, R. W. (1988) Measurement and evaluation of personal exposure to aerosols. In: Measurement of toxic and
related air pollutants: proceedings of the 1988 EPA/APCA international symposium; May; Research Triangle
Park, NC. Pittsburgh, PA: Air Pollution Control Association; pp. 84-88. (APCA publication VTP-10).
Wiener, R. W. (1989) Particle total exposure assessment methodology-an overview of planning and
accomplishments. In: Measurement of toxic and related air pollutants: proceedings of the 1989 U.S.
EPA/A&WMA international symposium; May; Raleigh, NC. Pittsburgh, PA: Air & Waste Management
Association; pp. 442-448. (A&WMA publication VIP-13).
Wiener, R. W.; Wallace, L.; Pahl, D.; Pellizzari, E.; Whittaker, D.; Spengler, J.; Ozkaynak, H. (1990) Review of the
particle TEAM 9 home field study. In: Measurement of toxic and related air pollutants: proceedings of the
1990 EPA/A&WMA international symposium, v. 2; May; Raleigh, NC. Pittsburgh, PA: Air & Waste
Management Association; pp. 452-460. (A&WMA publication VTP-17).
Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G. (1991a) Activity
patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
ARB/R93/487. Available from: NTIS, Springfield, VA.; PB94-108719.
Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991b) Study of children's activity
patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
Williams, R. W.; Watts, R. R.; Stevens, R. K.; Stone, C. L.; Lewtas, J. (1999) Evaluation of a personal air sampler
for twenty-four hour collection of fine particles and semivolatile organics. J. Exposure Anal. Environ.
Epidemiol. 9: 158-166.
Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
(2000a) The 1998 Baltimore Paniculate Matter Epidemiology-Exposure Study: part 1. Comparison of
ambient, residential outdoor, indoor and apartment paniculate matter monitoring. J. Exposure Anal. Environ.
Epidemiol. 10: 518-532.
Williams, R.; Suggs, J.; Creason, J.; Rodes, C.; Lawless, P.; Kwok, R.; Zweidinger, R.; Sheldon, L. (2000b) The
1998 Baltimore paniculate matter epidemiology-exposure study: part 2. Personal exposure assessment
associated with an elderly study population. J. Exposure Anal. Environ. Epidemiol.: 10: 533-543.
Williams, R.; Creason, J.; Zweidinger, R.; Watts, R.; Sheldon, L.; Shy, C. (2000c) Indoor, outdoor, and personal
exposure monitoring of paniculate air pollution: the Baltimore elderly epidemiology-exposure pilot study.
Atmos. Environ. 34: 4193-4204.
Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
Wilson, W. E.; Mage, D. T.; Grant, L. D. (2000) Estimating separately personal exposure to ambient and
nonambient paniculate matter for epidemiology and risk assessment: why and how. J. Air Waste Manage.
Assoc. 50: 1167-1183.
5-142
-------
Yakovleva, E.; Hopke, P. K.; Wallace, L. (1999) Receptor modeling assessment of particle total exposure
assessment methodology data. Environ. Sci. Technol. 33: 3645-3652.
Zartarian, V. G.; Ott, W. R.; Duan, N. (1997) A quantitative definition of exposure and related concepts. J. Exposure
Anal. Environ. Epidemiol. 7: 411-437.
Zeger, S. L.; Thomas, D.; Dominici, F.; Samet, J. M.; Schwartz, J.; Dockery, D.; Cohen, A. (2000) Exposure
measurement error in time-series studies of air pollution: concepts and consequences. Environ. Health
Perspect. 108: 419-426.
Zhu, Y.; Hinds, W. C.; Kim, S.; Shen, S.; Sioutas, C. (2002) Study of ultrafine particles near a major highway with
heavy-duty diesel traffic. Atmos. Environ. 36: 4323-4335.
Zmirou, D.; Masclet, P.; Boudet, C.; Dor, F.; Dechenaux, J. (2000) Personal exposure to atmospheric polycyclic
aromatic hydrocarbons in a general adult population and lung cancer risk assessment. J. Occup. Environ.
Med. 42: 121-126.
5-143
-------
United States
Environmental Protection
Agency
Please make all necessary changes in the below label,
detach copy or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HERE |_
detach copy or copy, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
National Center for
Environmental Assessment
Research Triangle Park, NC 27711
Official Business
Penalty for Private Use
$300
EPA/600/P-99/002aF
October 2004
-------
&EPA
United States
Environmental Protection
Agency
Please make all necessary changes in the below label,
detach copy or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HERE |_
detach copy or copy, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
National Center for
Environmental Assessment
Research Triangle Park, NC 27711
Official Business
Penalty for Private Use
$300
EPA/600/P-99/002aF
October 2004
------- |