United States
Environmental Protection
Agency
Office of Research
and Development
Washington, DC 20460
EPA 600/P-99/002aB
March 2001
Second External Review Draft
Air Quality Criteria for
Particulate Matter
Volume I
Notice
This document is a preliminary draft. It has not been formally
released by EPA and should not at this stage be construed to
represent Agency policy. It is being circulated for comment on
its technical accuracy and policy implications.
-------
EPA 600/P-99/002aB
March 2001
Second External Review Draft
Air Quality Criteria for
Participate Matter
Volume I
Notice
This document is a preliminary draft. It has not been formally
released by EPA and should not at this stage be construed to
represent Agency policy. It is being circulated for comment on its
technical accuracy and policy implications.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Disclaimer
This document is an external review draft for review purposes only and does not constitute
U.S. Environmental Protection Agency policy. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
March 2001 I-ii DRAFT-DO NOT QUOTE OR CITE
-------
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 and 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 //m) 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 Parti culate 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 //g/m3, 24-h; 50 //g/m3, annual average) were
retained in modified form and new standards (65 //g/m3, 24-h; 15 //g/m3, annual average) for
particles <2.5 //m (PM25) were promulgated in July 1997.
This Second External Review Draft of revised Air Quality Criteria for Parti culate Matter
assesses new scientific information that has become available mainly between early 1996 through
December 2000. The present draft is being 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 as a consequence of this assessment. Extensive additional pertinent
information is expected to be published during the next 6 to 9 mo (including results from a vastly
expanded EPA PM Research program and from other federal and state agencies, as well as other
March 2001 I-iii DRAFT-DO NOT QUOTE OR CITE
-------
partners in the general scientific community) and, as such, the findings and conclusions presented
in this draft document must be considered only provisional at this time. Public comments and
CASAC review recommendations will be taken into account, along with any pertinent newly
available information published or accepted for peer-reviewed publication by May/June 2001, in
making any appropriate further revisions to this document for incorporation into a Third External
Review Draft. That draft is expected to be released in September/October, 2001 for further
public comment and CASAC review (December 2001) in time for a final version to be
completed by early 2002. Evaluations contained in the present 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 options 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.
Preparation of this document was coordinated by staff of 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, and earlier drafts of this document were reviewed by experts from
federal and state government agencies, academia, industry, and NGO's for use by EPA in support
of decision making on potential public health and environmental risks of ambient PM. The
document describes the nature, sources, distribution, measurement, and concentrations of PM in
outdoor (ambient) and indoor 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 associated with alterations in
atmospheric transmission of solar radiation or its reflectance from the Earth's surface or
atmosphere (to support decision making on secondary PM NAAQS).
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.
March 2001 I-iv DRAFT-DO NOT QUOTE OR CITE
-------
Air Quality Criteria for Particulate Matter
VOLUME I
1. INTRODUCTION 1-1
2.
3.
PHYSICS, CHEMISTRY, AND MEASUREMENT OF
PARTICULATE MATTER
2-1
CONCENTRATIONS, SOURCES, AND EMISSIONS OF
ATMOSPHERIC PARTICULATE MATTER 3-1
Appendix 3A: Organic Composition of Particulate Matter 3A-1
Appendix 3B: Composition of Particulate Matter Source Emissions 3B-1
4. ENVIRONMENTAL EFFECTS OF PARTICULATE MATTER 4-1
Appendix 4A: Excerpted Key Points from the Executive Summary of
the World Meteorological Organization 1998 Assessment
of Stratospheric Ozone Depletion 4A-1
Appendix 4B: Excerpted Key Points from the Executive Summary of
the United Nations Environment Programme 1998
Assessment of Environmental Effects of Ozone
Depletion 4B-1
Appendix 4C: Excerpted Key Points from the Executive Summary of
the Special Report of the Intergovernmental Panel on
Climate Change Working Group II on the Regional
Impacts of Climate Change: An Assessment of
Vulnerability 4C-1
Appendix 4D: Excerpted Materials from the U.S. Global Change
Research Program Assessment Overview Report on
Climate Change Impacts on the United States and
Subsidiary Regional Assessment Reports 4D-1
Appendix 4E: Recent Model Proj ections of Excess Mortality Expected
in U.S. Cities During Summer and Winter Seasons
Because of Future Climate Change, Based on Kalkstein
and Greene (1997) 4E-1
5. HUMAN EXPOSURE TO PARTICULATE MATTER AND ITS
CONSTITUENTS
5-1
March 2001
I-v
DRAFT-DO NOT QUOTE OR CITE
-------
Air Quality Criteria for Particulate Matter
(cont'd)
VOLUME II
6. EPIDEMIOLOGY OF HUMAN HEALTH EFFECTS FROM
AMBIENT PARTICULATE MATTER ............................. 6-1
Appendix 6A: Demographic and Pollution Data for 90-City Analysis of
NMMAPS Project ................................ 6A-1
Appendix 6B: Heart Rate Variability as a Predictor of Serious
Cardiac Outcomes ................................. 6B-1
7. DOSIMETRY OF PARTICULATE MATTER ........................ 7-1
8. TOXICOLOGY OF PARTICULATE MATTER ...................... 8-1
9. INTEGRATIVE SYNTHESIS OF KEY POINTS: PARTICULATE
MATTER ATMOSPHERIC SCIENCE, AIR QUALITY, HUMAN
EXPOSURE, DOSIMETRY, AND HEALTH RISKS .................. 9-1
Appendix 9A: Key Quantitative Estimates of Relative Risk for
Particulate Matter-Related Health Effects Based on
Epidemiologic Studies of North American Cities
Assessed in the 1996 Particulate Matter Air Quality
Criteria Document ................................ 9A-1
EXECUTIVE SUMMARY ........................................... E-l
March 2001 I-vi DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
Page
List of Tables I-xiv
List of Figures I-xviii
Authors, Contributors, and Reviewers I-xxv
U.S. Environmental Protection Agency Project Team for Development of
Air Quality Criteria for Particulate Matter I-xxxv
1. INTRODUCTION 1-1
1.1 LEGISLATIVE REQUIREMENTS 1-1
1.2 HISTORY OF PREVIOUS PARTICULATE MATTER CRITERIA
AND NATIONAL AMBIENT AIR QUALITY STANDARDS REVIEWS 1-2
1.2.1 The 1997 Particulate Matter National Ambient Air Quality
Standards Revision 1-3
1.2.2 Presidential Memorandum: Next Particulate Matter Review
and Research 1-5
1.3 CURRENT PARTICULATE MATTER CRITERIA AND NATIONAL
AMBIENT AIR QUALITY STANDARDS REVIEW 1-6
1.3.1 Criteria Review 1-6
1.3.2 Methods and Procedures for Document Preparation 1-8
1.3.3 Approach 1-10
1.3.4 Key Issues of Concern 1-11
1.4 DOCUMENT CONTENT AND ORGANIZATION 1-13
REFERENCES 1-16
2. PHYSICS, CHEMISTRY, AND MEASUREMENT OF PARTICULATE MATTER ... 2-1
2.1 PHYSICS AND CHEMISTRY OF PARTICULATE MATTER 2-1
2.1.1 Definitions 2-1
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 2-3
2.1.2.3 Nuclei-Mode Particles 2-15
2.1.3 Chemistry of Atmospheric Particulate Matter 2-19
2.1.3.1 Chemical Composition and Its Dependence on Particle Size .... 2-19
2.1.3.2 Primary and Secondary Particulate Matter 2-20
2.1.3.3 Particle-Vapor Partitioning 2-27
2.13.4 Removal Processes 2-30
2.1.3.5 Particulate Matter and Welfare Effects 2-33
2.1.4 Summary 2-34
2.2 MEASUREMENT OF PARTICULATE MATTER 2-34
2.2.1 Problems in Measuring Particulate Matter 2-34
2.2.1.1 Treatment of Semivolatile Components of Particulate Matter ... 2-36
2.2.1.2 Upper CutPoint 2-38
March 2001 I-vii DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
2.2.1.3 Cut Point for Separation of Fine-Mode and Coarse-Mode
Particulate Matter 2-38
2.2.1.4 Treatment of Pressure, Temperature, and Relative Humidity .... 2-41
2.2.1.5 No Way to Determine Accuracy for Ambient Particulate
Matter Mass Measurement 2-42
2.2.2 Why Measure Particles 2-44
2.2.2.1 Attainment of a Standard 2-44
2.2.2.2 Implementation of a Standard 2-44
2.2.2.3 Determination of Health Effects 2-45
2.2.2.4 Determination of Ecological Effects 2-45
2.2.2.5 Determination of Radiative Effects 2-45
2.2.2.6 Particulate Matter Components/Parameters That Need
To Be Measured 2-45
2.2.3 Problems Associated with Semivolatile Particulate Matter 2-46
2.2.3.1 Particulate Nitrates 2-48
2.2.3.2 Semivolatile Organic Compounds 2-51
2.2.3.3 Use of Denuder Systems To Measure Semivolatile
Compounds 2-53
2.2.3.4 Particle-Bound Water 2-66
2.2.4 U. S. Environmental Protection Agency Monitoring Programs 2-71
2.2.4.1 The Federal Reference Methods for Equilibrated Mass 2-71
2.2.4.2 Speciation Monitoring 2-74
2.2.5 Continuous Monitoring 2-86
2.2.5.1 Tapered Element Oscillating Microbalance 2-86
2.2.5.2 Real-Time Total Ambient Mass Sampler 2-89
2.2.5.3 Continuous Ambient Mass Monitor 2-89
2.2.5.4 Light Scattering 2-91
2.2.5.5 Beta-Gauge Techniques 2-92
2.2.5.6 Measurements of Individual Particles 2-92
2.2.5.7 Automated Fine Particulate Nitrate 2-94
2.2.5.8 Semi-continuous Carbon Analysis 2-95
2.2.5.9 Determination of Aerosol Surface Area in Real Time 2-96
2.2.6 Data Quality 2-97
2.2.6.1 Errors in Gravimetric Analyses 2-97
2.2.6.2 Quality Assurance 2-98
2.3 SUMMARY 2-98
REFERENCES 2-106
3. CONCENTRATIONS, SOURCES, AND EMISSIONS OF ATMOSPHERIC
PARTICULATE MATTER 3-1
March 2001 I-viii DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
3.1 INTRODUCTION 3-1
3.2 TRENDS AND PATTERNS IN AMBIENT PM25 CONCENTRATIONS 3-2
3.2.1 Daily and Seasonal Variability 3-12
3.2.2 Diurnal (Circadian) Variability 3-19
3.2.3 Relations Among Particulate Matter in Different Size Fractions 3-24
3.2.4 Relations Between Mass and Chemical Component Concentrations 3-26
3.2.5 Spatial Variability 3-28
3.3 SOURCES OF PRIMARY AND SECONDARY PARTICULATE MATTER .... 3-34
3.3.1 Source Contributions to Ambient Particulate Matter 3-38
3.3.2 Long-Range Transport of Particulate Matter from Sources Outside
the United States 3-44
3.4 EMISSIONS ESTIMATES AND THEIR UNCERTAINTIES 3-49
3.4.1 Emissions Estimates for Primary Particulate Matter and Sulfur Dioxide,
Nitrogen Oxides, and Volatile Organic Compounds in the United States . . 3-49
3.4.2 Uncertainties of Emissions Inventories 3-54
3.5 SUMMARY AND CONCLUSIONS 3-57
REFERENCES 3-61
APPENDIX 3A: Organic Composition of Particulate Matter 3A-1
APPENDIX 3B: Composition of Particulate Matter Source Emissions 3B-1
4. ENVIRONMENTAL EFFECTS OF PARTICULATE MATTER 4-1
4.1 INTRODUCTION 4-1
4.2 EFFECTS ON VEGETATION AND ECOSYSTEMS 4-1
4.2.1 Direct Effects of Particulate Matter on Individual Plant Species 4-5
4.2.1.1 Effects of Coarse Particles 4-6
4.2.1.2 Effects of Fine Particles 4-9
4.2.2 Particulate Matter Effects on Natural Ecosystems 4-20
4.2.2.1 Introduction 4-20
4.2.2.2 Ecosystem Responses to Stress 4-22
4.2.2.3 Ecosystem Response to Direct Plant Effects 4-25
4.2.2.4 Indirect Effects of Particulate Matter in Ecosystems 4-33
4.2.3 Ecosystem Goods and Services and Their Economic Valuation 4-81
4.3 EFFECTS ON VISIBILITY 4-85
4.3.1 Introduction 4-85
4.3.2 Factors Affecting Atmospheric Visibility 4-85
4.3.2.1 Anthropogenic Pollutants 4-85
4.3.2.2 Human Vision 4-86
4.3.2.3 Characteristics of the Atmosphere 4-86
4.3.3 Optical Properties of Particles 4-88
March 2001 I-ix DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
4.3.4 Effect of Relative Humidity on Particle Size and Light-Scattering
Properties 4-91
4.3.5 Measures of Visibility 4-94
4.3.5.1 Human Observations 4-94
4.3.5.2 Light-Extinction Coefficient and Parameters Related to the
Light-Extinction Coefficient 4-94
4.3.5.3 Light-Scattering Coefficient 4-96
4.3.5.4 Fine Particulate Matter Concentrations 4-96
4.3.5.5 Discoloration 4-98
4.3.6 Visibility Monitoring Methods and Networks 4-99
4.3.7 Visibility Modeling 4-102
4.3.7.1 Regional Haze 4-102
4.3.7.2 Plume Models 4-106
4.3.7.3 Photographs 4-107
4.3.8 Trends in Visibility Impairment 4-108
4.3.9 Economics of Particulate Matter Visibility Effects 4-111
4.4 EFFECTS ON MATERIALS 4-114
4.4.1 Effects of Particles and Sulfur Dioxide on Man-Made Surfaces 4-115
4.4.1.1 Metals 4-115
4.4.1.2 Painted Finishes 4-116
4.4.1.3 Stone and Concrete 4-119
4.4.2 Soiling and Discoloration of Man-Made Surfaces 4-126
4.4.2.1 Stones and Concrete 4-126
4.4.2.2 Household and Industrial Paints 4-127
4.5 EFFECTS OF ATMOSPHERIC PARTICULATE MATTER ON CLIMATE
CHANGE PROCESSES AND THEIR POTENTIAL HUMAN HEALTH
AND ENVIRONMENTAL IMPACTS 4-128
4.5.1 Solar Ultraviolet Radiation Transmission Impacts on Human Health
and the Environment: Atmospheric Particulate Matter Effects 4-130
4.5.1.1 Bases for Concern Regarding Increased Ultraviolet
Radiation Transmission 4-130
4.5.1.2 Airborne Particle Impacts on Atmospheric Ultraviolet
Radiation Transmission 4-133
4.5.2 Global Warming Processes, Human Health and Environmental
Impacts, and Atmospheric Particle Roles 4-137
4.5.2.1 Bases for Concern Regarding Global Warming and Climate
Change 4-137
4.5.2.2 Airborne Particle Relationships to Global Warming and
Climate Change 4-150
4.6 SUMMARY 4-156
4.6.1 Particulate Matter Effects on Vegetation and Ecosystems 4-156
March 2001 I-x DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
4.6.2 Particulate Matter-Related Effects on Visibility 4-161
4.6.3 Particulate Matter-Related Effects on Materials 4-162
4.6.4 Effects of Particulate Matter on the Transmission of Solar Ultraviolet
Radiation and Global Warming Processes 4-163
REFERENCES 4-165
APPENDIX 4 A: Excerpted Key Points from the Executive Summary of the
World Meteorological Organization 1998 Assessment of
Stratospheric Ozone Depletion 4A-1
APPENDIX 4B: Excerpted Key Points from the Executive Summary of the
United Nations Environment Programme 1998 Assessment
of Environmental Effects of Ozone Depletion 4B-1
APPENDIX 4C: Excerpted Key Points from the Executive Summary of the
Special Report of the Intergovernmental Panel on Climate
Change Working Group n on the Regional Impacts of Climate
Change: An Assessment of Vulnerability 4C-1
APPENDIX 4D: Excerpted Materials from the U.S. Global Change Research
Program Assessment Overview Report on Climate Change
Impacts on the United States and Subsidiary Regional
Assessment Reports 4D-1
APPENDIX 4E:
Recent Model Projections of Excess Mortality Expected in
U.S. Cities During Summer and Winter Seasons Because of
Future Climate Change, Based on Kalkstein and Greene (1997)
4E-1
5. HUMAN EXPOSURE TO PARTICULATE MATTER AND ITS CONSTITUENTS .... 5-1
5.1 INTRODUCTION 5-1
5.1.1 Purpose 5-1
5.1.2 Particulate Matter Mass and Constituents 5-2
5.1.3 Relationship to Past Documents 5-3
5.2 STRUCTURE FOR THE CHAPTER 5-4
5.3 BASIC CONCEPTS OF EXPOSURE 5-4
5.3.1 Components of Exposure 5-4
5.3.2 Methods To Estimate Personal Exposure 5-5
5.3.2.1 Direct Measurement Methods 5-7
5 3.2.2 Indirect Methods (Modeling Methods) 5-8
5.3.2.3 Methods of Estimating Personal Exposure to Ambient
Particulate Matter 5-14
March 2001
I-xi
DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
5.4 SUMMARY OF PARTICULATE MATTER MASS DATA 5-19
5.4.1 Types of Particulate Matter Measurement Studies 5-19
5.4.2 Available Data 5-19
5.4.2.1 Personal Exposure Data 5-19
5.4.2.2 Microenvironmental Data 5-23
5.4.2.3 Interpretation of Parti culate Matter Exposure Data 5-24
5.4.3 Factors Influencing and Key Findings on Particulate Matter Exposures ... 5-30
5.4.3.1 Correlations of Personal/Microenvironmental Particulate
Matter with Ambient Particulate Matter 5-30
5.4.3.2 Factors That Affect Correlations 5-47
5.4.3.3 Impact of Ambient Sources on Exposures to Particulate Matter . . 5-64
5.4.3.4 Correlations of Parti culate Matter with Other Pollutants 5-67
5.5. SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA 5-70
5.5.1 Introduction 5-70
5.5.2 Monitoring Studies That Address Particulate Matter Constituents 5-70
5.5.3 Key Findings 5-73
5.5.3.1 Correlations of Personal and Indoor Concentrations with
Ambient Concentrations of Particulate Matter Constituents 5-73
5.5.4 Factors Affecting Correlations Between Ambient Measurements and
Personal or Microenvironmental Measurements of Particulate Matter
Constituents 5-79
5.5.5 Limitations of Available Data 5-80
5.6. IMPLICATIONS OF USING AMBIENT PARTICULATE MATTER
CONCENTRATIONS IN EPIDEMIOLOGIC STUDIES OF PARTICULATE
MATTERHEALTHEFFECTS 5-80
5.6.1 Potential Sources of Error Resulting from Using Ambient Particulate
Matter Concentrations in Epidemiologic Analyses 5-81
5.6.2 Associations Between Personal Exposures and Ambient Particulate
Matter Concentrations 5-83
5.6.3 Role of Compositional Differences in Exposure Characterization for
Epidemiology 5-85
5.6.4 Role of Spatial Variability in Exposure Characterization for
Epidemiology 5-87
5.6.5 Analysis of Exposure Measurement Error Issues in
Particulate Matter Epidemiology 5-88
5.6.5.1 Analysis of Exposure Measurement Errors in Time-Series
Studies 5-89
5.6.5.2 Analysis of Exposure Measurement Errors in
Long-Term Epidemiology Studies 5-92
5.6.5.3 Conclusions from Analysis of Exposure Measurement Errors
on Particulate Matter Epidemiology 5-93
March 2001 I-xii DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
5.7 SUMMARY OF KEY FINDINGS AND LIMITATIONS 5-95
REFERENCES 5-102
March 2001 I-xiii DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
Number Page
2-1 Comparison of Ambient Particles, Fine Mode and Coarse Mode 2-35
2-2 Particulate Matter Components/Parameters of Interest for Health, Ecological,
or Radiative Effects; for Source Category Apportionment Studies; or for Air
Quality Validation Studies 2-46
3-1 Gross Chemical Composition of PM25 Particles Obtained in Rural Areas
of the Eastern and Western United States by the IMPROVE Network and in
Mixed Rural, Suburban, and Urban Areas Obtained by Studies Summarized
in the 1996 Parti culate Matter Air Quality Criteria Document 3-3
3-2 Distribution of Ratios of PM25 to PM10 and Correlations Between PM25
and PM10, PM2 5, and PM(10_2 5), and PM(10_2 5) and PM10 Found at Collocated
Monitoring Sites in Seven Aerosol Characteristic (U.S. Environmental
Protection Agency/Health Effects Institute) Regions in 1999 3-25
3-3 Concentrations of PM2 5 and Selected Elements in the PM2 5 Size Range
and Correlations Between Elements and PM25 Mass 3-27
3-4 Pearson Correlation Coefficients for the Spatial Variation of PM25
Concentrations in Selected Metropolitan Statistical Areas 3-29
3-5 Correlation Coefficients for Spatial Variation of PM25 Mass and Different
Sources for Pairs of Sampling Sites in the South Coast Air Basin
(1986) 3-33
3-6 Correlation Coefficients for Spatial Variation of PM25 Mass and Different
Components for Pairs of Sampling Sites in Philadelphia (1994) 3-34
3-7 Constituents of Atmospheric Particles and Their Major Sources 3-35
3-8 Receptor Model Source Contributions to PM2 5 3-39
3-9 Receptor Model Source Contributions to PM10 3-40
3-10 Nationwide Changes in Ambient Concentrations and Emissions of PM10
and Gaseous Precursors to Secondary Particulate Matter
from 1989 to 1998 3-54
3A-1 Particulate Organic and Elemental Carbon Concentrations Based on
Studies Published After 1995 3A-2
March 2001 I-xiv DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number
3 A-2 Paniculate Organic Compound Concentrations Based on Studies Published
after 1990 at Selected Sites 3A-6
3B-1 Average Abundances of Major Elements in Soil and Crustal Rock 3B-2
3B-2 Summary of Particle-Phase Organic Compounds Present in Fine Particle
Road Dust Sample 3B-6
3B-3 Composition of Fine Particles Released by Various Stationary Sources in
the Philadelphia Area 3B-7
3B-4a Organic and Elemental Carbon Fractions of Diesel and Gasoline Engine
Paniculate Matter Exhaust 3B-13
3B-4b Contribution of Organic Carbon to Paniculate Matter Carbon Emissions
in Motor Vehicle Exhaust Collected from Vehicles Operated on Chassis
Dynamometers 3B-14
3B-5 Emission Rates for Constituents of Paniculate Matter from Gasoline and
Diesel Vehicles 3B-15
3B-6 Summary of Particle-Phase Organic Compounds Emitted from Motor
Vehicles 3B-17
3B-7 Mass Emissions, Organic Carbon, and Elemental Carbon Emissions from
Residential Combustion of Wood 3B-20
3B-8 Summary of Particle-Phase Organic Compounds Emitted from the
Combustion of Wood in Fireplaces 3B-21
3B-9 Mean Aerosol Composition at Tropical Site Affected Heavily by Biomass
Burning Emissions 3B-23
4-1 Relative Importance of Wet, Dry, Particulate, and Total Deposition to Three
Forest Sites 4-4
4-2 Ecosystem Services 4-20
4-3 Ecosystem Functions Impacted by Air Pollution Effects on Temperate Forest
Ecosystems 4-27
March 2001 I-xv DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number
4-4 Types of Plant Responses to Ultraviolet-B Radiation 4-40
4-5 Nitrogen-Saturated Forests in North America, Including Estimated Nitrogen
Inputs and Outputs 4-44
4-6 Primary Goods and Services Provided by Ecosystems 4-82
4-7 Averaged Regional PM2 5 Mass and Extinction Summaries for the Years
1988 to 1996 4-101
4-8 Corrosive Effects of Paniculate Matter and Sulfur Dioxide on Metals 4-117
4-9 Effects of Particulate Matter and Sulfur Dioxide on Stone 4-120
4D-1 Range of Projected Warming in the 21st Century 4D-5
4D-2 Types of Water Concerns Proj ected To Be Important for U. S. Regions
Consequent to Future Climate Change 4D-11
4D-3 Projected Future Climate-Change-Induced Impacts on Types of Ecosystems
of Concern to Different U.S. Regions 4D-12
4E-1 Modeled Projections of Direct Human Health Impacts of Climate Change:
Estimated Total Excess Mortality in U.S. Urban Areas For an Average
Summer Season, Assuming Full Acclimatization3 4E-2
4E-2 Modeled Projections of Direct Human Health Impacts of Global Climate
Change: Estimated Total Excess Mortality in U.S. Urban Areas for an
Average Winter Season, Assuming Full Acclimatizationa 4E-4
5-1 Classes of Parti culate Matter Exposure and Concentration Definitions 5-6
5-2 Activity Pattern Studies Included in the Consolidated Human Activity Database . . 5-11
5-3 Personal Exposure Models for Paniculate Matter 5-12
5-4 Summary of Recent Personal Exposure Studies 5-20
5-5 Summary of Recent Microenvironmental Measurement Studies 5-25
5-6 Papers Interpreting Particulate Matter Exposure Studies 5-29
March 2001 I-xvi DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number
5-7 Personal Monitoring Studies for Particulate Matter: Measured Concentrations
and Correlation Coefficients 5-33
5-8 Volume Mean Diameter and Maximum PM2 5 Concentrations of
Indoor Particle Sources 5-62
5-9 Correlations Between Personal PM25 and Ambient Pollutant Concentrations 5-68
5-10 Correlations Between Hourly Personal PM25 and Gaseous Pollutants 5-69
5-11 Studies That Have Measured Particulate Matter Constituents in
Personal Exposure Samples 5-71
5-12 Studies That Have Measured Particulate Matter Constituents in
Microenvironmental Samples 5-72
5-13 Summary Statistics for Personal, Indoor, and Outdoor Concentrations of
Selected Aerosol Components in Two Pennsylvania Communities 5-75
5-14 Statistical Correlation of Outdoor Versus Indoor Concentration for
Measured Species 5-77
March 2001 I-xvii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
Number Page
2-1 Number of particles as a function of particle diameter 2-5
2-2 Particle volume distribution as a function of particle diameter 2-6
2-3 Distribution of coarse, accumulation, and nuclei- or ultrafine-mode
particles by three characteristics: number, surface area, and volume
for the grand average continental size distribution 2-7
2-4 Volume size distribution, measured in traffic, showing fine-mode and
coarse-mode particles and the nuclei and accumulation modes within
the fine-particle mode 2-8
2-5 An idealized distribution of ambient particulate matter showing
fine-mode particles and coarse-mode particles and the fractions
collected by size-selective samplers 2-11
2-6 Specified particle penetration through an ideal inlet for five different
size-selective sampling criteria 2-12
2-7 Comparison of penetration curves for two PM10 beta gauge samplers using
cyclone inlets 2-15
2-8 Particle growth curves showing fully reversible hygroscopic growth of
sulfuric acid particles, deliquescent growth of ammonium sulfate particles
at about 80% relative humidity (RH), hygroscopic growth of ammonium
sulfate solution droplets at RH greater than 80%, and hysteresis (the
droplet remains supersaturated as the RH decreases below 80%) until the
crystallization point is reached 2-31
2-9 Theoretical predictions and experimental measurements of growth of
ammonium hydrogen sulfate particles at relative humidity between
95 and 100% 2-32
2-10 Schematic showing major nonvolatile and semivolatile components of PM2 5 .... 2-37
2-11 Particulate matter concentrations in Spokane, WA, during August 30, 1996,
dust storm 2-40
2-12 Amount of ammonium nitrate volatilized from Teflon filters, expressed as
a percentage of the measured PM2 5 mass, for the Southern California Air
Quality and CalTech studies, for spring and fall sampling periods 2-49
March 2001 I-xviii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
2-13 Aerosol water content expressed as a mass percentage, as a function of
relative humidity 2-69
2-14 This thermogram, for a sample containing rock dust (a carbonate source) and
diesel exhaust, shows three traces that correspond to temperature, filter
transmittance, and flame ionization detector response 2-82
2-15 Comparison of mass measurements with collocated real-time total ambient
mass sampler (real-time data), the Particle Concentration-Brigham Young
University Organic Sampling System (1-hour data), the Federal Reference
Method PM2 5 sampler (24-hour data), and a conventional tapered-element
oscillating microbalance monitor (real-time data) 2-90
2-16 Size distribution of particles divided by chemical classification into organic,
marine, and crustal 2-93
3-1 1999 annual mean PM10 concentrations 3-4
3-2 Nationwide trend in ambient PM10 concentration from 1989 through 1998 3-5
3-3 Trend in PM10 annual mean concentrations by U.S. Environmental Protection
Agency region, 1989 through 1998 3-6
3-4a 1999 annual mean PM25 concentrations 3-7
3-4b 1999 98th percentile 24-hour average PM2 5 concentrations 3-8
3-5 Annual distribution of 24-hour average PM25 concentrations observed in
U.S. and Canadian health studies 3-9
3-6 1999 annual mean PM(10.25) concentrations 3-12
3-7a Quarterly distributions of 24-hour average PM25 concentrations obtained in
eight eastern U.S. cities by the nationwide SLAMS/NAMS network of PM25
Federal Reference Method monitors during 1999 3-13
3-7b Quarterly distributions of 24-hour average PM25 concentrations obtained in six
central and mountain U.S. cities by the nationwide SLAMS/NAMS network
of PM25 Federal Reference Method monitors during 1999 3-14
March 2001 I-xix DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3-7c Quarterly distributions of 24-hour average PM25 concentrations obtained in
seven western U.S. cities and San Juan, PR, by the nationwide
SLAMS/NAMS network of PM2 5 Federal Reference Method monitors
during 1999 3-15
3-8 Concentrations of PM2 5 and PM10 measured in the four Metropolitan Acid
Aerosol Characterization Study cities 3-16
3-9 Frequency distribution of 24-hour average PM25 concentrations measured at
the PBY site in southwestern Philadelphia 3-17
3-10 Concentrations of 24-hour average PM2 5 measured at the U. S. Environmental
Protection Agency site in Phoenix, AZ 3-18
3-11 Frequency distribution of 24-hour average PM25 concentrations measured at
the U.S. Environmental Protection Agency site in Phoenix, AZ 3-18
3-12 Frequency distribution of 24-hour average PM25 measurements obtained
from all California Air Resources Board dichotomous sampler sites from
1989 to 1998 3-20
3-13 Frequency distribution of 24-hour average PM(10.25) concentrations obtained
from all California Air Resource Board Dichotomous sampler sites from
1989 to 1998 3-20
3-14 Concentrations of 24-hour average PM2 5 measured at the Riverside-Rubidoux
site 3-21
3-15 Frequency distribution of 24-hour average PM25 concentrations measured at
the Riverside-Rubidoux site 3-21
3-16 Intraday variability of hourly average PM25 concentrations across the
United States 3-23
3-17 PM2 5 chemical components in downtown Los Angeles and Burbank (1986)
have similar characteristics 3-31
3-18 Concentrations of PM2 5 chemical components in Rubidoux and downtown
Los Angeles (1986) 3-32
March 2001 I-xx DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3-19 Monthly average Saharan dust components of the aerosol sampled in Miami,
FL, during 1974 to 1996 3-46
3-20 PM2 5 and PM10 concentrations measured at Chilliwack Airport, located
in northwestern Washington State, just before and during the Asian desert
dust episode of April and May 1998 3-47
3-21 Time series of 24-hour average PM10 concentrations observed in the
Rio Grande Valley during May 1998 3-48
3-22 PM10 concentrations observed in St. Louis, MO, during May 1998 3-48
3-23 1998 directly emitted national particulate matter (PM25) emissions by
principal source categories for nonfugitive dust sources 3-50
3-24 Nationwide emissions of sulfur dioxide, nitrogen oxides, volatile organic
compounds, and ammonia from various source categories 3-52
3B-1 Size distribution of particles generated in a laboratory resuspension chamber .... 3B-3
3B-2 Size distribution of California source emissions, 1986 3B-4
3B-3 Chemical abundances for PM2 5 emissions from paved road dust in
Denver, CO 3B-5
3B-4 Chemical abundances for PM25 emissions from wood burning in Denver, CO . . . 3B-19
4-1 Effects of environmental stress on forest trees are presented on a hierarchial
scale for the leaf, branch, tree, and stand levels of organization 4-23
4-2 Nitrogen cycle 4-45
4-3 Diagrammatic overview of nitrogen excess in North America 4-48
4-4 Schematic of sources and sinks of hydrogen ions in a forest 4-56
4-5 Calcium deposition in >2-//m particles, <2-//m particles, and wet forms
and leaching in the Integrated Forest Study sites 4-64
4-6 Magnesium deposition in >2-//m particles, <2-//m particles, and wet forms
and leaching in the Integrated Forest Study sites 4-65
March 2001 I-xxi DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
4-7 Potassium deposition in >2-//m particles, <2-//m particles, and wet forms
and leaching in the Integrated Forest Study sites 4-66
4-8 Base cation deposition in >2-//m particles, <2-//m particles, and wet forms
and leaching in the Integrated Forest Study sites 4-67
4-9 Total cation leaching balanced by sulfate and nitrate, estimated from
particulate deposition and by other sources of sulfate and nitrate and by
other anions in the Integrated Forest Study sites 4-68
4-10 Soil exchangeable Ca2+ pools and net annual export of Ca2+ in the Integrated
Forest Study sites 4-69
4-11 Soil exchangeable Mg2+ pools and net annual export of Mg2+ in the Integrated
Forest Study sites 4-70
4-12 Soil exchangeable K2+ pools and net annual export of K2+ in the Integrated
Forest Study sites 4-71
4-13 Simulated soil solution mineral acid anions and base cations in the red spruce
site with no change, 50% N and S deposition, and 50% base cation deposition . . . 4-74
4-14 Simulated soil solution Al and soil base saturation in the red spruce site
with no change, 50% N and S deposition, and 50% base cation deposition 4-75
4-15 Simulated soil solution mineral acid anions and base cations in the Coweeta
site with no change, 50% N and S deposition, and 50% base cation deposition . . . 4-76
4-16 Relationship of plant nutrients and trace metals with vegetation 4-80
4-17 Light reflected from a target toward an observer 4-87
4-18 Humidity effect on scattering coefficients computed for internal and external
mixtures of the mixed-salt aerosol: Na2SO4(x2 = 0.5)-(NH4)2 SO4 (x3 = 0.5),
for two dry-salt particle size distributions, where x is the mass fraction of
the dry solutes 4-92
4-19 Scattering ratios for different chemical compositions as a function
of relative humidity 4-93
4-20 Comparison of extinction and visual range 4-95
March 2001 I-xxii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
4-21a,b Plots of the 10th, 50th, and 90th percentile groups for PM2 5 and deciview
at the Badlands National Park 4-97
4-22 Reduction in visual range as a function of increasing fine and coarse
particle concentrations 4-98
4-23 Five-season average haze patterns for the periods 1980 to 1985
and 1990 to 1995 4-110
4-24 Secular haze trends for 1940 to 1990 for six regions, summer and winter
in the 75th percentile 4-112
4-25 Processes involved in stratospheric ozone depletion because of man's
production of chlorofluorcarbons, halons, and other trace gases are shown 4-131
4-26 Bases for concern about global warming and climate change effects on the
environment and human health 4-142
4-27 Estimated global mean radiative forcing exerted by gas and various particle
phase species from 1850 to 1950 4-154
4D-1 Records of Northern Hemisphere surface temperatures, carbon dioxide
concentrations, and carbon emissions show a close correlation 4D-2
4D-2 Simulation of decadal average changes in temperature from leading
climate models on historic and projected changes in carbon dioxide
and sulfate atmospheric concentrations 4D-4
4D-3 Historic and projected changes in sea level based on the Canadian
and Hadley model simulations 4D-6
4D-4 This map is a preliminary classification of annual shoreline erosion
throughout the United States, in coarse detail and resolution 4D-6
4D-5 Breakout of United States regions as evaluated by the U.S. Global Change
Research Program, based on the overview assessment depiction 4D-9
4D-6 Projected climate-change impacts in the Mid-Atlantic Region of the
United States 4D-13
March 2001 I-xxiii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
5-1 Comparison of correlation coefficients for longitudinal analysis of
personal exposure versus ambient concentrations for individual subjects
for PM2 5 and sulfate 5-42
5-2 Personal exposure versus ambient concentrations for PM2 5 and sulfate 5-44
5-3 Regression analyses of aspects of daytime personal exposure to PM10 estimated
using data from the Particle Total Exposure Assessment Methodology study 5-46
5-4 Air-exchange rates measured in homes throughout the United States 5-53
5-5 Box plots of hourly air-exchange rates stratified by season in Boston, MA,
during 1998 5-54
5-6 Geometric mean infiltration factor (indoor/outdoor ratio) for hourly nighttime,
nonsource data for two seasons 5-55
5-7 Comparison of deposition rates from this study with literature values 5-57
5-8 Penetration efficiencies and deposition rates from models of nightly
average data 5-58
5-9 Mean hourly indoor/outdoor particle concentration ratio from an unoccupied
residence in Fresno, CA, during spring 1999 5-63
5-10 Personal versus outdoor SO4= in State College, PA 5-76
5-11 Plots of nonambient exposure to PM10, daytime individual values from the
Particle Total Exposure Assessment Methodology study data and daily-average
values from THEES data 5-86
March 2001 I-xxiv DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
CHAPTER 1. INTRODUCTION
Principal Authors
Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
CHAPTER 2. PHYSICS, CHEMISTRY, AND MEASUREMENT
OF PARTICULA TE MA TTER
Principal Authors
Dr. William Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Candis S. Claiborn—Washington State University, Laboratory for Atmospheric Research,
Department of Civil and Environmental Engineering, P.O. Box 642910, Pullman, WA 99164
Contributing Authors
Dr. Barbara J. Turpin, The State University of New Jersey, Rutgers, Department of
Environmental Sciences and Rutgers Cooperative Extension, New Brunswick, NJ 08901-8551
Dr. James J. Schauer, University of Wisconsin, College of Engineering, Department of Civil and
Environmental Engineering, Madison, WI 53706
Contributors and Reviewers
Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street,, Baltimore, MD 21205
Ms. Lee Byrd—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C,
Petaluma, CA 94549
March 2001 I-xxv DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, CA 92612
Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT 84602
Dr. Edward O. Edney—National Exposure Research Laboratory (MD-84)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
Engineering, 5531 Boelter Hall, Los Angeles, CA 90095
Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lynn Hildemann—Stanford University, Civil and Environmental Engineering Department,
Stanford, CA 94305
Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
One Brookings Drive, St. Louis, MO 63130
Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Scott Mathias—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom McCurdy—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001 I-xxvi DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street,
Waterbury, VT 05671
Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
Boston, MA 02461
Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Russell Weiner—National Exposure Research Laboratory (MD-46)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street,
Sacramento, CA 95814
CHAPTER 3. CONCENTRATIONS, SOURCES, AND
EMISSIONS OF A TMOSPHERIC PARTICULA TE MA TTER
Principal Author
Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001 I-xxvii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributing Authors
Dr. Barbara J. Turpin, The State University of New Jersey, Rutgers, Department of
Environmental Sciences and Rutgers Cooperative Extension, New Brunswick, NJ 08901-8551
Dr. James J. Schauer, University of Wisconsin, College of Engineering, Department of Civil and
Environmental Engineering, Madison, WI 53706
Contributors and Reviewers
Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street, Baltimore, MD 21205
Ms. Lee Byrd—Office of Air Quality Planning and Standards (MD-14),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C,
Petaluma, CA 94549
Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, C A 92612
Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT 84602
Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
Engineering, 5531 Boelter Hall, Los Angeles, CA 90095
Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lynn Hildemann—Civil and Environmental Engineering Department,
Stanford University, Stanford, CA 94305
Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
One Brookings Drive, St. Louis, MO 63130
March 2001 I-xxviii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Scott Mathias—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom McCurdy—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street,
Waterbury, VT 05671
Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
Boston, MA 02461
Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street, Sacramento, CA 95814
Dr. William Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001 I-xxix DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
CHAPTER 4. ENVIRONMENTAL EFFECTS OF PARTICULA TE MA TTER
Principal Authors
Ms. Beverly Comfort—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. J.H.B. Garner—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David A. Grantz—University of California/Riverside, Kearney Agricultural Center,
9240 South Riverbend Avenue, Parlier, CA 93648
Dr. Paul J. Hanson—Environmental Sciences Division, Oak Ridge National Laboratory,
P.O. Box 2008, Bethel Valley Road, Building 1059, Oak Ridge, TN 37831-6422
Dr. Dale W. Johnson—Environmental and Resource Science, 1000 Valley Road, University of
Nevada, Reno, NV 89512
Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William H. Smith—Yale University School of Forestry and Environmental Studies,
370 Prospect Street, New Haven, CT 06511
Ms. Debra Meyer Wefering—Duckterather Weg 61, Bergisch Gladbach, Germany 54169
(formerly with National Exposure Research Laboratory [MD-56], U.S. Environmental Protection
Agency, Research Triangle Park, NC 27711)
Contributors and Reviewers
Dr. Larry T. Cupitt—National Exposure Research Laboratory (MD-75),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001 I-xxx DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Russell R. Dickerson—University of Maryland, Department of Meteorology,
Stadium Drive, College Park, MD 20742
Dr. Anne Grambsch—National Center for Environmental Assessment (860ID),
U. S. Environmental Protection Agency, Washington, DC 20036
Dr. Sagar V. Krupa—University of Minnesota, Department of Plant Pathology,
495 BorlaugHall, 1991 Upper Buford Circle, St. Paul, MN 55108
Dr. Alan J. Krupnick—Quality of the Environment Division, Resources for the Future,
1616 P Street, NW, Washington, DC 20036
Mr. Paul T. Roberts—Sonoma Technology, Inc.,1360 Redwood Way - Suite C,
Petaluma, CA 94954
Mr. John Spence—1206 Sturdivant Drive, Cary, NC 27511
Dr. Richard Zepp—National Exposure Research Laboratory (IOD),
U. S. Environmental Protection Agency, Athens, GA
CHAPTER 5. HUMAN EXPOSURE TO PARTICULA TE MA TIER
AND ITS CONSTITUENTS
Principal Authors
Dr. David T. Mage—Institute for Survey Research, Temple University,
Philadelphia, PA 19122-6099 (formerly with the National Exposure Research
Laboratory (MD-56), U.S. Environmental Protection Agency, Research Triangle
Park,NC 27711)
Mr. Thomas McCurdy—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Linda S. Sheldon—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001 I-xxxi DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Principal Authors
(cont'd)
Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William E. Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributing Authors
Dr. Janet Burke—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Roy Fortmann—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Gary Norris—National Exposure Research Laboratory (MD-47),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Anne Rea—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Alan Vette—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributors and Reviewers
Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street,, Baltimore, MD 21205
Ms. Lee Byrd—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C, Petaluma, CA 94549
Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, C A 92612
Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT 84602
March 2001 I-xxxii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
Engineering, 5531 Boelter Hall, Los Angeles, CA 90095
Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Vic Hasselblad—29 Autumn Woods Drive, Durham, NC 27713
Dr. Lynn Hildemann—Civil and Environmental Engineering Department
Stanford University, Stanford, CA 94305
Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
One Brookings Drive, St. Louis, MO 63130
Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Allan Marcus—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001 I-xxxiii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street,
Waterbury, VT 05671
Mr. Harvey Richmond—Office of Air Quality Planning and Standards (MD-15),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
Boston, MA 02461
Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street, Sacramento, CA 95814
Dr. Jim Xue—Harvard School of Public Health, 665 Huntington Avenue, Boston, MA 02115
March 2001 I-xxxiv DRAFT-DO NOT QUOTE OR CITE
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
FOR PARTICULATE MATTER
Scientific Staff
Dr. Lester D. Grant—Director, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William E. Wilson—Air Quality Coordinator, Physical Scientist, National Center for
Environmental Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle
Park,NC 27711
Dr. Lawrence J. Folinsbee—Health Coordinator, Chief, Environmental Media Assessment
Group, National Center for Environmental Assessment (MD-52), U.S. Environmental Protection
Agency, Research Triangle Park, NC 27711
Dr. J.H.B. Garner—Ecological Scientist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis J. Kotchmar—Project Manager, Medical Officer, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Dr. Robert Chapman—Medical Officer, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly Comfort—Health Scientist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—Health Scientist, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David Mage—Physical Scientist, Institute for Survey Research, Temple University,
Philadelphia, PA 19122-6099
Dr. Allan Marcus—Statistician, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. James McGrath—Visiting Senior Health Scientist, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
March 2001 I-xxxv DRAFT-DO NOT QUOTE OR CITE
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
FOR PARTICULATE MATTER
(cont'd)
Scientific Staff
(cont'd)
Dr. Joseph P. Pinto—Physical Scientist, National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. James A. Raub—Health Scientist, National Center for Environmental Assessment (MD-52),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Technical Support Staff
Mr. Randy Brady—Deputy Directory, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Douglas B. Fennell—Technical Information Specialist, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Ms. Emily R. Lee—Management Analyst, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Diane H. Ray—Program Specialist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Eleanor Speh—Office Manager, Environmental Media Assessment Branch, National Center
for Environmental Assessment (MD-52), U.S. Environmental Protection Agency, Research
Triangle Park, NC 27711
Ms. Donna Wicker—Administrative Officer, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Richard Wilson—Clerk, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001 I-xxxvi DRAFT-DO NOT QUOTE OR CITE
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
FOR PARTICULATE MATTER
(cont'd)
Document Production Staff
Mr. John R. Barton—Document Processing Coordinator, OAO Corporation,
2222 Chapel Hill-Nelson Highway, Beta Building, Suite 100, Durham, NC 27713
Ms. Diane G. Caudill—Graphic Artist, OAO Corporation, 2222 Chapel Hill-Nelson Highway,
Beta Building, Suite 100, Durham, NC 27713
Ms. Yvonne A. Harrison—Word Processor, OAO Corporation, 2222 Chapel Hill-Nelson
Highway, Beta Building, Suite 100, Durham, NC 27713
Ms. Bettye B. Kirkland—Word Processor, OAO Corporation, 2222 Chapel Hill-Nelson
Highway, Beta Building, Suite 100, Durham, NC 27713
Mr. David E. Leonhard—Graphic Artist, OAO Corporation, 2222 Chapel Hill-Nelson Highway,
Beta Building, Suite 100, Durham, NC 27713
Ms. Phyllis H. Noell—Technical Editor, OAO Corporation, 2222 Chapel Hill-Nelson Highway,
Beta Building, Suite 100, Durham, NC 27713
Ms. Carolyn T. Perry—Word Processor, OAO Corporation, 2222 Chapel Hill-Nelson Highway,
Beta Building, Suite 100, Durham, NC 27713
Technical Reference Staff
Mr. R. David Belton—Reference Specialist, OAO Corporation, 2222 Chapel Hill-Nelson
Highway, Beta Building, Suite 100, Durham, NC 27713
Mr. John A. Bennett—Technical Information Specialist, OAO Corporation, 2222 Chapel
Hill-Nelson Highway, Beta Building, Suite 100, Durham, NC 27713
Ms. Sandra L. Hughey—Technical Information Specialist, OAO Corporation, 2222 Chapel
Hill-Nelson Highway, Beta Building, Suite 100, Durham, NC 27713
Mr. Jian Ping Yu—Reference Retrieval and Database Entry Clerk, OAO Corporation,
2222 Chapel Hill-Nelson Highway, Beta Building, Suite 100, Durham, NC 27713
Ms. Kun Zhang—Records Management Technician, OAO Corporation, 2222 Chapel Hill-Nelson
Highway, Beta Building, Suite 100, Durham, NC 27713
March 2001 I-xxxvii DRAFT-DO NOT QUOTE OR CITE
-------
i 1. INTRODUCTION
2
3
4 This document is an update of "Air Quality Criteria for Particulate Matter" published by the
5 U.S. Environmental Protection Agency (EPA) in 1996, and it will serve as the basis for
6 Congress!onally-mandated periodic review of the National Ambient Air Quality Standards
7 (NAAQS) for particulate matter (PM). The present document critically assesses the latest
8 scientific information relative to determining the health and welfare effects associated with
9 exposure to various concentrations of PM in ambient air. The document is not intended as a
10 complete and detailed literature review, but rather focuses on assessment and integration of
11 information most relevant to PM NAAQS criteria development, based on pertinent literature
12 mainly available through December 2000. This introductory chapter presents a brief summary of
13 the history of the PM NAAQS, provides an overview of issues addressed and procedures utilized
14 in the preparation of the present document, and provides orientation to the general organizational
15 structure of thi s document.
16
17
18 1.1 LEGISLATIVE REQUIREMENTS
19 Sections 108 and 109 of the U.S. Clean Air Act (CAA) (U.S. Code, 1991) govern the
20 establishment, review, and revision of National Ambient Air Quality Standards (NAAQS).
21 Section 108 directs the EPA Administrator to list pollutants that may reasonably be anticipated to
22 endanger public health or welfare and to issue air quality criteria for them. The air quality
23 criteria are to reflect the latest scientific information useful in indicating the kind and extent of all
24 exposure-related effects on public health and welfare that may be expected from the presence of
25 the pollutant in ambient air.
26 Section 109(a,b) directs the Administrator of EPA to propose and promulgate "primary"
27 and "secondary" NAAQS for pollutants identified under Section 108. Section 109(b)(l) defines
28 a primary standard as a level of air quality, the attainment and maintenance of which, in the
29 judgment of the Administrator, based on the criteria and allowing for an adequate margin of
30 safety, is requisite to protect the public health. Under Section 109(b) of the CAA, the EPA
March 2001 1-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 Administrator must consider available information to set secondary NAAQS that are based on
2 the criteria and are requisite to protect public welfare from any known or anticipated adverse
3 effects associated with the presence of such pollutants. Welfare effects include effects on
4 vegetation, crops, soils, water, animals, manufactured materials, visibility, weather, and climate,
5 as well as damage to and deterioration of property, hazards to transportation, and effects on
6 economic value and personal comfort and well-being. Section 109(d) also requires periodic
7 review and, if appropriate, revision of existing criteria and standards, and it requires an
8 independent committee of non-EPA experts, the Clean Air Scientific Advisory Committee
9 (CASAC), to provide the EPA Administrator advice and recommendations regarding the
10 scientific soundness and appropriateness of criteria and NAAQS for PM and other "criteria air
11 pollutants" (e.g., ozone, nitrogen oxides, sulfate dioxide, carbon monoxide, lead) regulated under
12 CAA Sections 108-109.
13
14
15 1.2 HISTORY OF PREVIOUS PARTICULATE MATTER CRITERIA AND
16 NATIONAL AMBIENT AIR QUALITY STANDARDS REVIEWS
17 On April 30, 1971 (Federal Register, 1971), EPA promulgated the original primary and
18 secondary NAAQS for PM under Section 109 of the CAA. The reference method for measuring
19 attainment of these standards was the "high-volume" sampler (Code of Federal Regulations,
20 1977), which collects ambient PM up to a nominal size of 25 to 45 micrometers (//m) (i.e.,
21 so-called "total suspended paniculate" or "TSP"). Thus, TSP was the original indicator for the
22 PM NAAQS. The primary standards for PM (measured as TSP) were 260 Mg/m3 (24-h average),
23 not to be exceeded more than once per year, and 75 //g/m3 (annual geometric mean). The
24 secondary standard (measured as TSP) was 150 //g/m3 (24-h average), not to be exceeded more
25 than once per year. The next review of PM air quality criteria and standards was completed in
26 July 1987, when the original TSP NAAQS set in 1971 were revised to protect against adverse
27 health effects of inhalable airborne particles with an upper 50% cut-point of 10-//m aerodynamic
28 diameter (PM10), which can be deposited in the lower (thoracic) regions of the human respiratory
29 tract (Federal Register, 1987). Identical primary and secondary PM10 standards were set for two
30 averaging times: 150 //g/m3 (24-h average), with no more than one expected exceedance per
31 year; and 50 //g/m3 (expected annual arithmetic mean), averaged over 3 years.
March 2001 1-2 DRAFT-DO NOT QUOTE OR CITE
-------
l 1.2.1 The 1997 Particulate Matter National Ambient Air Quality Standards
2 Revision
3 The EPA initiated the last previous review of the air quality criteria and standards for PM
4 in April 1994 by announcing its intention to develop revised Air Quality Criteria for Particulate
5 Matter (henceforth, the "PM Air Quality Criteria Document" or PM AQCD). Thereafter, the
6 EPA presented its plans for review of the criteria and standards for PM under a highly
7 accelerated, court-ordered schedule at a public meeting of the CASAC in December 1994.
8 A court order entered in American Lung Association v. Browner, CIV-93-643-TUC-ACM (U.S.
9 District Court of Arizona, 1995), as subsequently modified, required publication of EPA's final
10 decision on the review of the PM NAAQS by July 19, 1997.
11 Several workshops were held by EPA's National Center for Environmental Assessment
12 RTF Division (NCEA-RTP) in November 1994 and January 1995 to discuss important new
13 health effects information useful in preparing initial PM AQCD draft materials. External review
14 drafts of the PM AQCD then were made available for public comment and were reviewed by
15 CASAC at public meetings held in August 1995, December 1995, and February 1996. The
16 CASAC came to closure in its review of the PM AQCD, advising the EPA Administrator in a
17 March 15, 1996, closure letter (Wolff, 1996) that "although our understanding of the health
18 effects of PM is far from complete, a revised Criteria Document which incorporates the Panel's
19 latest comments will provide an adequate review of the available scientific data and relevant
20 studies of PM." Revisions in response to public and CASAC comments were incorporated as
21 appropriate in the final 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a). A PM
22 Staff Paper (SP), prepared by EPA's Office of Air Quality Planning and Standards (O AQPS) and
23 drawing on the 1996 PM AQCD and other exposure and risk assessments to pose options for PM
24 NAAQS decisions, also underwent similar CASAC review and public comment, with consequent
25 revision to its July 1996 final form (U.S. Environmental Protection Agency, 1996b).
26 The SP analyses served as key inputs to subsequently published proposals for revision of
27 the primary PM NAAQS. Taking into account information and assessments presented in the PM
28 AQCD and the SP, advice and recommendations of CASAC, and public comments received on
29 the proposal, the EPA Administrator revised the PM NAAQS by adding new PM2 5 standards and
30 by revising the form of the 24-h PM10 standard. Specifically, in July 1997, the Administrator
31 made the following revisions to the PM NAAQS:
March 2001 1-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 (1) The suite of PM standards was revised to include an annual primary PM2 5 standard and a
2 24-h PM2 5 standard.
3 (2) The 24-h PM25 standard is met when the 3-year average of the 98th percentile of 24-h PM2 5
4 concentrations at each population-oriented monitor within an area is less than or equal to
5 65 //g/m3, with fractional parts of 0.5 or greater rounding up.
6 (3) The annual PM2 5 standard is met when the 3-year average of the annual arithmetic mean
7 PM25 concentrations, from single or multiple community-oriented monitors is less than or
8 equal to 15 //g/m3, with fractional parts of 0.05 or greater rounding up.
9 (4) The form of the 24-h PM10 (150 //g/m3) standard was revised to be based on the 3-year
10 average of the 99th percentile of 24-h PM10 concentrations at each monitor within an area.
11 (5) In addition, the Administrator retained the annual PM10 standard at the level of 50 //g/m3,
12 which is met when the 3-year average of the annual arithmetic mean PM10 concentrations at
13 each monitor within an area is less than or equal to 50 //g/m3, with fractional parts of 0.5 or
14 greater rounding up.
15 The principal focus of the last review of the air quality criteria and standards for PM was on
16 recent epidemiological evidence reporting associations between ambient concentrations of PM
17 and a range of serious health effects. Particular attention was given to several size-specific
18 classes of particles, including PM10 and the principal fractions of PM10, referred to as the fine
19 (PM2 5) and coarse (PM10_2 5) fractions. PM2 5 refers to particles with an upper 50% cutpoint of
20 2.5-//m aerodynamic diameter. PM10_2 5 refers to those particles with an upper 50% cutpoint of
21 10 //m and a lower 50% cut point of 2.5-//m aerodynamic diameter. In other words, the coarse
22 fraction (PM10_25) refers to the inhalable particles that remain if fine (PM25) particles are removed
23 from a sample of PM10 particles. As discussed in the 1996 PM AQCD, fine and coarse fraction
24 particles can be differentiated by their sources and formation processes and by their chemical and
25 physical properties, including behavior in the atmosphere. Detailed discussions of atmospheric
26 formation, ambient concentrations, and health effects of ambient air PM, as well as quantitative
27 estimates of human health risks associated with exposure to ambient PM, can be found in the
28 1996 PM AQCD and in the 1996 OAQPS SP (U.S. Environmental Protection Agency, 1996b).
29
30
March 2001 1-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.2.2 Presidential Memorandum: Next Particulate Matter Review
2 and Research
3 On July 18, 1997, the EPA published a final rule revising the NAAQS for PM (Federal
4 Register, 1997a) and, on the same day, a final rule revising the Ozone NAAQS (Federal Register,
5 1997b). A Presidential Memorandum (Federal Register, 1997c) also was published outlining the
6 Administration's goals for implementing the revised PM and Ozone NAAQS. The
7 Memorandum directed EPA to provide to CASAC within 90 days and to publish a notice
8 outlining its schedule for the next periodic review of PM and to complete the next review,
9 including review by CASAC, within 5 years after issuance of the revised standards (i.e., by July
10 2002). Such a schedule would ensure that EPA's review of newly emerging scientific
11 information, which forms the criteria on which the standards are based, and of the standards
12 themselves will have been completed prior to any areas being designated as "nonattainment"
13 under the newly established standards for fine particles (i.e., PM2 5 standards) and prior to the
14 imposition of any new controls related to the revised standards. The Presidential Memorandum
15 also directed EPA and other relevant Federal agencies to develop and implement a greatly
16 expanded, coordinated research program. To facilitate timely scientific research within this
17 review period, EPA initiated certain activities immediately, as noted below in the discussion of
18 the PM Research Program.
19
20 Particulate Matter Research Program
21 The EPA broadened its ongoing PM research activities by developing, in partnership with
22 other Federal agencies, a coordinated interagency PM research program. This interagency
23 program has and continues to focus mainly on expanding scientific knowledge of ambient PM
24 exposure and health effects, as well as including development of improved monitoring methods
25 and cost-effective mitigation strategies. The interagency effort also promotes further
26 coordination with other research organizations, including the Health Effects Institute and other
27 state-, university-, and industry-sponsored research groups. Beginning in the fall of 1997, public
28 participation has been and continues to be encouraged through workshops and review of program
29 documentation.
30 To aid identification of needed research efforts, EPA published a particulate matter health
31 risk research needs document (U.S. Environmental Protection Agency, 1998a). That document
March 2001 1-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 identifies research needed to improve scientific information supporting future health risk
2 assessment and review of the PM NAAQS. The document aimed to provide a foundation for PM
3 research coordination among Federal agencies and other research organizations and served as one
4 useful input to National Research Council PM research deliberations. In January 1998, the
5 National Research Council (NRC) established its Committee on Research Priorities for Airborne
6 Particulate Matter in response to a request from Congress in the Fiscal 1998 appropriation to
7 EPA. This Committee is charged to identify the most important research priorities relevant to
8 setting paniculate matter standards, to develop a conceptual plan for particulate matter research,
9 and to monitor research progress toward improved understanding of the relationship between
10 particulate matter and public health. The Committee issued its first report in early 1998
11 (National Research Council, 1998) and a second one in 1999 (National Research Council, 1999).
12 The EPA's PM Research Program includes studies to improve understanding of the
13 formation and composition of fine PM, the characteristics or components of PM that are
14 responsible for its health effects, the mechanisms by which these effects are produced, and
15 improved measurements and estimation of population exposures to PM. Specific EPA research
16 efforts include controlled human exposure studies, in vivo and in vitro toxicology, epidemiology,
17 atmospheric sciences including monitoring and modeling studies, development of data on
18 emissions of fine particles from stationary and mobile sources, and identification and evaluation
19 of risk management options. The results from these efforts, as well as related efforts by other
20 Federal agencies and the general scientific community, are expected to enhance substantially the
21 scientific and technical bases for future decisions on the PM NAAQS and for the implementation
22 of PM monitoring and control efforts.
23
24
25 1.3 CURRENT PARTICULATE MATTER CRITERIA AND NATIONAL
26 AMBIENT AIR QUALITY STANDARDS REVIEW
27 1.3.1 Criteria Review
28 As with all NAAQS reviews, the purpose is to update the criteria and to determine whether
29 it is appropriate to revise existing standards in light of new scientific and technical information.
March 2001 1-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 Although the EPA concluded in its most recent final rule on the PM standards (Federal Register,
2 1997a) that the current scientific knowledge provides a strong basis for the revised PM standards,
3 including the establishment of PM25 standards, there remain scientific uncertainties associated
4 with the health effects of PM and with the means of reducing such effects. Recognizing the
5 importance of developing a better understanding of the effects of fine particles on human health,
6 including their causes and mechanisms, as well as the species and sources of PM25, EPA has and
7 will continue to sponsor research to address these uncertainties even as this criteria review
8 progresses.
9 As with other NAAQS reviews, a rigorous assessment of relevant scientific information is
10 to be presented in this updated, revised PM AQCD being prepared by EPA's NCEA-RTP.
11 Development of the document has and will continue to involve substantial external peer review
12 through (a) public workshops involving the general aerosol scientific community, (b) iterative
13 reviews of successive drafts by CAS AC, and (c) comments from the public. The final document
14 will reflect input received through these reviews and will serve to evaluate and integrate the latest
15 available scientific information to ensure that the review of the PM standards is based on sound
16 science. An earlier (October 1999) First External Review Draft of this updated document was
17 released in the fall of 1999 for public comment and CAS AC review. This Second External
18 Review Draft takes into account the earlier public comments and CASAC review
19 recommendations and includes consideration of relevant new peer-reviewed scientific studies
20 published or accepted for publication from January 1996 through December 2000.
21 Following CASAC review of the First External Review Draft of this revised PM AQCD in
22 December 1999, EPA's OAQPS started to prepare an SP for the EPA Administrator. Drawing
23 on information in this newly revised PM AQCD, the SP will evaluate policy implications of the
24 key studies and scientific information contained in the AQCD and identify critical elements that
25 EPA staff believes should be considered in reviewing the PM standards. The SP is intended to
26 bridge the gap between the scientific review in the AQCD and the public health and welfare
27 policy judgments required of the Administrator in reviewing the PM NAAQS. For that purpose,
28 the SP will present technical analyses, including air quality analyses and a quantitative health risk
29 assessment, and other factors relevant to the evaluation of the PM NAAQS, as well as staff
30 conclusions and recommendations of options for the EPA Administrator's consideration. The SP
March 2001 1-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 also will be reviewed by CASAC and the public, and the final SP will reflect the input received
2 through these reviews.
3 Following completion of the final SP, the Administrator will then announce in the Federal
4 Register proposals for retaining or revising the current PM NAAQS, and opportunities will be
5 provided for public comment and CASAC review of those proposals. Taking into account public
6 comments and CASAC recommendations, final decisions regarding the current PM NAAQS
7 review are scheduled to be promulgated by July 2002.
8
9 1.3.2 Methods and Procedures for Document Preparation
10 The procedures followed for developing this revised PM AQCD build on the knowledge
11 and methods derived from the most recent previous PM, Ozone, and CO AQCD preparation
12 efforts. Briefly, the respective responsibilities for production of the present PM AQCD are as
13 follows. An NCEA-RTP PM team was formed to be responsible for developing and
14 implementing the project plan for preparation of the PM AQCD, taking into account inputs from
15 individuals in other EPA program and policy offices identified as part of the EPA PM Work
16 Group. The resulting project plan (i.e., the PM Document Development Plan) was then
17 discussed with CASAC (May 1998) and appropriately revised. An ongoing literature search has
18 continued to be conducted to identify, to the extent possible, all PM literature published since
19 early 1996. Additionally, EPA published (1) a request for information in the Federal Register
20 asking for recently available research information on PM that may not yet be published and
21 (2) a request for individuals with the appropriate type and level of expertise to contribute to the
22 writing of PM AQCD materials to identify themselves (U.S. Environmental Protection Agency,
23 1998b). Specific authors of chapters or sections of the proposed document were selected on the
24 basis of their expertise on the subject areas and their familiarity with the relevant literature; these
25 include both EPA and non-EPA scientific experts. The project team defined critical issues and
26 topics to be addressed by the authors and provided direction in order to emphasize evaluation of
27 those studies most clearly identified as important for standard setting.
28 The main focus of this revised criteria document is the evaluation and interpretation of air
29 quality data, human exposure information, and health and welfare effects information newly
30 published since that assessed in the 1996 PM AQCD and likely to be useful in deriving criteria
31 for PM NAAQS. Initial draft versions of AQCD chapters were evaluated via expert workshops
March 2001 1-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 and/or expert written peer reviews, which focused on the selection of pertinent studies included
2 in the chapters, the potential need for additional information to be added to the chapters, and the
3 quality of the summarization and interpretation of the literature. The authors of the draft chapters
4 then revised them on the basis of the workshop and/or written expert review recommendations.
5 These and other integrative summary materials were incorporated into the First External Review
6 Draft of the PM AQCD (October, 1999), which was released for public comment and reviewed at
7 a December 1999 CAS AC public meeting. Necessary revisions, based on public comments and
8 the recommendations derived from the December 1999 CAS AC review, as well as evaluation of
9 newly emerging research results, have been incorporated into this Second External Review Draft.
10 The final version of the newly revised PM AQCD will incorporate changes made in response to
11 public comments and CAS AC review of this Second External Review Draft.
12 New research results are being incorporated into this document as they become available.
13 In order to foster timely presentation and publication of newly emerging PM research findings,
14 EPA co-sponsored an Air and Waste Management Association International Speciality
15 Conference, entitled "PM 2000: Particulate Matter and Health", which was held in Charleston,
16 SC, in January 2000. The conference was co-sponsored in cooperation with several other
17 government agencies and/or private organizations that also fund PM research. Topics covered
18 included new research results concerning the latest advances in PM atmospheric sciences (e.g.,
19 PM formation, transport, transformation), PM exposure, PM dosimetry and extrapolation
20 modeling, PM toxicology (e.g., mechanisms, laboratory animal models, human clinical
21 responses), and PM epidemiology. The main purpose of the conference was to facilitate having
22 the latest scientific information available in time for incorporation into this revised draft EPA
23 PM AQCD so as to allow for its release for public comment and CASAC review by December
24 2000. Arrangements were made for scientists to submit written manuscripts on papers or posters
25 presented at the PM 2000 Conference for expedited peer-review by several major journals, so
26 that decisions on acceptance for publication could be made by mid-2000. The evaluations and
27 findings set forth in this Second External Review Draft of the revised PM AQCD include
28 consideration of such PM 2000 papers and extensive additional information published elsewhere
29 since completion of the previous First External Review Draft.
30
March 2001 1-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.3.3 Approach
2 The approach to organization and content of this revised PM AQCD is somewhat different
3 from those used for previous criteria documents. Because the most recent prior document (U.S.
4 Environmental Protection Agency, 1996a) provides an extensive discussion of most topic areas,
5 this new document focuses more specifically on critical issues that have been identified as areas
6 needed to improve the scientific basis (criteria) for PM NAAQS, particularly for those areas in
7 which the information database has continued to evolve rapidly.
8 An initial step was to review the available scientific literature and to focus on the selection
9 of pertinent issues to include in the document as the basis for the development of PM NAAQS
10 criteria. Preliminary issues were identified by the NCEA PM Team and through input from other
11 EPA program and policy offices. Identification of issue topics was derived from the 1996 PM
12 AQCD and SP, their CAS AC and public reviews, from the standard promulgation process, and
13 from EPA's PM Research Needs Document. Further identification and clarification of issues
14 resulted from the NRC review and reports on PM research priorities. The CAS AC review of the
15 PM AQCD Development Plan and public comments on draft AQCD materials at various stages
16 of their development also has played an important role in issue identification.
17 In developing draft materials for inclusion in the revised PM AQCD, detailed review of key
18 new research was undertaken as a first step. However, instead of presenting a comprehensive
19 review of all the literature, emphasis in this revised AQCD is placed on (1) the concise summary
20 of key findings derived from previous PM criteria reviews and (2) evaluation of the most
21 pertinent new key information, with greater emphasis on more interpretive assessment. This
22 approach reflects recommendations made by CASAC.
23 Building on the previous PM AQCD, most of the scientific information selected for review
24 and discussion in the text is from literature published since completion of the previous
25 PM AQCD (U.S. Environmental Protection Agency, 1996a). To aid in development of a concise
26 document, compilation of summary tables of the relevant published literature and selective
27 discussion of the literature has been undertaken, and increased emphasis has been placed in text
28 discussions on interpretive evaluation and integration of key points derived from the newly
29 summarized research results.
30
March 2001 1-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.3.4 Key Issues of Concern
2 Several broad topics related to the main issues of concern addressed by this revised
3 PM AQCD are summarized below. The document reviews and assesses available data bearing
4 on each of the issues identified below.
5 1. Causality. Evaluation of the evidence for or against a causal relationship between health
6 outcomes and ambient PM and/or specific physical-chemical components.
7 • Specific components of interest include size classes such as PM10, PM10_25, PM25, and
8 ultrafme particles. Chemical components include transition metals, acidity, sulfates,
9 nitrates, and organics.
10 • Expand review of foundations of causal inference for associated PM air pollution health
11 effects.
12 • Access new long-term PM exposure and health data to broaden interpretation of long-term
13 exposure findings.
14 • Review data exploring potential mechanisms of response to PM physical-chemical
15 characteristics, response pathway, and exposure-dose-response relationships (laboratory and
16 clinical research).
17 2. Uncertainties. In carrying out overall assessment, address the following types of uncertainty.
18 • Uncertainties between stationary PM monitoring instruments and personal exposure to PM
19 of ambient origin, especially for susceptible groups and their related activity patterns.
20 Specific topics include measurement error in outdoor monitors themselves, use of central
21 monitors for estimates of community concentrations, and the use of community
22 concentrations as a surrogate for personal exposure to particles of ambient origin.
23 • Uncertainties related to particulate matter size fraction, particle number, surface area, and
24 content of semi-volatile components.
25 • Uncertainties about the effects of long-term PM exposure, such as life shortening, and
26 development and progression of disease.
27 • Uncertainties because of coexposure to other pollutants such as O3, SO2, CO, and NO2, and
28 because of meterological factors.
29 • Uncertainties because of potential confounding in epidemiologic studies (e.g., economic
30 factors, demographic and lifestyle attributes, genetic susceptibility factors, occupational
31 exposure, medical care).
March 2001 1-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Uncertainty about shape of concentration-response (CR) relationships and associated
2 community risks (linear and threshold models for CR).
3 • Uncertainty about methods for synthesis of health outcome studies and evaluation of
4 sensitivity and confounding aspects, including but not limited to meta-analyses.
5 3. Biological Mechanisms of Action. Evaluate data examining mechanisms underlying health
6 outcomes of PM. Mechanistic information aids judgment about causality.
7 • New studies have examined mechanisms of action of PM constituents, including transition
8 metals, airborne allergens, and the generation of reactive oxygen species. Different cell
9 types have differing responses to PM components.
10 • Newly published studies also have identified potential mechanisms for the production of
11 cardiac arrhythmias by PM constituents, especially in animal models of disease and suggest
12 that particular attention should be accorded to PM metal constituents.
13 • Although many new animal toxicology studies involve instillation of previously collected
14 particles and this technique is appropriate to study mechanisms of action, extrapolation to
15 human equivalent exposure/doses is uncertain.
16 • Ongoing work on the effects of lung inflammation and PM phagocytosis on subsequent
17 systemic effects, especially cardiac or vascular effects, is needed to provide further
18 information on the relationship between inhaled pollutants and cardiac events.
19 • Interpretation of concentrated ambient particles studies. Newly available information is
20 examined from toxicology studies using devices that concentrate (to variable extents)
21 ambient PM to determine PM concentration-response relationships. Again, difficulties are
22 encountered regarding extrapolation to comparable human exposures to ambient PM levels.
23 4. Susceptible Populations. Examine health outcome data to determine specific risk groups that
24 are more susceptible than normal healthy adults to adverse effects from PM exposure.
25 • Preexisting respiratory or cardiovascular disease in conjunction with advanced age appear to
26 be important factors in PM mortality susceptibility.
27 • For morbidity health endpoints, children and asthmatics potentially may display increased
28 sensitivity to PM exposure. Data will be examined for coherence.
29 • Patterns of respiratory tract deposition, clearance, and retention in susceptible populations
30 have been studied recently and provide evidence of differences in respiratory tract PM
31 deposition for children and small-sized adults and for those with lung diseases.
March 2001 1-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Animal models of lung disease exposed to PM constituents suggest a role for PM in cardiac
2 death.
3 5. Environmental Effects. Evaluate several types of PM welfare effects.
4 • Vegetation and ecosystem effects.
5 • Visibility effects.
6 • Materials damage.
7 • Role of PM in atmospheric radiative transfer and potential consequences for penetration of
8 biologically harmful UVB to the earth's surface and for climate change.
9 6. Background Information Topics Useful in Evaluating Health Risks. Topics include the
10 following.
11 • New monitoring methods, especially methods used in epidemiology studies.
12 • Indicator topics such as PM2 5 versus PMLO; ultrafine; and PM2 5 versus PM10_2 5.
13 • New data patterns of daily and annual concentrations for PM25, PM10_25, and PM10.
14
15
16 1.4 DOCUMENT CONTENT AND ORGANIZATION
17 The present draft document attempts to critically review and assess relevant scientific
18 literature on PM through December 2000. The material selected for review and comment in the
19 text generally comes from the more recent literature published since early 1996, with emphasis
20 on studies conducted at or near PM pollutant concentrations found in ambient air. Literature
21 discussed in detail in the previous 1996 EPA PM AQCD (U.S. Environmental Protection
22 Agency, 1996a) generally is not discussed in depth in this document. However, some limited
23 treatment is included of the earlier studies judged to be potentially useful in deriving PM
24 NAAQS. Key literature is presented mainly in tables and overall interpretive points are
25 discussed mainly in the text.
26 The primary emphasis is on consideration of published material that has undergone
27 scientific peer review. However, in the interest of admitting new and important information
28 expected to become available shortly, some material not yet fully published in the open literature
29 but meeting other standards of scientific reporting (i.e., peer review, quality assurance) are now
30 provisionally included. As noted earlier, emphasis has been placed on studies in the range of
31 current ambient levels. However, studies examining effects of higher concentrations have been
March 2001 1-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 included if they contain unique data or documentation of a previously unreported effect or
2 mechanism. In reviewing and summarizing the literature, an attempt has been made to present
3 alternative points of view where scientific controversy exists.
4 The present document includes nine chapters presented in two volumes. Volume 1
5 contains this general introduction (Chapter 1). It also includes Chapters 2 and 3, which provide
6 background information on physical and chemical properties of PM and related compounds;
7 sources and emissions; atmospheric transport, transformation, and fate of PM; methods for the
8 collection and measurement of PM; and ambient air concentrations. Next, Chapter 4 describes
9 PM environmental effects on vegetation and ecosystems, visibility, man-made materials, and
10 climate, as well as economic impacts of such welfare effects. Chapter 5, which discusses factors
11 affecting exposure of the general population to ambient PM, is also included in Volume 1. The
12 second volume contains Chapters 6 through 9 and the Executive Summary for the entire
13 document. Chapters 6 through 8 evaluate information concerning the health effects of PM
14 (Chapter 6 discusses epidemiological studies; Chapter 7, dosimetry of inhaled particles in the
15 respiratory tract, and Chapter 8, the toxicology of specific types of PM constituents, including
16 laboratory animal studies and controlled human exposure studies). Chapter 9 integrates key
17 information on exposure, dosimetry, and critical health risk issues derived from studies reviewed
18 in the prior chapters.
19 Neither control techniques nor control strategies for abatement of PM are discussed in this
20 document, although some topics covered may be incidentally relevant to control strategies.
21 Issues germane to the scientific basis for control strategies, but not pertinent to the development
22 of NAAQS criteria, are addressed in numerous other documents issued by EPA's OAQPS.
23 Technologies for controlling PM emissions also are discussed in other documents issued by
24 OAQPS. Also, certain issues of direct relevance to standard setting are not addressed explicitly
25 in this document, but instead are analyzed in documentation prepared by OAQPS as part of its
26 regulatory analyses materials. Such analyses include (1) delineation of particular adverse effects
27 that the primary and secondary NAAQS are intended to protect against, (2) exposure analyses
28 and assessment of consequent risk, and (3) discussion of factors to be considered in determining
29 an adequate margin of safety. Key points and conclusions from such analyses will be presented
30 in the PM SP prepared by OAQPS for review by CASAC. Although scientific data contribute
31 significantly to decisions regarding the above issues, their resolution cannot be achieved solely
March 2001 1-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 on the basis of experimentally acquired information. Final decisions on items (1) and (3) are
2 made by the EPA Administrator, as mandated by the CAA.
3 A fourth issue directly pertinent to standard setting is identification of populations at risk,
4 which is basically a selection by EPA of the subpopulation(s) to be protected by the promulgation
5 of a given standard. This issue is addressed only partially in this document. For example,
6 information is presented on factors, such as preexisting disease, that may biologically predispose
7 individuals and subpopulations to adverse effects from exposures to PM. The characterization of
8 population risk, however, requires information above and beyond data on biological
9 predisposition (e.g., information on estimated exposure, activity patterns, and personal habits).
10 Such information is typically addressed in the SP developed by OAQPS.
11
12
March 2001 1-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Code of Federal Regulations. (1977) Appendix B—Reference method for the determination of suspended
3 particulates in the atmosphere (high-volume method). C. F. R. 40: §50.
4 Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30)
5 36:8186-8201.
6 Federal Register. (1987) Revisions to the national ambient air quality standards for paniculate matter. F. R. (July 1)
7 52:24,634-24,669.
8 Federal Register. (1997a) National ambient air quality standards for paniculate matter; final rule. F. R. (July 18)
9 62: 38,652-38,752.
10 Federal Register. (1997b) National ambient air quality standards for ozone; final rule. F. R. (July 18)
11 62:38,856-38,896.
12 Federal Register. (1997c) Implementation of revised air quality standards for ozone and paniculate matter. F. R.
13 (July 18) 62: 38,421-38,422.
14 National Research Council. (1998) Research priorities for airborne paniculate matter. I. Immediate priorities and a
15 long-range research portfolio. Washington, DC: National Academy Press.
16 National Research Council. (1999) Research priorities for airborne paniculate matter. II. Evaluating research
17 progress and updating the portfolio. Washington, DC: National Academy Press.
18 U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air
19 quality standards. U. S. C. 42: §§7408-7409.
20 U.S. District Court of Arizona. (1995) American Lung Association v. Browner. West's Federal Supplement 884
21 F.Supp. 345 (No. CIV 93-643 TUC ACM).
22 U.S. Environmental Protection Agency. (1996a) Air quality criteria for paniculate matter. Research Triangle Park,
23 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
24 U.S. Environmental Protection Agency. (1996b) Review of the national ambient air quality standards for paniculate
25 matter: policy assessment of scientific and technical information. OAQPS staff paper. Research Triangle
26 Park, NC: Office of Air Quality Planning and Standards; report no. EPA/452/R-96-013. Available from:
27 NTIS, Springfield, VA; PB97-115406REB.
28 U.S. Environmental Protection Agency. (1998a) Paniculate matter research needs for human health risk assessment
29 to support future reviews of the national ambient air quality standards for paniculate matter. Research
30 Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
31 U.S. Environmental Protection Agency. (1998b) Review of national ambient air quality standards for paniculate
32 matter. Commer. Bus. Daily: February 19. Available: http://cbdnet.access.gpo.gov/index.html [1999,
33 November 24].
34 Wolff, G. T. (1996) Closure by the Clean Air Scientific Advisory Committee (CASAC) on the staff paper for
35 paniculate matter [letterto Carol M. Browner, Administrator, U.S. EPA]. Washington, DC: U.S.
36 Environmental Protection Agency, Clean Air Scientific Advisory Committee;
37 EPA-SAB-CASAC-LTR-96-008; June 13.
March 2001 1-16 DRAFT-DO NOT QUOTE OR CITE
-------
i 2. PHYSICS, CHEMISTRY, AND MEASUREMENT
2 OF PARTICULATE MATTER
3
4
5 An extensive review of the physics and chemistry of particulate matter (PM) was included
6 in Chapter 3 of the 1996 EPA document Air Quality Criteria for Parti culate Matter (U.S.
7 Environmental Protection Agency, 1996). Chapter 2 of this new version of the PM Air Quality
8 Criteria Document (PM AQCD) provides background information on the physics and chemistry
9 of atmospheric particles that may be useful in reading subsequent sections and chapters.
10 New information needed to understand risk assessment is discussed, with emphasis placed on
11 differences between fine and coarse particles and differences between the nuclei mode and the
12 accumulation mode within fine particles.
13 Chapter 4 of the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996) contained
14 a review of the state-of-the-art of PM measurement technology. Since that time, considerable
15 progress has been made in understanding problems in the measurement of PM mass, chemical
16 composition, and physical parameters. There also has been some progress in developing new and
17 improved measurement techniques. Therefore, a more extensive survey on measurement
18 problems and on newly developed measurement techniques is included below in Section 2.2.
19 For more detail and older references, the reader is referred to Chapter 3 and 4 of the 1996 PM
20 AQCD (U.S. Environmental Protection Agency, 1996).
21
22
23 2.1 PHYSICS AND CHEMISTRY OF PARTICULATE MATTER
24 2.1.1 Definitions
25 Atmospheric particles originate from a variety of sources and possess a range of
26 morphological, chemical, physical, and thermodynamic properties. Examples include
27 combustion-generated particles, such as diesel soot or fly ash; photochemically produced
28 particles, such as those found in urban haze; salt particles formed from sea spray; and soil-like
29 particles from resuspended dust. Some particles are liquid; some are solid. Others may contain a
30 solid core surrounded by liquid. Atmospheric particles contain inorganic ions, metallic
March 2001 2-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 compounds, elemental carbon, organic compounds, and crustal compounds. Some atmospheric
2 particles are hygroscopic and contain particle-bound water. The organic fraction is especially
3 complex, containing hundreds of organic compounds. Primary particles are emitted directly from
4 sources. Secondary particles are formed in the atmosphere from products of chemical reactions
5 of gases from natural and anthropogenic sources such as SO2, NOX, and certain organic
6 compounds. The particle formation process includes nucleation of particles from low-vapor
7 pressure gases emitted from sources or formed in the atmosphere by chemical reactions;
8 condensation of low-vapor pressure gases on existing particles; and coagulation of particles.
9 Thus, any given particle may contain PM from many sources.
10 The composition and behavior of airborne particles are fundamentally linked with those of
11 the surrounding gas. Aerosol may be defined as a suspension of solid or liquid particles in air.
12 The term aerosol includes both the particles and all vapor or gas phase components of air.
13 However, the term aerosol is often used to refer to the suspended particles only. "Particulate" is
14 an adjective and should only be used as a modifier, as in parti culate matter.
15 A complete description of the atmospheric aerosol would include an accounting of the
16 chemical composition, morphology, and size of each particle and the relative abundance of each
17 particle type as a function of particle size (Friedlander, 1970). However, most often the physical
18 and chemical characteristics of particles are measured separately. Size distributions by particle
19 number, from which surface area and volume distributions are calculated, often are determined
20 by physical means, such as electrical mobility or light scattering of suspended particles.
21 Chemical composition usually is determined by analysis of collected samples, although sulfate
22 can be measured in situ. The mass and average chemical composition of particles, segregated
23 according to aerodynamic diameter by cyclones or impactors, can also be determined. However,
24 recent developments in single particle analysis techniques, by electron microscopy with X-ray
25 analysis of single particles (but not agglomerates) collected on a substrate or by mass
26 spectroscopy of suspended particles passing through a sensing volume, provide elemental
27 composition of individual particles by particle size and, thus, are bringing the description
28 envisioned by Friedlander (1970) closer to reality.
29
30
31
March 2001 2-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.1.2 Physical Properties and Processes
2 2.1.2.1 Definitions of Particle Diameter
3 The diameter of a spherical particle may be determined by optical or electron microscopy,
4 by light scattering and Mie theory, by its electrical mobility, or by its aerodynamic behavior.
5 However, atmospheric particles often are not spherical. Therefore, their diameters are often
6 described by an "equivalent" diameter (i.e., that of a unit density sphere that would have the same
7 physical behavior). The aerodynamic diameter is important for particle transport, collection,
8 and respiratory tract deposition. The aerodynamic diameter (Da) depends on the density of the
9 particle. It is defined as the diameter of a spherical particle with a settling velocity equal to that
10 of the particle in question, but with a density of 1 g/cm3. Particles with the same physical size
11 and shape but different densities will have different aerodynamic diameters. Detailed definitions
12 of the various sizes and their relationships are given in standard aerosol textbooks (e.g.,
13 Friedlander [1977], Reist [1984, 1993], Seinfeld and Pandis [1998], Hinds [1999], Vincent
14 [1989, 1995], Willeke and Baron [1993], and Fuchs [1964, 1989]).
15
16 2.1.2.2 Aerosol Size Distributions
17 Particle size, as indexed by one of the "equivalent" diameters, is an important parameter in
18 determining the properties, effects and fate of atmospheric particles. The atmospheric deposition
19 rates of particles, and therefore their residence times in the atmosphere, are a strong function of
20 their aerodynamic diameters. The aerodynamic diameter also influences deposition patterns of
21 particles within the lung. Light scattering is strongly dependent on the optical particle size.
22 Particle size distributions, therefore, have a strong influence on atmospheric visibility and,
23 through their effect on radiative balance, on climate. Studies using impactors or cyclones
24 measure the particle-size distribution directly in aerodynamic diameter. The diameters of
25 atmospheric particles range from 1 nm to 100 //m, thus spanning 5 orders of magnitude.
26 A variety of different instruments, measuring a variety of equivalent diameters, are required to
27 cover this range.
28 Older particle counting studies used optical particle counters to cover the range of 0.3 to
29 30 /u,m diameter. Diameters of particles below 0.5 //m were measured as mobility diameters.
30 The particle diameters used in size distribution graphs from these studies usually are given as
31 physical diameters rather than aerodynamic diameters. In recent years, aerodynamic particle
March 2001 2-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 sizers, which give a direct measurement of the aerodynamic diameter in the range of
2 approximately 0.7 to 10 //m diameter, have been used with electrical mobility analyzers, which
3 measure the mobility diameter from approximately 0.5 //m to very small particles of the order of
4 0.005 //m, to cover the range of regulatory interest. Unfortunately, there is no agreed-upon
5 technique for combining the various equivalent diameters. Some workers use various
6 assumptions to combine the various measurements into one presentation; others report each
7 instrument separately. Therefore, the user of size distribution data must be careful to determine
8 exactly which equivalent diameter is reported. Aerodynamic diameter is the most widely used
9 equivalent diameter. Therefore, particle diameters, unless otherwise indicated, refer to the
10 aerodynamic diameter in the discussions which follow below.
11
12 Particle Size Distribution Functions
13 The distribution of particles with respect to size is an important physical parameter
14 governing their behavior. Because atmospheric particles cover several orders of magnitude in
15 particle size, size distributions often are expressed in terms of the logarithm of the particle
16 diameter, on the X-axis, and the measured differential concentration on the Y-axis:
17 AN/A(logDp) = the number of particles per cm3 of air having diameters in the size range from
18 log Dp to log(Dp + ADp). Because logarithms do not have dimensions, it is necessary to think of
19 the distribution as a function of log(Dp/Dp0), where the reference diameter Dp0 = 1 //m is not
20 explicitly stated. If AN/A(logDp) is plotted on a linear scale, the number of particles between
21 Dp and Dp + ADp is proportional to the area under the curve of AN/A(logDp) versus logDp.
22 Similar considerations apply to distributions of surface, volume, and mass. It has been found that
23 atmospheric aerosol size distributions frequently may be approximated by a sum of log-normal
24 distributions corresponding to the various modes or fractions. When approximated by a function,
25 the distributions are usually given as dN/d(log Dp) rather than AN/A(log Dp).
26
27 Atmospheric Aerosol Size Distributions
28 Averaged atmospheric size distributions are shown in Figures 2-1 through 2-3 (Whitby,
29 1978; Whitby and Sverdrup, 1980). Figure 2-1 describes the number of particles as a function of
30 particle diameter for rural, urban-influenced rural, urban, and freeway-influenced urban aerosols.
31 For some of the same data, the particle volume distribution is shown in Figure 2-2. Figure 2-3
March 2001 2-4 DRAFT-DO NOT QUOTE OR CITE
-------
1,000,000 -
10,000 -
1-
0.01 H
0.0001 -
0.000001 -
\ r
Clean Rural
Urban Influenced Rural
Average Urban
Urban + Freeway
200,000
0.01 0.1 1 10 100
Particle Diameter, Dp (|jm)
150,000
O
_o
T3
100,000 -
50,000 -
0.01 0.1 1 10
Particle Diameter, Dp (|jm)
Figure 2-1. Number of particles as a function of particle diameter: (a) number
concentrations are shown on a logarithmic scale to display the wide range by
site and size; (b) number concentrations for the average urban distribution are
shown on a linear scale for which the area under any part of the curve is
proportional to particle number in that size range.
Source: Whitby and Sverdrup (1980).
1 shows the number, surface, and volume distribution for the grand average continental size
2 distribution. Note that the particle diameter is always shown on a logarithmic scale. The particle
3 number is frequently shown on a logarithmic scale in order to display the wide range in number
4 concentration for different particle sizes and different sites. Volume and surface area, and
5 sometimes number, are shown on an arithmetic scale with the distributions plotted such that the
6 volume, surface area, or number of particles in any specified size range is proportional to the
7 corresponding area under the curve. These distributions show that most of the particles are quite
8 small, below 0.1 //m, whereas most of the particle volume (and therefore most of the mass) is
9 found in particles >0.1 //m.
March 2001
2-5
DRAFT-DO NOT QUOTE OR CITE
-------
n.
Q
D)
%
1
/U
65 -
60 -
55 -
50 -
35 -
30 -
25 -
20 -
15 -
10 -
5 -
o H
;'
,' '. Uru3D lnflU6DCGQ
• '; Rural
: • New Mexico
/ v' A\
/ / \\
/ ^x""x^X
. .j s ^ |»-/ ! ,^..X
70
65
60 -
55 -
50 -
^40 H
^35 H
"3
9. 30 -
O)
=§ 25 '
^ 20 -
15 -
10 -
5 -
0
Average Urban
Urban + Freeway
0.01 0.1 1 10
Particle Diameter, Dp (pm)
100 0.01 0.1 1 10
Particle Diameter, Dp (|jm)
100
Figure 2-2. Particle volume distribution as a function of particle diameter: (a) for the
averaged rural and urban-influenced rural number distributions shown in
Figure 2-1 and a distribution from south central New Mexico, and (b) for the
averaged urban and freeway-influenced urban number distributions shown in
Figure 2-1.
Source: Whitby and Sverdrup (1980) and Kim et al. (1993).
1 An important feature of the mass or volume size distributions of atmospheric aerosols is
2 their multimodal nature. Volume distributions, measured in ambient air in the United States, are
3 almost always found to be bimodal, with a minimum between 1 and 3 //m. The distribution of
4 particles that are mostly larger than the minimum is termed "coarse." The distribution of
5 particles that are mostly smaller than the minimum is termed "fine." Whitby and Sverdrup
6 (1980), Whitby (1978), and Willeke and Whitby (1975) identified three modes: (1) nuclei,
7 (2) accumulation, and (3) coarse. The three modes are most apparent in the freeway-influenced
8 size distribution of Figure 2-2b, in the surface area distribution of Figure 2-3b, and in the
9 in-traffic volume distribution of Figure 2-4. However, the nuclei mode, corresponding to
March 2001
2-6
DRAFT-DO NOT QUOTE OR CITE
-------
CD
_Q
E
CO
^
CO
X
CO
o
Q
O)
o
a.
Q
O)
o
c/
15-
10 -
5-
600-
400-
200-
30 H
(l)
E
Q
O)
20 H
10-
Nn = 7.7x 10
DGNn = 0.013
agn=1.7
Na = 1.3x 10
Nc=4.2
DGNC= 0.97
agc=2.15
i | 11111
Sa = 535
DGSa = 0.19
DGSn = 0.023
Sc=41
DGSC= 3.1
Vn = 0.33
DGVn = 0.031
0.001
• I I I I l|
0.01
(a)
(b)
0.1
1.0
10
100
Figure 2-3. Distribution of coarse (c), accumulation (a), and nuclei- or ultrafine (n) -mode
particles by three characteristics, a) number (N), b) surface area (S) and
c) volume (V) for the grand average continental size distribution. DGV =
geometric mean diameter by volume; DGS = geometric mean diameter by
surface area; DGN = geometric mean diameter by number; Dp = geometric
diameter.
Source: Whitby (1978).
March 2001
2-7
DRAFT-DO NOT QUOTE OR CITE
-------
E
o
7 -
6 -
5 -
O)
2 -
Mechanically
Generated
DGV = 0.018
o =1.6
1 ' I ""I
0.002 0.01
Nuclei Mode
0.1 1
Particle Diameter, Dp, |jm
Accumulation Mode
Fine-Mode Particles
1 ' l ''"I ' r
10
Coarse Mode
Coarse-Mode Particles
100
Figure 2-4. Volume size distribution, measured in traffic, showing fine-mode and
coarse-mode particles and the nuclei and accumulation modes within the
fine-particle mode. DGV (geometric mean diameter by volume, equivalent to
volume median diameter) and og (geometric standard deviation) are shown for
each mode. Also shown are transformation and growth mechanisms (e.g.,
nucleation, condensation, and coagulation).
Source: Adapted from Wilson and Suh (1997).
1 particles below about 0.1 //m, may not be noticeable in volume or mass distributions. The
2 middle mode, from 0.1 to 1 or 2 //m, is the accumulation mode. Fine particles include both the
3 accumulation and the nuclei modes. The third mode, containing particles larger than 1 or 2 //m,
4 is known as the coarse particle mode. The number concentrations of coarse particles are usually
5 too small to be seen in arithmetic plots (Figures 2-lb and 2-3a) but can be seen in a logarithmic
6 plot (Figure 2-lb). Whitby and Sverdrup (1980) observed that rural aerosols, not influenced by
7 sources, have a small accumulation mode and no observable nuclei mode. For urban aerosols,
March 2001
2-8
DRAFT-DO NOT QUOTE OR CITE
-------
1 the accumulation and coarse particle modes are comparable in volume. The nuclei mode is small
2 in volume but it dominates the number distributions of urban aerosols. Whitby's conclusions
3 were based on extensive studies of size distributions in a number of western and midwestern
4 locations during the 1970s (Whitby, 1978; Whitby and Sverdrup, 1980). No size-distribution
5 studies of similar scope have been published since then. Newer results from particle counting
6 and impactor techniques, including data from Europe (U.S. Environmental Protection Agency,
7 1996) and Australia (Keywood et al., 1999, 2000), show similar results.
8
9 Definitions of Particle Size Fractions
10 In the preceding discussion several subdivisions of the aerosol size distribution were
11 identified. Aerosol scientists use four different approaches or conventions in the classification of
12 particles by size: (1) modes, based on the observed size distributions and formation mechanisms;
13 (2) cut point, usually based on the 50% cut point of the specific sampling device; (3) dosimetry
14 or occupational health sizes, based on the entrance into various compartments of the respiratory
15 system; and (4) legally specified, regulatory sizes for air quality standards.
16
17 Modal. The modal classification, first proposed by Whitby (1978), is shown in Figure 2-3.
18 The nuclei mode can be seen clearly in the volume distribution only in traffic or near traffic or
19 other sources of nuclei mode particles (Figure 2-4). The observed modal structure is frequently
20 approximated by several log-normal distributions. Definitions of terms used to describe size
21 distributions in modal terms are given below.
22
23 Coarse Mode: The distribution of particles with diameters mostly greater than the
24 minimum in the particle mass or volume distributions, which generally occurs between
25 1 and 3 //m. These particles are usually mechanically generated (e.g., from wind erosion of
26 crustal material).
27
28 Fine Mode: The distribution of particles with diameters mostly smaller than the minimum
29 in the particle mass or volume distributions, which generally occurs between 1 and 3 //m.
30 These particles are generated in combustion or formed from gases. The fine mode includes
31 the accumulation mode and the nuclei mode.
March 2001 2-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nuclei Mode: That portion of the fine particle mode with diameters below about 0.1 //m.
2 The nuclei mode can be observed as a separate mode in mass or volume distributions only
3 in clean or remote areas or near sources of new particle formation by nucleation.
4 Toxicologists and epidemiologists use ultrafine to refer to particles in the nuclei-mode size
5 range. Aerosol physicists and material scientists tend to use nanoparticles to refer to
6 particles in this size range generated in the laboratory.
7
8 Accumulation Mode: That portion of the fine particle mode with diameters above about
9 0.1 //m. Accumulation-mode particles normally do not grow into the coarse mode.
10 Nuclei-mode particles grow by coagulation (two particles combining to form one) or by
11 condensation (low-equilibrium vapor pressure gas molecules condensing on a particle) and
12 "accumulate" in this size range.
13
14 Over the years, the terms fine and coarse, as applied to particle sizes, have lost the precise
15 meaning given in Whitby's (1978) definition. In any given article, therefore, the meaning of fine
16 and coarse, unless defined, must be inferred from the author's usage. In particular, PM2 5 and
17 fine-mode particles are not equivalent. In this document, the term mode is used with fine and
18 coarse when it is desired to specify the distribution of fine-mode particles or coarse-mode
19 particles as shown in Figures 2-4 and 2-5.
20
21 Sampler Cut Point. Another set of definitions of particle size fractions arises from
22 considerations of size-selective sampling. Size-selective sampling refers to the collection of
23 particles below or within a specified aerodynamic size range, usually defined by the upper 50%
24 cut point size, and has arisen in an effort to measure particle size fractions with some special
25 significance (e.g., health, visibility, source apportionment, etc.). Dichotomous samplers split the
26 particles into smaller and larger fractions, which may be collected on separate filters. However,
27 some fine particles (~ 10%) are collected with the coarse particle fraction. Cascade impactors use
28 multiple size cuts to obtain a distribution of size cuts for mass or chemical composition
29 measurements. One-filter samplers with a variety of upper size cuts also have been used.
30
March 2001 2-10 DRAFT-DO NOT QUOTE OR CITE
-------
E
ro
E
70
60
50 -
JB 40
D)
o
30 -
05
05
< 20
10 -
0.1
Fine-Mode Particles
Coarse-Mode Particles
0.2
0.5 1.0 2 5 10 20
Aerodynamic Particle Diameter Da, [jm
Total Suspended Particles (TSP)
PNL
'10
PM
'2.5
PM
(10-2.5)
TSP
HiVol
WRAC
50 100
Figure 2-5. An idealized distribution of ambient particulate matter showing fine-mode
particles and coarse-mode particles and the fractions collected by size-selective
samplers. (WRAC is the Wide Range Aerosol Classifier which collects the
entire coarse mode [Lundgren and Burton, 1995].)
Source: Adapted from Wilson and Suh (1997).
4
5
Occupational Health or Dosimetric Size Cuts. The occupational health community has
defined size fractions for use in the protection of human health. This convention classifies
particles into inhalable, thoracic, and respirable particles according to their upper size cuts.
However, these size fractions may also be characterized in terms of their entrance into various
compartments of the respiratory system. Thus, inhalable particles enter the respiratory tract,
March 2001
2-11
DRAFT-DO NOT QUOTE OR CITE
-------
1 including the head airways. Thoracic particles travel past the larynx and reach the lung airways
2 and the gas-exchange regions of the lung. Respirable particles are a subset of thoracic particles
3 which are more likely to reach the gas-exchange region of the lung. In the past exact definitions
4 of these terms have varied among organizations. As of 1993, a unified set of definitions was
5 adopted by the American Conference of Governmental Industrial Hygienists (ACGIH) (1994),
6 the International Standards Organization (ISO), and the European Standardization Committee
7 (CEN). The curves which define inhalable (IPM), thoracic (TPM), and respirable (RPM)
8 particulate matter are shown in Figure 2-6.
9
• IPM
• TPM
ORPM
V PM
2.5 4 10 20 50
Aerodynamic Diameter (|jm)
100
Figure 2-6. Specified particle penetration (size-cut curves) through an ideal (no-particle-
loss) inlet for five different size-selective sampling criteria. PM10 is defined in
the Code of Federal Regulations (1991a). 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).
March 2001
2-12
DRAFT-DO NOT QUOTE OR CITE
-------
1 Regulatory Size Cuts. In 1987, the NAAQS for PM were revised to use PM10, rather than
2 total suspended particulate matter (TSP), as the indicator for the NAAQS for PM (Federal
3 Register, 1987). The use of PM10 as an indicator is an example of size-selective sampling based
4 on a regulatory size cut (Federal Register, 1987). The selection of PM10 as an indicator was
5 based on health considerations and was intended to focus regulatory concern on those particles
6 small enough to enter the thoracic region of the human respiratory tract. The PM2 5 standard, set
7 in 1997, is also an example of size-selective sampling based on a regulatory size cut (Federal
8 Register, 1997). The PM25 standard was based primarily on epidemiological studies using
9 concentrations measured with PM2 5 samplers as an exposure index. However, the PM2 5 sampler
10 was not designed to collect respirable particles. It was designed to collect fine-mode particles
11 because of their different sources (Whitby et al., 1974). Thus, the need to attain a PM2 5 standard
12 will tend to focus regulatory concern on control of sources of fine-mode particles.
13 Prior to 1997, the indicator for the NAAQS for PM was TSP. TSP is defined by the design
14 of the High Volume Sampler (hivol), which collects all of the fine particles but only part of the
15 coarse particles. The upper cut-off size of the hivol depends on the wind speed and direction and
16 may vary from 25 to 40 //m. The Wide Range Aerosol Classifier (WRAC) was designed
17 specifically to collect the entire coarse mode (Lundgren and Burton, 1995).
18 An idealized distribution, showing the normally observed division of ambient aerosols into
19 fine-mode particles and coarse-mode particles and the size fractions collected by the WRAC,
20 TSP, PM10, PM25 and PM(10.25) samplers, is shown in Figure 2-5. PM10 samplers, as defined in
21 Appendix J to 40 Code of Federal Regulations (CFR) Part 50 (Code of Federal Regulations,
22 199 la; Federal Register, 1987), collect all of the fine particles and part of the coarse particles.
23 The upper cut point is defined as having a 50% collection efficiency at 10 ± 0.5 Mm aerodynamic
24 diameter. The slope of the collection efficiency curve is defined in amendments to 40 CFR,
25 Part 53, (Code of Federal Regulations, 1991b). An example of a PM10 size-cut curve is shown in
26 Figure 2-6.
27 An example of a PM2 5 size-cut curve is also shown in Figure 2-6. The PM2 5 size-cut
28 curve, however, is defined by the design of the Federal Reference Method Sampler. The basic
29 design of the FRM is given in the Federal Register (1997, 1998) and as 40 CFR Part 50,
30 Appendix L in the Code of Federal Regulations (Code of Federal Regulations, 1999a).
31 Additional performance specifications are given in 40 CFR Parts 53 and 58 (Code of Federal
March 2001 2-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 Regulations, 1999b). Each actual PM25 reference method, as represented by a specific sampler
2 design and associated manual operational procedures, must be designated as a reference method
3 under Part 53 (see Section 1.2 of Appendix L). Thus there may be many somewhat different
4 PM2 5 FRMs (currently, 6 have been designated).
5 Papers discussing PM10 or PM2 5 frequently insert an explanation such as PMX (particles less
6 than x //m diameter) or PMX (nominally, particles with aerodynamic diameter x are collected and not all particles
-------
c
O
'-I—'
03
-i—<
-------
1 number and a significant portion of particle surface area, some further attention to nuclei-mode
2 particles is justified.
3
4 Formation and Growth of Fine Particles
5 Several processes influence the formation and growth of particles. New particles may be
6 formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
7 material condenses on existing particles. Particles also may grow by coagulation as two particles
8 combine to form one. Gas phase material condenses preferentially on smaller particles, and the
9 rate constant for coagulation of two particles decreases as the particle size increases. Therefore,
10 nuclei mode particles grow into the accumulation mode, but accumulation mode particles do not
11 grow into the coarse mode (see Figure 2-4). More information and references on formation and
12 growth of fine particles may be found in the AQC PM 1996 (U.S. Environmental Protection
13 Agency, 1996).
14
15 Equilibrium Vapor Pressures
16 An important parameter in particle nucleation and in particle growth by condensation is the
17 saturation ratio S, defined as the ratio of the partial pressure of a species, p, to its equilibrium
18 vapor pressure above a flat surface, p0: S = p/p0. For either condensation or nucleation to occur,
19 the species vapor pressure must exceed its equilibrium vapor pressure. For particles, the
20 equilibrium vapor pressure is not the same as p0. Two effects are important: (1) the Kelvin
21 effect, which is an increase in the equilibrium vapor pressure above the surface due to its
22 curvature; thus very small particles have higher vapor pressures and will not be stable to
23 evaporation until they attain a critical size; and (2) the solute effect, which is a decrease in the
24 equilibrium vapor pressure of the liquid due to the presence of other compounds in solution.
25 Organic compounds may also be adsorbed on ultrafme carbonaceous particles.
26 For an aqueous solution of a nonvolatile salt, the presence of the salt decreases the
27 equilibrium vapor pressure of the water over the droplet. This effect is in the opposite direction
28 of the Kelvin effect, which increases the equilibrium vapor pressure above a droplet because of
29 its curvature. The existence of an aqueous solution will also influence the vapor pressure of
30 water-soluble species. The vapor pressure behavior of mixtures of several liquids or of liquids
31 containing several solutes is complex.
March 2001 2-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 New Particle Formation
2 When the vapor concentration of a species exceeds its equilibrium concentration (expressed
3 as its equilibrium vapor pressure), it is considered condensable. Condensable species can either
4 condense on the surface of existing particles or can form new particles. The relative importance
5 of nucleation versus condensation depends on the rate of formation of the condensable species
6 and on the surface or cross-sectional area of existing particles (McMurry and Friedlander, 1979).
7 In ambient urban environments, the available particle surface area is sufficient to rapidly
8 scavenge the newly formed condensable species. Formation of new particles (nuclei mode) is
9 usually not important except near sources of condensable species. Wilson et al. (1977) report
10 observations of the nuclei mode in traffic. New particle formation also can be observed in
11 cleaner, remote regions. Bursts of new particle formation in the atmosphere under clean
12 conditions usually occur when aerosol surface area concentrations are low (Covert et al., 1992).
13 High concentrations of nuclei mode particles have been observed in regions with low particle
14 mass concentrations, indicating that new particle formation is inversely related to the available
15 aerosol surface area (Clarke, 1992).
16
17 Sources of Nuclei-Mode Particles
18 Nuclei mode particles are the result of nucleation of gas phase species to form condensed
19 phase species with very low equilibrium vapor pressure. In the atmosphere there are four major
20 classes of sources that yield particulate matter with equilibrium vapor pressures low enough to
21 form nuclei mode particles:
22 (1) Particles containing heavy metals. Nuclei mode particles of metal oxides or other
23 metal compounds are generated when metallic impurities in coal or oil are vaporized during
24 combustion and the vapor undergoes nucleation. Metallic ultrafine particles also may be
25 formed from metals in lubricating oil or fuel additives that are vaporized during
26 combustion of gasoline or diesel fuels. Nuclei-mode metallic particles were discussed in
27 Section 6.9 of the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996).
28 (2) Elemental carbon or soot (EC). EC particles are formed primarily by condensation of
29 C2 molecules generated during the combustion process. Because EC has a very low
30 equilibrium vapor pressure, ultrafine EC particles can nucleate even at high temperatures
31 (Kittelson, 1998; Morawska et al., 1998a).
March 2001 2-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 (3) Sulfcites and nitrates. Sulfuric acid (H2SO4), or its neutralization products with
2 ammonia (NH3), ammonium sulfate ((NH4)2SO4) or ammonium acid sulfate (NH4HSO4),
3 are generated in the atmosphere by conversion of sulfur dioxide (SO2) to H2SO4. As H2SO4
4 is formed, it can either nucleate to form new ultrafme particles, or it can condense on
5 existing nuclei mode or accumulation mode particles (Clark and Whitby, 1975; Whitby,
6 1978). The possible formation of ultrafme NH4NO3 by reaction of NH3 and HNO3
7 apparently has not been investigated.
8 (4) Organic carbon. Recent smog chamber studies and indoor experiments show that
9 atmospheric oxidation of certain organic compounds found in the atmosphere can produce
10 highly oxidized organic compounds with an equilibrium vapor pressure sufficiently low to
11 result in nucleation (Kamens et al., 1999; Weschler and Shields, 1999).
12
13 Concentration of Nuclei-Mode Particles: A Balance Between Formation and Removal
14 Nuclei-mode particles may be removed by dry deposition or by growth into the
15 accumulation mode. This growth takes place as other low vapor pressure material condenses on
16 the particles or as nuclei-mode particles coagulate with themselves or with accumulation mode
17 particles. Because the rate of coagulation would vary with the concentration of accumulation-
18 mode particles, it might be expected that the concentration of nuclei-mode particles would
19 increase with a decrease in accumulation-mode mass. On the other hand, the concentration of
20 particles would be expected to decrease with a decrease in the rate of generation of particles by
21 reduction in emissions of metal and carbon particles or a decrease in the rate of generation of
22 H2SO4 or condensable organic vapor. The rate of generation of H2SO4 depends on the
23 concentration of SO2 and OH, which is generated primarily by the photolysis of O3. Thus, the
24 reductions in SO2 and O3 that are expected to form a major basis for attaining PM2 5 and O3
25 standards and the implementation of Title n and Title IV Clean Air Act programs should lead to
26 a decrease in the rate of generation of H2SO4 and condensable organic vapor and to a decrease in
27 the concentration of nuclei-mode particles.
28 The balance between formation and removal is uncertain. However, these processes can be
29 modeled using a general dynamic equation for particle size distribution (Friedlander, 1977) or by
30 aerosol dynamics modules in newer air quality models (Binkowski and Shanker, 1995;
31 Binkowski and Ching, 1995).
March 2001 2-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 Further research is important due to the possibility of enhanced toxicity of ultrafine
2 particles. It is possible that freshly generated ultrafine particles relatively near significant sources
3 could present an additional risk to health, above those associated with particle mass per se.
4 It will, therefore, be important to monitor particle number and surface as well as mass to further
5 delineate the relative effectiveness of strategies for reducing particle mass, surface, and number.
6
7 2.1.3 Chemistry of Atmospheric Particulate Matter
8 The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
9 ions; particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
10 material. Atmospheric PM also contains a large number of elements in various compounds and
11 concentrations. More information and references on the composition of PM, measured in a large
12 number of studies in the United States, may be found in 1996 PM AQCD (U.S. Environmental
13 Protection Agency, 1996). The composition and concentrations of PM are discussed in
14 Chapter 3 of this document.
15
16 2.1.3.1 Chemical Composition and Its Dependence on Particle Size
17 Studies conducted in most parts of the United States indicate that sulfate, ammonium, and
18 hydrogen ions; elemental carbon, secondary organic compounds and some primary organic
19 compounds; and certain transition metals are found predominantly in the fine particle mode.
20 Crustal materials such as calcium, aluminum, silicon, magnesium, and iron are found
21 predominately in the coarse particles. Some organic materials such as pollen, spores, and plant
22 and animal debris are also found predominantly in the coarse mode. Some components such as
23 potassium and nitrate may be found in both the fine and coarse particle modes but from different
24 sources or mechanisms. Potassium in coarse particles comes from soil. Potassium also is found
25 in fine particles in emissions from burning wood or cooking meat. Nitrate in fine particles comes
26 primarily from the reaction of gas-phase nitric acid with gas-phase ammonia to form particulate
27 ammonium nitrate. Nitrate in coarse particles comes primarily from the reaction of gas-phase
28 nitric acid with preexisting coarse particles.
29
30
31
March 2001 2-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.1.3.2 Primary and Secondary Particulate Matter
2 Particulate material can be primary or secondary. PM is called primary if it is in the same
3 chemical form in which it was emitted into the atmosphere. PM is called secondary if it is
4 formed by chemical reactions in the atmosphere. Primary coarse particles are usually formed by
5 mechanical processes. This includes material emitted in particulate form such as wind-blown
6 dust, sea salt, road dust, and combustion-generated particles such as fly ash and soot. Primary
7 fine particles are emitted from sources, either directly as particles or as vapors that rapidly
8 condense to form ultrafme or nuclei-mode particles. This includes soot from diesel engines,
9 a great variety of organic compounds condensed from incomplete combustion or cooking, and
10 compounds of As, Se, Zn, etc., which condense from vapor formed during combustion or
11 smelting. The concentration of primary particles depends on their emission rate, transport and
12 dispersion, and removal rate from the atmosphere.
13 Secondary PM is formed by chemical reactions of free, adsorbed, or dissolved gases. Most
14 secondary fine PM is formed from condensable vapors generated by chemical reactions of
15 gas-phase precursors. Secondary formation processes can result in either the formation of new
16 particles or the addition of particulate material to preexisting particles. Most of the sulfate and
17 nitrate and a portion of the organic compounds in atmospheric particles are formed by chemical
18 reactions in the atmosphere. Secondary aerosol formation depends on numerous factors
19 including the concentrations of precursors; the concentrations of other gaseous reactive species
20 such as ozone, hydroxyl radical, peroxy radicals, or hydrogen peroxide; atmospheric conditions
21 including solar radiation and relative humidity; and the interactions of precursors and preexisting
22 particles within cloud or fog droplets or on or in the liquid film on solid particles. As a result, it
23 is considerably more difficult to relate ambient concentrations of secondary species to sources of
24 precursor emissions than it is to identify the sources of primary particles. A significant effort is
25 currently being directed toward the identification and modeling of organic products of
26 photochemical smog including the conversion of gases to particulate matter.
27
28 Formation of Sulfcites and Nitrates
29 A substantial fraction of the fine particle mass, especially during the warmer months of the
30 year, is secondary sulfate and nitrate, formed as a result of atmospheric reactions. Such reactions
31 involve the gas phase conversion of SO2 to H2SO4 by OH radicals and aqueous-phase reactions of
March 2001 2-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 SO2 with H2O2, O3, or O2 (catalyzed by Fe and Mn). These heterogeneous reactions may occur in
2 cloud and fog droplets or in films on atmospheric particles. The NO2 portion of NOX can be
3 converted to HNO3 by reaction with OH radicals during the day. At night, NOX also is oxidized
4 to nitric acid by a sequence of reactions initiated by O3, that include nitrate radicals (NO3) and
5 dinitrogenpentoxide (N2O5). Both H2SO4 and HNO3 react with atmospheric ammonia (NH3).
6 Gaseous NH3 reacts with gaseous HNO3 to form particulate NH4NO3. Gaseous NH3 reacts with
7 H2SO4 to form acidic HSO4 (in NH4 HSO4) as well as in SO4 in (NH4)2SO4. In addition, acid
8 gases such as SO2 and HNO3 may react with coarse particles to form coarse secondary PM
9 containing sulfate and nitrate. Examples include reactions with basic compounds resulting in
10 neutralization (e.g., CaCo3 + 2NHO3 - Ca (NO3)2 + H2CO31) or with salts of volatile acids
11 resulting in release of the volatile acid (e.g., SO2 + 2NaCl + H2O - Na2SO3 + 2HC11).
12 If particulate NH4NO3 coagulates with an acidic sulfate particle (H2SO4 or HSO4), gaseous
13 HNO3 will be released and the NH3 will increase the neutralization of the acidic sulfate. Thus, in
14 the eastern United States where PM tends to be acidic, sulfate is usually a larger fraction of PM
15 mass than nitrate. However, in the western United States, where higher NH3 and lower SO2
16 emissions permit complete neutralization of H2SO4, the concentration of nitrate may be higher
17 than that of sulfate. As SO2 concentrations in the atmosphere in the eastern United States are
18 reduced, the NH3 left in the atmosphere after neutralization of H2SO4 will be able to react with
19 HNO3 to form NH4NO3. Therefore, a reduction in SO2 emissions, especially without a reduction
20 in NOX emissions, could lead to an increase in NH4NO3 concentrations (West et al., 1999; Ansari
21 and Pandis, 1998). Thus, possible environmental effects of NH4NO3 are of interest for both the
22 western and eastern United States.
23 Chemical reactions of SO2 and NOX within plumes are an important source of H+, SO4, and
24 NO3. These conversions can occur by gas-phase and aqueous-phase mechanisms. In power-
25 plant or smelter plumes containing SO2 and NOX, the gas-phase chemistry depends on plume
26 dilution, sunlight, and volatile organic compounds, either in the plume or in the ambient air
27 mixing into and diluting the plume. For the conversion of SO2 to H2SO4, the gas-phase rate in
28 such plumes during summer midday conditions in the eastern United States typically varies
29 between 1 and 3% h"1 but in the cleaner western United States rarely exceeds 1% h"1. For the
30 conversion of NOX to HNO3, the gas-phase rates appear to be approximately three times faster
March 2001 2-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 than the SO2 conversion rates. Winter rates for SO2 conversion are approximately an order of
2 magnitude lower than summer rates.
3 The contribution of aqueous-phase chemistry to particle formation in point-source plumes
4 is highly variable, depending on the availability of the aqueous phase (wetted aerosols, clouds,
5 fog, and light rain) and the photochemically generated gas-phase oxidizing agents, especially
6 H2O2 for SO2 chemistry. The in-cloud conversion rates of SO2 to SO4 can be several times
7 larger than the gas-phase rates given above. Overall, it appears that SO2 oxidation rates to SO4
8 by gas-phase and aqueous-phase mechanisms may be comparable in summer, but aqueous phase
9 chemistry may dominate in winter.
10 In the western United States, markedly higher SO2 conversion rates have been reported in
11 smelter plumes than in power plant plumes. The conversion is predominantly by a gas-phase
12 mechanism. This result is attributed to the lower NOX in smelter plumes. In power plant plumes,
13 NO2 depletes OH and competes with SO2 for OH.
14 In urban plumes, the upper limit for the gas-phase SO2 conversion rate appears to be about
15 5% h"1 under the more polluted conditions. For NO2, the rates appear to be approximately three
16 times faster than the SO2 conversion rates. Conversion rates of SO2 and NOX in background air
17 are comparable to the peak rates in diluted plumes. Neutralization of H2SO4 formed by SO2
18 conversion increases with plume age and background NH3 concentration. If the NH3
19 concentrations are more than sufficient to neutralize H2SO4 to (NH4)2SO4, the HNO3 formed from
20 NOX conversions may be converted to NH4NO3.
21
22 Organic Aerosol
23 Organic compounds contribute from 20 to 60% of the dry fine particle mass in the
24 atmosphere (Gray et al., 1984; Shah et al., 1986; U.S. Environmental Protection Agency, 1996).
25 However, organic PM concentrations, composition, and formation mechanisms are poorly
26 understood. Particulate organic matter is an aggregate of hundreds of individual compounds
27 spanning a wide range of chemical and thermodynamic properties (Saxena and Hildemann,
28 1996). Some of the organic compounds are "semivolatile" (i.e., they have atmospheric
29 concentrations and saturation vapor pressures), such that both gaseous and condensed phases
30 exist in equilibrium in the atmosphere. The presence of semivolatile or multiphase organic
31 compounds complicates the sampling process. Organic compounds, originally in the gas phase,
March 2001 2-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 may be absorbed on glass or quartz filter fibers (positive artifact). Semivolatile compounds,
2 originally present in the condensed phase, may evaporate from particles collected on glass,
3 quartz, or Teflon filters (negative artifact). In addition, no single analytical technique is currently
4 capable of analyzing the entire range of organic compounds present in the atmosphere in PM.
5 Even rigorous analytical methods are able to identify only 10 to 20% of the organic PM mass on
6 the molecular level (Rogge et al., 1993a). Even in smog chamber studies of specific compounds,
7 only about 50% of the condensed phase compounds could be identified (Forstner et al., 1997a,b).
8 Measurement techniques are discussed in Section 2.2.3.2. Information on the identification and
9 concentration of the many different organic compounds identified in atmospheric samples is
10 given in Chapter 3.
11
12 Formation of Secondary Organic Particulate Matter
13 Atmospheric reactions involving volatile organic compounds such as alkanes, alkenes,
14 aromatics, cyclic olefms, and terpenes (or any reactive organic gas that contains at least seven
15 carbon atoms) yield organic compounds with low ambient temperature, saturation vapor
16 pressures. Such reactions may occur in the gas phase, in fog or cloud droplets (Graedel and
17 Goldberg, 1983; Faust, 1994) or possibly in aqueous aerosols (Aumont et al., 2000). Reaction
18 products from the oxidation of reactive organic gases also may nucleate to form new particles or
19 condense on existing particles to form secondary organic PM. Organic compounds with two
20 double bounds may react to form dicarboxylic acids, which, with four or more carbon atoms, also
21 may condense. Both biogenic and anthropogenic sources contribute to primary and secondary
22 organic particulate matter (Grosjean, 1992; Hildemann et al., 1996; Mazurek et al., 1997;
23 Schauer et al., 1996). Oxalic acid was the most abundant organic acid found in PM25 in
24 California (Poore, 2000).
25 Although the mechanisms and pathways for forming inorganic secondary particulate matter
26 are fairly well known, those for forming secondary organic PM are not as well understood.
27 Ozone and the hydroxyl radical are thought to be the major initiating reactants. However, HO2
28 and NO3 radicals also may initiate reactions and organic radicals may be nitrated by HNO2,
29 HNO3, or NO2. Pun et al. (2000) discuss formation mechanisms for highly oxidized,
30 multifunctional organic compounds. The production of such species has been included in a
31 photochemical model by Aumont et al. (2000). Understanding the mechanisms of formation of
March 2001 2-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 secondary organic PM is important because secondary organic PM can contribute in a significant
2 way to ambient PM levels, especially during photochemical smog episodes. Experimental
3 studies of the production of secondary organic PM in ambient air have focused on the Los
4 Angeles Basin. Turpin and Huntzicker (1991, 1995) and Turpin et al. (1991) provided strong
5 evidence that secondary PM formation occurs during periods of photochemical ozone formation
6 in Los Angeles and that as much as 70% of the organic carbon in ambient PM was secondary in
7 origin during a smog episode in 1987. Schauer et al. (1996) estimated that 20 to 30% of the total
8 organic carbon PM in the <2.1 //m size range in the Los Angeles airshed is secondary in origin
9 on an annually averaged basis.
10 Pandis et al. (1992) identified three mechanisms for formation of secondary organic PM:
11 (1) condensation of oxidized end-products of photochemical reactions (e.g., ketones, aldehydes,
12 organic acids, hydroperoxides), (2) adsorption of organic gases onto existing solid particles (e.g.,
13 polycyclic aromatic hydrocarbons), and (3) dissolution of soluble gases that can undergo
14 reactions in particles (e.g., aldehydes). The first and third mechanisms are expected to be of
15 major importance during the summertime when photochemistry is at its peak. The second
16 pathway can be driven by diurnal and seasonal temperature and humidity variations at any time
17 of the year. With regard to the first mechanism, Odum et al. (1996) suggested that the products
18 of the photochemical oxidation of reactive organic gases are semivolatile and can partition
19 themselves onto existing organic carbon at concentrations below their saturation concentrations.
20 Thus, the yield of secondary organic PM depends not only on the identity of the precursor
21 organic gas but also on the ambient levels of organic carbon capable of absorbing the oxidation
22 product.
23 Haagen-Smit (1952) first demonstrated that hydrocarbons irradiated in the presence of NOX
24 produce light scattering aerosols. The aerosol forming potentials of a wide variety of individual
25 anthropogenic and biogenic hydrocarbons were compiled by Pandis et al. (1992) based mainly on
26 estimates made by Grosjean and Seinfeld (1989) and data from Pandis et al. (1991) for p-pinene
27 and Izumi and Fukuyama (1990) for aromatic hydrocarbons. Zhang et al. (1992) examined the
28 oxidation of a-pinene. Pandis et al. (1991) found no aerosol products formed in the
29 photochemical oxidation of isoprene, although they and Zhang et al. (1992) found that the
30 addition of isoprene to reaction mixtures increased the reactivity of the systems studied. Further
31 details about the oxidation mechanisms and secondary organic PM yields from various reactive
March 2001 2-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 organic gases given in the above studies. Estimates of the production rate of secondary organic
2 PM in the Los Angeles airshed are provided in the previous PM AQCD (U.S. Environmental
3 Protection Agency, 1996).
4 More recently, Odum et al. (1997a,b) have found that the aerosol formation potential of
5 whole gasoline vapor can be accounted for solely by summing the contributions of the individual
6 aromatic compounds in the fuel. In general, data for yields for secondary organic PM formation
7 can be broken into two distinct categories. The oxidation of toluene and aromatic compounds
8 containing ethyl or propyl groups (i.e., ethylbenzene, ethyltoluene, n-propylbenzene) produced
9 higher yields of secondary organic PM than did the oxidation of aromatic compounds containing
10 two or more methyl groups (i.e., xylenes, di-, tri-, tetra-methylbenzenes). Yields in the first
11 group ranged from about 7 to 10% and in the second group were generally between 3 and 4%
12 within a range of existing organic carbon levels between 13 and 100 //g/m"3. This grouping is
13 consistent with those found by Izumi and Fukuyama (1990). Reasons for the differences in
14 secondary organic PM yields found between the two classes of compounds are not clear.
15 Kao and Friedlander (1995) examined the statistical properties of a number of PM
16 components in the South Coast Air Basin (Los Angeles area). They found that, regardless of
17 source type and location within their study area, the concentrations of nonreactive, primary
18 components of PM10 had approximately log normal frequency distributions with constant values
19 of the geometric standard deviations (GSDs). However, aerosol constituents of secondary origin
20 (e.g., SO4=, NH4+, and NO3") were found to have much higher GSDs. Surprisingly, the GSDs of
21 organic (1.87) and elemental (1.74) carbon were both found to be within la (0.14) of the mean
22 GSD (1.85) for nonreactive primary species, compared to GSD's of 2.1 for sulfate, 3.5 for
23 nitrate, and 2.6 for ammonium. These results suggest that most of the organic carbon seen in
24 ambient samples in the South Coast Air Basin was of primary origin. Pinto et al. (1995) found
25 similar results for data obtained during the summer of 1994. Further studies are needed to
26 determine if these relations are valid at other locations and to what extent the results might be
27 influenced by the evaporation of volatile constituents during or after sampling. It must be
28 emphasized that the inferences drawn from field studies in the Los Angeles Basin are unique to
29 that area and cannot be extrapolated to other areas of the country.
30
31
March 2001 2-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 Secondary Organic Aerosol Formation from Oxidation Products ofBiogenic Hydrocarbons
2 The formation of atmospheric aerosols from biogenic emissions has been of interest for
3 many years. Recently, more quantitative results have been reported. Hoffmann et al. (1997)
4 found secondary organic PM yields of «5% for open-chain biogenic hydrocarbons such as
5 ocimene and linalool, 5 to 25% for monounsaturated cyclic monoterpenes such as a-pinene,
6 d-3 carene and terpinene-4-ol, and -40% for a cyclic monoterpene with two double bonds such
7 as d-limonene. Secondary organic PM yields of close to 100% were observed during the
8 photochemical oxidation of one sesquiterpene, trans-caryophyllene. These results were all
9 obtained for initial hydrocarbon mixing ratios of 100 ppb.
10 Kamens et al. (1999) observed secondary organic PM yields of 20 to 40% for a-pinene.
11 Using information on the composition of secondary PM formed from a-pinene (Jang and
12 Kamens, 1999), they were able to calculate formation rates with a kinetic model including
13 formation mechanisms for O3 + a-pinene reaction products. Griffin et al. (1999) introduced the
14 concept of incremental aerosol reactivity, the change in the secondary organic aerosol mass
15 produced (in //g/m3) per unit change of parent organic reacted (in ppb), as a measure of the
16 aerosol-forming capability of a given parent organic compound in a prescribed mixture of other
17 organic compounds. They measured the incremental aerosol reactivity for a number of aromatic
18 and biogenic compounds for four initial mixtures. Incremental aerosol reactivity ranged from
19 0.133 to 10.352 //gm"3 ppb"1 and varied by almost a factor of two depending on the initial
20 mixture.
21 Recent laboratory and field studies support the concept that nonvolatile and semivolatile
22 oxidation products from the photooxidation of biogenic hydrocarbons contribute significantly to
23 ambient PM concentrations in both urban and rural environments. A number of multifunctional
24 oxidation products have been identified in laboratory studies (Yu et al., 1998; Glasius et al.,
25 2000; Christoffersen et al., 1998; Koch et al., 2000; Leach et al., 1999). Many of these
26 compounds have subsequently been identified in field investigations (Yu et al., 1999; Kavouras
27 et al., 1998, 1999a,b; Casimiro et al., 2001; Castro et al., 1999). However, further investigations
28 are needed to accurately access their overall contributions to PM2 5 concentrations.
29 Sampling and characterization of PM in the ambient atmosphere and in important
30 microenvironments is required to address important issues in exposure, toxicology, and
31 compliance. Currently, it is not possible to fully quantify the concentration, composition, or
March 2001 2-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 sources of the organic components. Many of the secondary organic aerosol components are
2 highly oxidized, difficult to measure, multifunctional compounds. Additional laboratory studies
3 are needed to identify such compounds, strategies need to be developed to sample and measure
4 such compounds in the atmosphere, and models of secondary organic aerosol formation need to
5 be improved and added to air quality models in order to address compliance issues related to
6 reducing PM mass concentrations that affect human exposure.
7 A high degree of uncertainty is associated with all aspects of the calculation of secondary
8 organic PM concentrations. This is compounded by the volatilization of organic carbon from
9 filter substrates during and after sampling as well as potential positive artifact formation from the
10 absorption of gaseous hydrocarbon on quartz filters. Significant uncertainties always arise in the
11 interpretation of smog chamber data because of wall reactions. Limitations also exist in
12 extrapolating the results of smog chamber studies to ambient conditions found in urban airsheds
13 and forest canopies. Concentrations of terpenes and NOX are much lower in forest canopies
14 (Altshuller, 1983) than the levels commonly used in smog chamber studies. The identification of
15 aerosol products of terpene oxidation has seldom been a specific aim of field studies, making it
16 difficult to judge the results of model calculations of secondary organic PM formation.
17 Uncertainties also arise because of the methods used to measure biogenic hydrocarbon emissions.
18 Khalil and Rasmussen (1992) found much lower ratios of terpenes to other hydrocarbons (e.g.,
19 isoprene) in forest air than were expected, based on their relative emissions strengths and rate
20 coefficients for reaction with OH radicals and O3. They offered two explanations: (1) either the
21 terpenes were being removed rapidly by some heterogeneous process, or (2) emissions were
22 enhanced artificially by feedbacks caused by the bag enclosures they used. If the former
23 consideration is correct, then the production of aerosol carbon from terpene emissions could be
24 substantial; if the latter is correct, then terpene emissions could have been overestimated by the
25 techniques used.
26
27 2.1.3.3 Particle-Vapor Partitioning
28 Several atmospheric aerosol species, such as ammonium nitrate and certain organic
29 compounds, are semivolatile and are found in both gas and particle phases. A variety of
30 thermodynamic models have been developed to predict the temperature and relative humidity
31 dependence of the ammonium nitrate equilibria with gaseous nitric acid and ammonia. However,
March 2001 2-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 under some atmospheric conditions, such as cool, cold, or very clean air, the relative
2 concentrations of the gas and solid phases are not accurately predicted by equilibrium
3 considerations alone, and transport kinetics can be important. The gas-particle distribution of
4 semivolatile organic compounds depends on the equilibrium vapor pressure of the compound,
5 total particle surface area, particle composition, atmospheric temperature, and relative humidity.
6 Although it generally is assumed that the gas-particle partitioning of semivolatile organics is in
7 equilibrium in the atmosphere, neither the equilibria nor the kinetics of redistribution are well
8 understood. Diurnal temperature fluctuations, which cause gas-particle partitioning to be
9 dynamic on a time scale of a few hours, can cause semivolatile compounds to evaporate during
10 the sampling process. The pressure drop across the filter can also contribute to loss of
11 semivolatile compounds. The dynamic changes in gas-particle partitioning, caused by changes in
12 temperature, pressure, and gas-phase concentration, both in the atmosphere and after collection,
13 cause serious sampling problems that are discussed in Section 3.2.3.
14
15 Equilibria with Water Vapor
16 As a result of the equilibrium of water vapor with liquid water in hygroscopic particles,
17 many ambient particles contain liquid water (particle-bound water). Unless removed, this
18 particle-bound water will be measured as a component of the particle mass. Particle-bound water
19 is important in that it influences the size of the particles and in turn their aerodynamic properties
20 (important for deposition to surfaces, to airways following inhalation, and in sampling
21 instrumentation) and their light scattering properties. The aqueous solution provides a medium
22 for reactions of dissolved gases, including reactions that do not take place in the gas phase. The
23 aqueous solutions also may act as a carrier to convey soluble toxic species to the gas-exchange
24 regions of the respiratory system, including species that would be removed by deposition in the
25 upper airways if in the gas phase (Friedlander and Yeh, 1998; Kao and Friedlander, 1995;
26 Wilson, 1995). An extensive review of this equilibrium as it pertains to ambient aerosols was
27 given in Chapter 3 of the 1996 PM AQCD (U.S. Environmental Protection, Agency, 1996).
28 The interaction of particles with water vapor may be described briefly as follows.
29 As relative humidity increases, particles of crystalline soluble salts, such as (NH4) 2SO4,
30 NH4HSO4, or NH4NO3, undergo a phase transition to become aqueous solution particles.
31 According to the phase rule, for particles consisting of a single component, this phase transition
March 2001 2-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 is abrupt, taking place at a relative humidity that corresponds to the vapor pressure of water
2 above the saturated solution (the deliquescence point). With further increase in relative
3 humidity, the solution particle adds water (and the concentration of the solute decreases) so that
4 the vapor pressure of the solution is maintained equal to that of the surrounding relative
5 humidity; thus, the solution particle tends to follow the equilibrium growth curve. As relative
6 humidity decreases, the solution particle follows the equilibrium curve to the deliquescence
7 point. However, rather than crystallizing at the deliquescence relative humidity, the solute
8 remains dissolved in a supersaturated solution to considerably lower relative humidities.
9 Ultimately the solution particle abruptly loses its water vapor (efflorescence), returning typically
10 to the initial crystalline form.
11 For particles consisting of more than one component, the solid to liquid transition will take
12 place over a range of relative humidities, with an abrupt onset at the lowest deliquescence point
13 of the several components, and with subsequent growth as crystalline material in the particle
14 dissolves according to the phase diagram for the particular multicomponent system. Under such
15 circumstances, a single particle may undergo several more or less abrupt phase transitions until
16 the soluble material is fully dissolved. At decreasing relative humidity, such particles tend to
17 remain in solution to relative humidities well below the several deliquescence points. In the case
18 of the sulfuric acid-ammonium sulfate-water system, the phase diagram is fairly completely
19 worked out. Mixed anion systems containing nitrate are more difficult because of the
20 equilibrium between paniculate NH4NO3 and gaseous NH3 and HNO3. For particles of
21 composition intermediate between NH4HSO4 and (NH4)2SO4, this transition occurs in the range
22 from 40% to below 10%, indicating that for certain compositions the solution cannot be dried in
23 the atmosphere. At low relative humidities, particles of this composition would likely be present
24 in the atmosphere as supersaturated solution droplets (liquid particles) rather than as solid
25 particles. Thus, they would exhibit hygroscopic rather than deliquescent behavior during relative
26 humidity cycles.
27 Other pure compounds, such as sulfuric acid (H2SO4), are hygroscopic (i.e., they form water
28 solutions at any relative humidity and maintain a solution vapor pressure over the entire range of
29 relative humidity). Soluble organic compounds may also contribute to the hygroscopicity of the
30 atmospheric aerosol (Saxena et al., 1995; Saxena and Hildeman, 1996), but the equilibria
31 involving organic compounds and water vapor, and especially for mixtures of salts, organic
March 2001 2-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 compounds, and water, are not so well understood. These equilibrium processes may cause an
2 ambient particle to significantly increase its diameter at relative humidities above about 40%
3 (Figure 2-8). A particle can grow to five times its dry diameter as the RH approaches 100%
4 (Figure 2-9). The Federal Reference Methods, for filter measurements of PM2 5 and PM10 mass,
5 require, after collection, equilibration at a specified, low relative humidity (-40% RH). This
6 equilibration removes much of the particle-bound water and provides a stable PM mass (see
7 Section 2.2 for details and references). Otherwise, particle mass would be a function of relative
8 humidity and, at higher relative humidities, the particle mass would be largely particle-bound
9 water.
10 Continuous monitoring techniques must remove particle-bound water before measurement,
11 either by heating or dehumidification. Semivolatile material may be lost during sampling or
12 equilibration. It is certainly lost when the collected sample is heated above ambient. In addition
13 to problems due to the loss of semivolatile species, recent studies have shown that significant
14 amounts of particle-bound water are retained in particles collected on impaction surfaces even
15 after equilibration and that the amount of retained particle-bound water increases with relative
16 humidity during collection (Hitzenberger et al., 1997). Large increases in mass with increasing
17 relative humidity were observed for the accumulation mode. The change in particle size with
18 relative humidity also means that particle measurements such as surface area or volume, or
19 composition as a function of size, all must be made at the same RH if the results are to be
20 comparable. These problems are addressed below in more detail, in Section 2.2 on Measurement
21 of Paniculate Matter.
22
23 2.1.3.4 Removal Processes
24 The lifetimes of particles vary with size. Coarse particles can settle rapidly from the
25 atmosphere within hours, and normally travel only short distances. However, when mixed high
26 into the atmosphere, as in dust storms, the smaller-sized coarse-mode particles may have longer
27 lives and travel distances. Nuclei mode particles rapidly grow into the accumulation mode.
28 However, the accumulation mode does not grow into the coarse mode. Accumulation-mode fine
29 particles are kept suspended by normal air motions and have very low deposition rates to
30 surfaces. They can be transported thousands of km and remain in the atmosphere for a number of
31 days. Coarse-mode particles of less than ~ 10 //m diameter, as well as accumulation-mode and
March 2001 2-30 DRAFT-DO NOT QUOTE OR CITE
-------
90
Figure 2-8. Particle growth curves showing fully reversible hygroscopic growth of sulfuric
acid (H2SO4) particles, deliquescent growth of ammonium sulfate [(NH4)2 SO4]
particles at about 80% relative humidity (RH), hygroscopic growth of
ammonium sulfate solution droplets at RH greater than 80%, and hysteresis
(the droplet remains supersaturated as the RH decreases below 80%) until the
crystallization point is reached.
Source: National Research Council (1993) adapted from Tang (1980).
1 nuclei-mode (or ultrafme) particles, all have the ability to penetrate deep into the lungs and to be
2 removed by deposition in the lungs. Dry deposition rates are expressed in terms of a deposition
3 velocity that varies with particle size, reaching a minimum between 0.1 and 1.0 //m aerodynamic
4 diameter. Accumulation-mode particles are removed from the atmosphere primarily by cloud
5 processes. Fine particles, especially particles with a hygroscopic component, grow as the relative
March 2001
2-31
DRAFT-DO NOT QUOTE OR CITE
-------
o
Q
CO
a:
o
(5
5-
4 -
3 -
2 -
T
T
0
Theoretical Prediction at 22°c
o o o o o Experimental Measurements
RH=99.8%
50 100 150
NH4 HSO4 Dry Particle Diameter (nm)
216
-125
-64
-27
-8
0
200
o
Figure 2-9. Theoretical predictions and experimental measurements of growth of
NH4HSO4 particles at relative humidity between 95 and 100%.
Source: Lietal. (1992).
1 humidity increases, serve as cloud condensation nuclei, and grow into cloud droplets. If the
2 cloud droplets grow large enough to form rain, the particles are removed in the rain. Falling rain
3 drops impact coarse particles and remove them. Ultrafine or nuclei mode particles are small
4 enough to diffuse to the falling drop, be captured, and be removed in rain. Falling rain drops,
5 however, are not effective in removing accumulation-mode particles.
March 2001
2-32
DRAFT-DO NOT QUOTE OR CITE
-------
1 2.1.3.5 Particulate Matter and Welfare Effects
2 The EPA is required by law to set primary standards to protect human health and secondary
3 standards to mitigate welfare effects. The role of particles in reducing visibility and affecting
4 radiative balance through scattering and absorption of light is evident, as are the effects of
5 particles in soiling and damaging materials. Visibility effects are addressed through regional
6 haze regulations and are considered in establishing secondary NAAQS. The direct effects of
7 particles in scattering and absorbing light and the indirect effects of particles on clouds are being
8 addressed in climate change programs in several government agencies with the lead role assigned
9 to the Department of Energy. These welfare effects are discussed briefly in Chapter 4. The
10 effects on vegetation resulting from the direct and indirect effects of particles on light flux also
11 are discussed in Chapter 4.
12 Concerns over the possible ecological effects of acid deposition in the United States led to
13 the creation of a major research program in 1980 under the new National Acid Precipitation
14 Assessment Program (NAPAP). However, the role of PM in acid deposition has not always been
15 recognized. Acid deposition and PM are intimately related, however, first because particles
16 contribute significantly to the acidification of rain and secondly because the gas phase species
17 that lead to dry deposition of acidity are also precursors of particles. Therefore, reductions in
18 SO2 and NOX emissions will decrease both acidic deposition and PM concentrations.
19 Sulfate, nitrate, and some partially oxidized organic compounds are hygroscopic and act as
20 nuclei for the formation of cloud droplets. These droplets serve as chemical reactors in which
21 (even slightly) soluble gases can dissolve and react. Thus, SO2 can dissolve in cloud droplets and
22 be oxidized to sulfuric acid by dissolved ozone or hydrogen peroxide. These reactions do not
23 take place in the gas phase but only in solution in water. Sulfur dioxide also may be oxidized by
24 dissolved oxygen. This process will be faster if metal catalysts such as iron or manganese are
25 present in solution. If the droplets evaporate, larger particles are left behind. If the droplets grow
26 large enough, they will fall as rain, and the particles will be removed from the atmosphere with
27 potential effects on the materials, plants, or soil on which the rain falls. (Similar considerations
28 apply to dew.) Atmospheric particles that nucleate cloud droplets also may contain other soluble
29 or nonsoluble materials such as metal salts and PNA organic compounds that may add to the
30 toxicity of the rain. Thus, the adverse effects of acid deposition on soils, plants, and trees as well
March 2001 2-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 as lakes, streams and fish must be taken into account in setting secondary PM standards. These
2 effects are discussed in Chapter 4.
3 Sulfuric acid, ammonium nitrate, ammonium sulfates, and organic particles also are
4 deposited on surfaces by dry deposition. The utilization of ammonium by plants leads to the
5 production of acidity. Therefore, dry deposition of particles can also contribute to the ecological
6 damages caused by acid deposition.
7
8 2.1.4 Summary
9 The physical and chemical properties of ultrafine mode, accumulation mode, and coarse
10 mode particles are summarized in Table 2-1.
11
12
13 2.2 MEASUREMENT OF PARTICULATE MATTER
14 The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996) summarized sampling
15 and analytical techniques for PM and acid deposition that had appeared in the literature since the
16 earlier 1982 PM AQCD (U.S. Environmental Protection Agency, 1982). Excellent reviews have
17 been published by Chow (1995) and McMurry (2000). This section discusses problems in
18 measuring PM; new techniques that attempt to alleviate these problems or measure problem
19 species; the current EPA monitoring program (including measurements with Federal Reference
20 Methods, speciation monitors, and continuous monitors); and the importance of intercomparison
21 studies in the absence of any reference standard for suspended atmospheric particles.
22
23 2.2.1 Problems in Measuring Particulate Matter
24 The EPA decision to revise the PM standards by adding daily and yearly standards for
25 PM2 5 has led to a renewed interest in the measurement of atmospheric particles and also to a
26 better understanding of the problems in developing precise and accurate measurements of
27 particles. Unfortunately, it is very difficult to measure and characterize particles suspended in
28 the atmosphere.
29 The U.S. Federal Reference Methods (FRM) for PM25 and PM10 provide relatively precise
30 (±10 %) methods for determining the mass of material remaining on a Teflon filter after
March 2001 2-34 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2-1. COMPARISON OF AMBIENT PARTICLES,
FINE MODE (Nuclei Mode Plus Accumulation Mode) AND COARSE MODE
Fine
Coarse
Nuclei
Accumulation
Formed from:
Formed by:
Combustion, high-temperature
processes, and atmospheric reactions
Composed of:
Solubility:
Sources:
Atmospheric
half-life:
Removal
Processes:
Nucleation
Condensation
Coagulation
Sulfates
Elemental carbon
Metal compounds
Organic compounds
with very low,
saturation vapor
pressure at ambient
temperature
Probably less
soluble than
accumulation mode
Combustion
Atmospheric
transformation of
SO2 and some
organic compounds
High temperature
processes
Minutes to hours
Grows into
accumulation mode
Travel distance: <1 to 10s of km
Condensation
Coagulation
Evaporation of fog and
cloud droplets in which
gases have dissolved and
reacted
Sulfate, SC>4
Nitrate, NO3
Ammonium, NH^,
Hydrogen ion, H+
Elemental carbon,
Large variety of organic
compounds
Metals: compounds of Pb,
Cd, V, Ni, Cu, Zn, Mn, Fe,
etc.
Particle-bound water
Largely soluble,
hygroscopic, and
deliquescent
Combustion of coal, oil,
gasoline, diesel fuel, wood
Atmospheric transformation
products of NOX, SO2, and
organic compounds,
including biogenic organic
species (e.g., terpenes)
High-temperature
processes, smelters, steel
mills, etc.
Days to weeks
Forms cloud droplets and
rains out
Dry deposition
100s to 1000s of km
Break-up of large solids/droplets
Mechanical disruption (crushing,
grinding, abrasion of surfaces)
Evaporation of sprays
Suspension of dusts
Reactions of gases in or on particles
Suspended soil or street dust
Fly ash from uncontrolled combustion
of coal, oil, and wood
Nitrates/chlorides fromHNO3/HCl
Oxides of crustal elements
(Si, Al, Ti, Fe)
CaCO3, NaCl, sea salt
Pollen, mold, fungal spores
Plant and animal fragments
Tire, brake pad, and road wear debris
Largely insoluble and nonhygroscopic
Resuspension of industrial dust and
soil tracked onto roads and streets
Suspension from disturbed soil (e.g.,
farming, mining, unpaved roads)
Construction and demolition
Uncontrolled coal and oil combustion
Ocean spray
Biological sources
Minutes to hours
Dry deposition by fallout
Scavenging by falling rain drops
-------
1 equilibration. However, numerous uncertainties remain as to the relationship between the mass
2 and composition of material remaining on the filter, as measured by the FRMs, and the mass and
3 composition of material that existed in the atmosphere as suspended PM. The goal of a PM
4 indicator might be to measure accurately what exists as a particle in the atmosphere. However,
5 this is not currently possible, in part because of the difficulty of creating a reference standard for
6 particles suspended in the atmosphere. As a result, EPA defines accuracy for PM measurements
7 in terms of agreement of a candidate sampler with a reference sampler. Therefore,
8 intercomparisons of samplers become very important in determining how well various samplers
9 agree and how various design choices influence what is actually measured.
10 There are five general areas where choices must be made in designing an aerosol indicator.
11 These include (1) treatment of semivolatile components; (2) selection of an upper cut point;
12 (3) separation of fine-mode and coarse-mode PM; (4) treatment of pressure, temperature, and
13 relative humidity; and (5) assessment of the reliability of the measurement technique. In many
14 cases, choices have been made without adequate knowledge or understanding of the
15 consequences. As a result, measurement methods developed by different organizations may give
16 different results when sampling the same atmosphere, even though the techniques appear to be
17 identical.
18
19 2.2.1.1 Treatment of Semivolatile Components of Particulate Matter
20 Current filtration-based mass measurements lead to significant evaporative losses, during
21 and possibly after collection, of a variety of semivolatile components (i.e., species that exist in
22 the atmosphere in dynamic equilibrium between the condensed phase and gas phase). Important
23 examples include ammonium nitrate, semivolatile organic compounds, and particle-bound water.
24 This problem is illustrated in Figure 2-10.
25 Possible approaches that have been used to address the problem of potentially lost
26 semivolatile components include those that follow, which will be discussed in more detail in
27 subsequent sections.
28 1. Collect/measure all components present in the atmosphere in the condensed phase except
29 particle-bound water. (Examples: Brigham Young absorptive sampler, Harvard pressure drop
30 monitor. Both require preconcentration of the accumulation mode and reduction of ambient
31 humidity.)
March 2001 2-36 DRAFT-DO NOT QUOTE OR CITE
-------
Should be
retained
Particle-bound water
should be removed
Mineral/Metal
0.1
1.0
2.5
Aerodynamic Diameter, jjrn
Semivolatile components subject to evaporation during or after sampling
Figure 2-10. Schematic showing major nonvolatile and semivolatile components of PM25.
Semivolatile components are subject to partial to complete loss during
equilibration or heating. The optimal technique would be to remove all
particle-bound water but no ammonium nitrate or semivolatile organic PM.
1 2. Stabilize PM at a specified temperature high enough to remove all particle-bound water. This
2 results in loss of most of the semivolatile PM. (Examples: TEOM operated at 50 °C beta
3 gauge with heated inlet.)
4 3. Equilibrate collected material at fixed, near-room temperature and moderate relative humidity
5 (RH) to remove most particle-bound water. Accept the loss of an unknown but possibly
6 significant fraction of semivolatile PM. (Example: U.S. Federal Reference Method and most
7 filter-weighing techniques.) (Note: Equilibration originally was designed to remove adsorbed
March 2001
2-37
DRAFT-DO NOT QUOTE OR CITE
-------
1 water vapor from glass fiber filters in order to maintain a stable filter weight. The designated
2 RH (40%) was a compromise. If the RH is too low, electrostatic charging becomes a
3 problem. The equilibration process does help provide a stable and reproducible mass. It also
4 reduces the particle-bound water. However, it may not remove all particle-bound water.
5 The amount of semivolatile material lost is dependent on the concentration and
6 composition of the semivolatile components and is, therefore, also dependent on season and
7 location. The amount of semivolatile material lost has been found to be significant in air sheds
8 with high nitrate, wood smoke, or secondary organic aerosols.
9
10 2.2.1.2 Upper Cut Point
11 A technique must be used that gives an upper cut-point, and its standard deviation, that is
12 independent of wind speed and direction (the classical high volume sampler head was
13 unsatisfactory because of radial asymmetry). A separation that simulates the removal of particles
14 by the upper part of the human respiratory system would appear to be a good choice (i.e.,
15 measure what gets into the lungs). The ACGIH-ISO-CEN penetration curve for thoracic
16 particles, with a 50% cut-point at 10 //m aerodynamic diameter (AD), would be an appropriate
17 choice. (Thoracic particles are able to pass the larynx and penetrate into the bronchial and
18 alveolar regions of the lung.) Some countries, however, use PM10 to refer not to samplers with a
19 50% cut at 10 //m AD but samplers with 100% rejection of all particles greater than 10 //m AD.
20 Such samplers miss too much of the thoracic PM. The U.S. PM10 separation curve, while sharper
21 than the thoracic curve, is probably satisfactory both for regulatory and health risk monitoring.
22 It has the advantage of reducing the problem of maintaining the finite collection efficiency
23 specified by the thoracic penetration curve for particles larger than 10 //m AD. (See Figure 2-6
24 and Section 2.1.2.2.)
25
26 2.2.1.3 Cut Point for Separation of Fine-Mode and Coarse-Mode Particulate Matter
27 Fine-mode and coarse-mode particles differ not only in size and morphology (e.g., smooth
28 droplets versus rough solid particles) but also in formation mechanisms; sources; and chemical,
29 physical, and biological properties. They also differ in terms of dosimetry (deposition in the
30 respiratory system), toxicity, and health effects as observed by epidemiologic studies. The many
31 reasons for wanting to collect fine and coarse particles separately and considerations as to the
March 2001 2-3 8 DRAFT-DO NOT QUOTE OR CITE
-------
1 appropriate outpoint for separating fine and coarse particles were discussed in Chapter 3 of the
2 1996PM AQCD (U.S. Environmental Protect!on Agency, 1996). A review of atmospheric
3 particle-size-distribution data did not provide a clear or obvious rationale for selection of an
4 appropriate cutpoint. Depending on conditions, a significant amount of either fine- or
5 coarse-mode material may be found in the intermodal region between 1 and 3 //m. However, the
6 analysis of the existing data did demonstrate the important role of relative humidity in
7 influencing the size of particles in the accumulation mode.
8 At high relative humidity, such as that found in fog and clouds, hygroscopic fine-mode
9 particles will increase in size due to accumulation of particle-bound water. Under such
10 conditions, some, originally submicrometer, fine-mode PM may be found with an AD above
11 1 //m. At very low relative humidity, coarse-mode particles may be fragmented into smaller
12 sizes, and small amounts of coarse-mode PM may be found with an AD below 1 //m (Lundgren
13 et al., 1984). Thus, a PM2 5 sample will contain most of the fine-mode material, except during
14 periods of RH near 100 %. However, especially under conditions of low RH, it may contain 5 to
15 20% of the coarse-mode material below 10 //m in diameter. A cut point of 1.0 //m would reduce
16 the misclassification of coarse-mode material as fine, but under high RH conditions could result
17 in some fine-mode material being misclassified as coarse. A reduction in RH, either
18 intentionally or inadvertently, will reduce the size of the fine mode. A sufficient reduction in RH
19 will yield a dry fine-particle mode with very little material above 1.0 //m. Studies of the changes
20 in particle size with changes in relative humidity suggest that only a small fraction of
21 accumulation mode particles will be above 1 //m in diameter at RH below 60% but a substantial
22 fraction will grow above 1 //m for RH above 80% (Hitzenberger et al., 1997; McMurry and
23 Stolzenburg, 1989; U.S. Environmental Protection Agency, 1996).
24 It is desirable to separate fine-mode PM and coarse-mode PM as cleanly as possible in
25 order to properly allocate health effects to either fine-mode PM or coarse-mode PM and to
26 correctly determine sources by factor analysis and/or chemical mass balance. For example,
27 sulfate in the fine-mode is associated with hydrogen or ammonium ions; sulfate in the coarse
28 mode is associated with basic metal ions. The sources are different and the health effects may be
29 different. Transition metals in the coarse mode are likely to be associated with soil and tend to
30 be less soluble than transition metals in the fine mode, which may be found in fresh combustion
31 particles.
March 2001 2-39 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
In areas where winds cause high concentrations of wind blown soil, the current practice of
separating fine-mode and coarse-mode particles at 2.5 //m AD may not provide the best
separation for exposure or epidemiologic studies. 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 PM(2 54) go up. During the peak of the dust storm, PM<2^.1)
was 88% of PM25. For the 24-h period, PM(2 54) was 54% of PM25. However, PMj was not
biased by the intrusion of coarse-mode particles.
1 ' I ' I ' I ' I ' I ' I ' I ' I ' I ' I ' T
12am 2am 4am 6am Sam 10am 12pm 2pm 4pm 6pm 8pm 10pm
Local Time, August 30,1996
Figure 2-11. Particulate matter concentrations in Spokane, WA, during the August 30,
1996 dust storm.
Source: Claiborn et al. (2000).
1 Under conditions of high relative humidity, a cut point near 1 //m AD may reject some
2 fine-mode material. Under these circumstances, a monitor using a 1.0 //m AD cut point can
March 2001
2-40
DRAFT-DO NOT QUOTE OR CITE
-------
1 achieve better modal separation if the air stream is dehumidified to some fixed humidity that
2 would remove all or most particle-bound water without evaporating semivolatile components.
3 New techniques have been developed for both integrated and continuous measurement of fine
4 particulate matter minus particle-bound water, but including semivolatile nitrate and organic
5 compounds. These techniques (see Section 2.2.5) require reduction of RH prior to collection.
6 With such techniques, PMX would be a good indicator of fine-mode particles.
7
8 2.2.1.4 Treatment of Pressure, Temperature, and Relative Humidity
9 There are a variety of techniques for defining (or ignoring) the pressure, temperature, and
10 relative humidity during or after sampling.
11 Temperature and Pressure
12 (a) Sample volume based on mass or volumetric flow corrected to standard temperature
13 and pressure (273 K and 1 atm.).
14 (b) Sample volume based on volumetric flow at ambient conditions of temperature and
15 pressure.
16 Temperature During Collection
17 (a) Heat enough to remove all particle-bound water (i.e., TEOM at 50 °C).
18 (b) Heat several degrees to prevent condensation of water in sampling system.
19 (c) Try to maintain sampler near ambient temperature.
20 (d) Maintain sampler at constant temperature inside heated/air conditioned shelter.
21 Temperature After Collection
22 (a) No control
23 (b) Constant Temperature (room temperature)
24 (c) Store at cool temperature (4 °C)
25 Relative Humidity
26 Changes in relative humidity cause changes in particle size of hygroscopic or deliquescent
27 particles. Changing relative humidity by adding or removing water vapor affects
28 measurements of the following items.
29 (a) Particle number, particle surface area and particle size distribution
30 (b) Amount of overlap of fine-mode and coarse-mode particles
March 2001 2-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 Changing relative humidity by intentional or inadvertent changes in temperature affects
2 above measurements plus the following.
3 (c) Amount of loss of ammonium nitrate and semivolatile organic compounds.
4 Studies of relationships between personal/indoor/outdoor measurements present special
5 problems. Indoor environments are typically dryer than outdoors and may be warmer or, if
6 air-conditioned, cooler. These differences may change particle size and the amount of
7 volatilization of semivolatile components. Such changes between indoors and outdoors will
8 complicate the comparison of indoor to outdoor concentrations, the modeling of personal
9 exposure to all particles, and exposure apportionment by the disaggregation of personal exposure
10 into exposure to particles of ambient origin and exposure to particles of indoor origin.
11
12 2.2.1.5 No Way To Determine Accuracy for Ambient Particulate Matter Mass
13 Measurement
14 Precision is typically determined by comparison of collocated samplers or through replicate
15 analyses, while accuracy is determined through the use of traceable calibration standards.
16 Unfortunately, no standard reference calibration material or procedure has been developed for
17 suspended, atmospheric PM. It is possible to determine the accuracy of certain components of
18 the PM measurement system (e.g., flow control, inlet aspiration, PM25 cut, weighing, etc.). The
19 absolute accuracy for collecting a test aerosol can also be determined by isokinetic sampling in a
20 wind tunnel. However, it is not currently feasible to provide a simulated atmospheric aerosol
21 with naturally occurring semivolatile components. It is particularly challenging to develop an
22 atmospheric aerosol calibration standard suitable for testing samplers in the field. Therefore, it is
23 not possible at the present time to establish the absolute accuracy of a PM monitoring technique.
24 Intercomparison studies, to establish the precision of identical monitors and the extent of
25 agreement between different types of monitors, are essential for establishing the reliability of PM
26 measurements. Intercomparison studies have contributed greatly to our understanding of the
27 problems in PM measurement. Such studies will be discussed as they apply to specific
28 measurement problems, monitoring instruments, or analytical techniques.
29 Some measurement errors of concern in PM10 sampling, including those that arise due to
30 uncertainty tolerances in cutpoint, particle bounce and reentrainment, impactor surface
31 overloading, and losses to sampler internal surfaces, were discussed in detail in the 1996 PM
March 2001 2-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 AQCD (U.S. Environmental Protection Agency, 1996). Other measurement errors of concern in
2 PM2 5 sampling arise because of our inability to assess accuracy in an absolute sense due to a lack
3 of an atmospheric aerosol calibration standard, because of the inclusion in PM2 5 of a small
4 amount of coarse particles, and because of problems associated with the definition of PM25 as
5 what remains after collection on a filter and equilibration rather than the mass of particles as they
6 exist in the air. However, it is possible to measure PM indicators with high precision.
7 Because of the difficulties associated with determining the accuracy of PM measurements,
8 EPA has sought to make FRM measurements equivalent by specifying operating conditions and,
9 in the case of PM25 samplers, by specifying details of the sampler design. Thus, both the PM10 as
10 well as the PM25 standards are defined with consistency of measurement technique, rather than
11 accuracy of the true mass concentration measurement, in mind (McMurry, 2000). It is
12 acknowledged in the Federal Register (1997) that, "because the size and volatility of the particles
13 making up ambient particulate matter vary over a wide range and the mass concentration of
14 particles varies with particle size, it is difficult to define the accuracy of PM25 measurements in
15 an absolute sense...." Thus, accuracy is defined as the degree of agreement between a field PM2 5
16 sampler and a collocated PM25 reference method audit sampler (McMurry, 2000). The Federal
17 Reference Method (FRM) for PM2 5 is discussed below in Section 2.2.3.3. As mentioned earlier,
18 volatilization losses, during sampling or post-sampling handling, of some organics as well as
19 ammonium nitrate can lead to significant underestimation of the true fine parti culate mass
20 concentration in some locations. Sources of error in the measurement of true PM2 5 mass also
21 arise because of adsorption or desorption of semivolatile vapors onto or from collected PM, filter
22 media, or other sampler surfaces; neutralization of acid or basic vapors on either filter media or
23 collected PM; and artifacts associated with particle-bound water.
24 During the past 25 years, there have been advancements in the generation and classification
25 of monodisperse aerosols, as well as in the development of electron microscopy and imaging
26 analysis, that have contributed to the advancement in aerosol calibration (Chen, 1993). Still, one
27 of the limitations in PM sampling and analysis remains the lack of primary calibration standards
28 for evaluating analytical methods and for intercomparing laboratories. Klouda et al. (1996)
29 examined the possibility of resuspending the NIST Standard Reference Material 1649 (Urban
30 Dust) in air for collection on up to 320 filters simultaneously, using SRI, International's dust
31 generation and collection system. However, the fine component is not resuspended and the
March 2001 2-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 semivolatile component has evaporated so this material is not a suitable standard for suspended
2 PM. Little additional work in this area has been reported.
3 Methods validation was discussed in the 1996 PM AQCD (U.S. Environmental Protection
4 Agency, 1996), and the usefulness of intercomparisons and "internal redundancy" was
5 emphasized. For example, a number of internal consistency checks are applied to the IMPROVE
6 network (Malm et al., 1994). These include mass balances, sulfur measurements by both proton
7 induced X-ray emission (PIXE) and ion chromatography (1C), and comparison of organic matter
8 by combustion and by proton elastic scattering analysis (PESA) analysis of hydrogen. Mass
9 balances compare the gravimetrically determined mass with the mass calculated from the sum of
10 the major chemical components (i.e., crustal elements plus associated oxygen, organic carbon,
11 elemental carbon, sulfate, nitrate, ammonium, and hydrogen ions). Mass balances are useful
12 validation techniques; however, they do not check for, or account for, artifacts associated with
13 the absorption of gases during sampling, or the loss of semi-volatile material during sampling.
14 The mass balance check may appear reasonable even if such artifacts are present because only the
15 material collected on the filter is included in the balance.
16
17 2.2.2 Why Measure Particles
18 2.2.2.1 Attainment of a Standard
19 A critical need for particle measurements is to determine if a location is in compliance with
20 an existing standard and to determine if trends show improvements in air quality. For this
21 purpose, precision of the measurement by the variety of indicators in use is the most important
22 consideration. Therefore, intercomparisons of various potential indicators, under a variety of
23 atmospheric and air quality conditions, are essential.
24
25 2.2.2.2 Implementation of a Standard
26 In order to reduce pollution to attain a standard, local agencies and national research
27 organizations need measurements to identify source categories and to develop and validate air
28 quality models. For these purposes, PM parameters other than mass, such as chemical
29 composition and size distribution, must also be measured. Moreover, measurements are needed
30 with shorter time resolution in order to match changes in pollution with diurnal changes in the
31 boundary layer.
March 2001 2-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.2.3 Determination of Health Effects
2 PM measurements are needed to determine exposure for use in epidemiological studies, to
3 assess exposure for risk assessment, and to determine components of PM to guide planning and
4 interpretation of toxicologic experiments. For these purposes, size and chemical composition
5 may be needed. For exposure assessment, different measurement time intervals may be needed.
6 For epidemiologic studies of acute (short-term PM exposures), 1-h or continuous measurements
7 may be needed as well as 24-h measurements. However, for epidemiologic studies of chronic
8 PM exposures, measurements that integrate over longer intervals (e.g., a week to a month) may
9 be more cost effective. For dosimetric studies and modeling, information will be needed on the
10 particle size distribution and on the behavior of particles as the relative humidity and temperature
11 are increased to those in the respiratory system.
12
13 2.2.2.4 Determination of Ecological Effects
14 Measurements of particles, and of the chemical components of particulate matter in rain,
15 fog and dew, are needed to understand the contributions of PM to soiling of surfaces and damage
16 to materials and to understand the wet and dry deposition of acidity and toxic substances to
17 surface water, soil, and plants. Some differentiation into particle size is needed to determine dry
18 deposition. Information on chemical composition is also needed to understand materials damage
19 and ecological damage.
20
21 2.2.2.5 Determination of Radiative Effects
22 Particles reduce visibility by scattering and absorbing light. They also have a direct effect
23 on the climate by scattering visible and ultraviolet light back into space and, indirectly, as cloud
24 condensation nuclei, by changing the albedo and stability of clouds. For understanding these
25 effects, information is needed on refractive index (including ratio of scattering to absorption),
26 size distribution, and change in particle size with change in relative humidity.
27
28 2.2.2.6 Particulate Matter Components/Parameters That Need To Be Measured
29 PM parameters and components of PM that need to be measured for various purposes are
30 summarized in Table 2-2.
31
March 2001 2-45 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2-2. PARTICULATE MATTER COMPONENTS/PARAMETERS
OF INTEREST FOR HEALTH, ECOLOGICAL, OR RADIATIVE EFFECTS;
FOR SOURCE CATEGORY APPORTIONMENT STUDIES;
OR FOR AIR QUALITY VALIDATION STUDIES
Particle number
Particle surface area
Particle size distribution
PM mass (fine-mode [PMj 0] and coarse-mode [PM1(M] mass as well as PM2 5 and PM10);
nonvolatile mass, Federal Reference mass, and 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
l 2.2.3 Problems Associated with Semivolatile Particulate Matter
2 It is becoming increasingly apparent that the semivolatile component of PM may impact
3 significantly the quality of the measurement and can lead to both positive and negative sampling
4 artifacts. Losses of semivolatile species, like ammonium nitrate and many organic species, may
5 occur during sampling because of changes in temperature, relative humidity, or composition of
6 the aerosol or because of the pressure drop across the filter (McMurry, 1999). Gas phase organic
7 species, both volatile and semivolatile, may adsorb onto or react with filter media or collected
8 PM, leading to a positive sampling artifact. Quartz fiber filters have a large specific surface area
March 2001 2-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 on which adsorption of gases can occur. A number of other types of filters (e.g., stretched Teflon
2 membrane filters) have much smaller exposed surface areas (Turpin et al., 1994) and appear to be
3 subject to less adsorption (Kirchstetter et al., 2000; Turpin et al., 1994). Tsai and Huang (1995)
4 observed positive sulfate and nitrate artifacts on high-volume PM10 quartz filters and attributed
5 the artifacts to interactions between acidic gases SO2, HONO, and HNO3 and both the filter
6 media (either glass fiber or quartz) and the coarse particles collected on the filter. Volatilization
7 losses also have been reported to occur during sample transport and storage (Chow, 1995).
8 Evaporative losses of paniculate nitrates have been investigated in laboratory and field
9 experiments (e.g., Wang and John, 1988), and in theoretical studies (Zhang and McMurry, 1992).
10 It has been known for some time that volatilization losses of SVOC can be significant (e.g.,
11 Eatough et al., 1993).
12 The theory describing phase equilibria of SVOC continues to be developed. Liang et al.
13 (1997), Jang et al. (1997), and Strommen and Kamens (1997) have modeled the gas/particle
14 partitioning of SVOC on inorganic, organic, and ambient smog aerosols.
15 Adsorption of organic vapors onto quartz filters is recognized as a source of positive
16 sampling error. This artifact has been examined in experiments in which two quartz fiber filters
17 were deployed in series. The second quartz filter may indicate gaseous VOC adsorbed on both
18 filters (positive artifact) or SVOC evaporated from particles on the first filter and subsequently
19 adsorbed on the second filter (negative artifact), or a combination of both effects. Unless the
20 individual compounds are identified, the investigator does not know what to do with the loading
21 value on the second filter (i.e., to add or subtract from the first filter loading value).
22 The developing state of the art in which diffusion denuder technology is being applied to
23 SVOC sampling (e.g., Eatough et al., 1993; Gundel et al., 1995), as well as for sampling of gas
24 and particulate phase organic acids (Lawrence and Koutrakis, 1996a,b), holds promise for
25 improving the understanding of SVOC sampling artifacts. In a denuder-based system, gas-phase
26 organics are removed by diffusion to an adsorbent surface (e.g., activated carbon, special
27 polymer resins, etc.). Particles then are collected on a filter downstream of the denuder and the
28 remaining organic vapors (i.e., from denuder breakthrough and volatile losses from the collected
29 particles) are collected in an adsorbent downstream of the filter (e.g., charcoal or carbon-
30 impregnated filters, polyurethane foam, or polystyrene-divinylbenzene resin [XAD]).
March 2001 2-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 Finally, Eatough et al. (1999a) have reported on a batch sampler (the Particle Concentrator-
2 Brigham Young University Organic Sampling System [PC-BOSS]) and a continuous sampler
3 (Real-Time Air Monitoring System or RAMS), which attempt to correct simultaneously for
4 volatilization losses of both nitrate and SVOC. These samplers are discussed in more detail in
5 Section 2.2.3.2.
6
7 2.2.3.1 Particulate Nitrates
8 It is well known that volatilization losses of particulate nitrates (e.g., Zhang and McMurry
9 [1992]; see also Hering and Cass [1999] and references therein) occur during sampling on Teflon
10 filters. The impact on the accuracy of atmospheric particulate measurements from these
11 volatilization losses is more significant for PM2 5 than for PM10. The FRM for PM2 5 suffers loss
12 of nitrates, similar to the losses experienced with other simple filter collection systems.
13 Sampling artifacts resulting from the loss of particulate nitrates represents a significant problem
14 in areas such as southern California that experience high amounts of nitrates. Hering and Cass
15 (1999) examined the errors in PM25 mass measurements because of volatilization of particulate
16 nitrate by looking at data from two field measurement campaigns conducted in southern
17 California: (1) the Southern California Air Quality Study (SCAQS, Lawson, 1990) and (2) the
18 1986 CalTech study (Solomon et al., 1992). In both these studies, side-by-side sampling of PM25
19 was conducted. One sampler collected particles directly onto a Teflon filter. The second
20 sampler consisted of a denuder to remove gaseous nitric acid followed by a nylon filter that
21 absorbs the HNO3 which evaporates from ammonium nitrate. In both studies, the denuder
22 consisted of MgO-coated glass tubes (Appel et al., 1981). Fine particulate nitrate collected on
23 the Teflon filter was compared to fine particulate nitrate collected on the denuded nylon filter.
24 In both studies, the PM2 5 mass lost because of volatilization of ammonium nitrate represented a
25 significant fraction of the total PM2 5 mass. The fraction of mass lost was higher during summer
26 than during fall (17% versus 9% during the SCAQS study and 21% versus 13% during the
27 CalTech study) (Figure 2-12). In regard to percentage loss of nitrate, as opposed to percentage
28 loss of mass discussed above, Hering and Cass (1999) found that nitrate remaining on the Teflon
29 filter samples was on average 28% lower than that on the denuded nylon filters.
30 Hering and Cass (1999) also analyzed these data by extending the evaporative model
31 developed by Zhang and McMurry (1987). The extended model utilized by Hering and Cass
March 2001 2-48 DRAFT-DO NOT QUOTE OR CITE
-------
0)
E
'>
ro
O
•5
80%
60% -
40% -
SCAQS Data Set
o Summer Measurements
• Fall Measurements
20% -
50
100
150
200
250
PM2.5 Gravimetric Mass (ng/m )
80%
-------
1 during the summer measurements. When the impactor collection was compared to the Teflon
2 filter collection for a nonvolatile species (sulfate), the results were in agreement.
3 It should be noted that during these intercomparison studies, filters or collection surfaces
4 were removed immediately after sampling and placed into vials containing a basic extraction
5 solution. Therefore, losses that might occur during handling, storage, and equilibration of filters
6 or impaction surfaces were avoided. The loss of nitrate observed from Teflon filters and
7 impaction surfaces in this study, therefore, is a lower limit compared to losses that might occur
8 during the normal processes involved in equilibration and weighing of filters and impaction
9 surfaces. Brook and Dann (1999) measured particulate nitrate in Windsor and Hamilton,
10 Ontario, Canada, by three techniques: (1) a single Teflon filter in a dichotomous sampler, (2) the
11 Teflon filter in an annular denuder system (ADS), and (3) total nitrate including both the Teflon
12 filter and the nylon back-up filter from the ADS. The dicot Teflon filter averaged only 13% of
13 the total nitrate. The Teflon filter from the ADS averaged 46% of the total nitrate. The authors
14 conclude that considerable nitrate was lost from the dicot filters during handling, which included
15 weighing and XRF measurement in a vacuum.
16 Kim et al. (1999) also examined nitrate sampling artifacts by comparing denuded and
17 undenuded quartz and nylon filters, during the PM10 Technical Enhancement Program (PTEP) in
18 the South Coast Air Basin of California. They observed negative nitrate artifacts (losses) for
19 most measurements; however, for a significant number of measurements they observed positive
20 nitrate artifacts. Kim et al. (1999) pointed out that random measurement errors make it difficult
21 to measure true amounts of nitrate loss.
22 Several diffusion denuder samplers have been developed to account for the nitrate lost
23 because of volatilization from filters, many of which were discussed in the 1996 PM AQCD
24 (U.S. Environmental Protection Agency, 1996). Eatough et al. (1999a) developed a high-volume
25 diffusion denuder system in which diffusion denuder and particle concentrator techniques were
26 combined (see Section 2.2.3.2). The particle concentrator reduces the flow through the denuder
27 so that the denuder can be operated for weeks without a loss of collection efficiency, thus making
28 the sampler suitable for routine field sampling. The system was evaluated for the collection of
29 fine particulate sulfate and nitrate in Riverside, CA (Eatough et al., 1999b). Concentrations of
30 PM25 nitrate obtained from the PC-BOSS agreed with those obtained using the Harvard-EPA
31 Annular Denuder Sampler, HEADS (Koutrakis et al., 1988a).
March 2001 2-50 DRAFT-DO NOT QUOTE OR CITE
-------
1 In atmospheres with high sulfate and low ammonia, the PM tends to be acidic (NH4HSO4
2 or H2SO4), and nitric acid remains in the vapor phase. In atmospheres with lower sulfate and
3 higher ammonia, there may be sufficient ammonia to fully neutralize the H2SO4 and also react
4 with HNO3 to form NH4NO3 particles. In the United States, therefore, loss of nitrate will be a
5 bigger problem in the western United States than in the eastern United States. However, as SO2
6 emissions are reduced in the eastern United States, nitrate may become a larger fraction of the
7 suspended PM.
8
9 2.2.3.2 Semivolatile Organic Compounds
10 Semivolatile organic compounds (SVOC) can similarly be lost from Teflon filters because
11 of volatilization, causing the PM2 5 mass to be significantly underestimated (negative artifact).
12 Like particulate nitrates, the FRM for PM2 5 will suffer loss of SVOC, similar to the losses
13 experienced with other simple filter collection systems. When PM is collected on a quartz filter
14 in a system without a denuder, the quartz filter may adsorb some gas phase organic compounds
15 (positive artifact) as well as SVOC that evaporate from collected particles. A second quartz
16 filter, placed directly after either a quartz or Teflon first filter, could also collect some gas phase
17 organic compounds passing through the first filter as well as SVOC that evaporated from
18 particles collected on the first filter. Some workers (Turpin et al., 2000) suggest subtracting the
19 organic carbon mass on the quartz second filter from that on the quartz first filter to correct for
20 the positive artifact. However, if some SVOC, lost by evaporation from particles collected on
21 the first filter, are adsorbed on the quartz second filter, the negative artifact would be doubled
22 (Eatough et al., 1994; Cui et al., 1998). Using their multichannel diffusion denuder sampling
23 system (BOSS), Eatough et al. (1995) reported that, for samples collected at the South Coast Air
24 Quality Management District sampling site at Azusa, CA, changes in the phase distribution of
25 SVOC could result in a loss on average of 35% of the particulate organic material. Cui et al.
26 (1998) found that losses of SVOC from particles in the Los Angeles Basin during the summer
27 were greater during the night (average, 62%) than during the day (average, 42%).
28 The percent SVOC lost from the front filter in a filter-denuder system may be greater than
29 that lost in a filter-only system such as the FRM. In a filter-denuder system, the gas-phase
30 component of the SVOC is removed. Because of the absence of the gas phase, SVOC collected
31 on the filter might evaporate more rapidly in a filter-denuder system than in a filter-only
March 2001 2-51 DRAFT-DO NOT QUOTE OR CITE
-------
1 collection system. To determine the fraction of SVOC lost from a Teflon filter in a filter-only
2 system, it is necessary to compare the amount measured by a nondenuder system with that
3 measured by a denuder system. At present, little information is available on the volatilization
4 losses of SVOC. However, in one study (Pang et al., 2000), the total mass on denuded and
5 undenuded filters were compared and found to be identical within error limits (R2 = 0.816, slope
6 = 0.961 ± 0.027 for total mass compared to R2 = 0.940, slope = 0.986 ± 0.020 for sulfate). This
7 suggests that the major cause of loss of SVOC is the pressure drop across the filter.
8 In addition to their contribution to suspended PM mass, SVOC are also of interest because
9 of their possible health effects. SVOC include products of incomplete combustion such as
10 polycyclic aromatic hydrocarbons (PAHs) and poly cyclic organic matter, which has been
11 identified as a hazardous air pollutant. PAHs also have been suggested as alternative particulate
12 tracers for automobile emissions, because the phase-out of organo-lead additives to gasoline
13 means that lead is no longer a good tracer for automobiles (Venkataraman et al., 1994). PAHs
14 also are emitted during biomass burning, including burning of cereal crop residues and wood
15 fuels (Jenkins et al., 1996; Roberts and Corkill, 1998).
16 Several investigators have observed that collection of particles on a filter can result in
17 underestimation of particulate organic compounds because of losses of semivolatile organic
18 material during sample collection (negative sampling artifact) (Eatough et al., 1993; Tang et al.,
19 1994; Eatough et al., 1995; Gundel et al., 1995; Finn et al., 2000). Positive sampling artifacts
20 also can occur because of the adsorption of gases onto the filter materials (e.g., Gundel et al.,
21 1995). There appears to be a larger positive artifact because of adsorption of organic vapor onto
22 quartz fiber filters than to Teflon filters (Turpin et al., 1994; Chow et al., 1994, 1996; Eatough et
23 al., 1996; Finn et al., 2000). When samples for organic analysis are collected on quartz fiber
24 filters, the amount of adsorbed organic vapor on the quartz filter is sometimes estimated by the
25 amount collected on a second quartz fiber filter behind the first, or by the amount collected on a
26 quartz fiber filter placed behind a Teflon filter in a parallel sampling port (Novakov et al., 1997).
27 Many, but by no means all, investigators subtract this adsorption estimate from the front filter
28 quantity to obtain the mass of collected particulate organic (Turpin et al., 2000).
29 Kirchstetter et al. (2000) report that adsorptive properties of quartz fiber filters vary with lot
30 number, and therefore front and back-up filters should be taken from the same lot. Recent
31 literature suggests that a Teflon-quartz back-up filter appears to provide a better estimate of the
March 2001 2-52 DRAFT-DO NOT QUOTE OR CITE
-------
1 adsorption of gases on a quartz fiber front filter than does a quartz-quartz backup, and that the
2 difference between these two adsorption estimates can be substantial for short durations
3 (Kirchstetter et al., 2000; Turpin et al., 2000). The typically lower organic carbon loadings on
4 quartz-quartz back-up filters, relative to Teflon-quartz back-up samples collected concurrently, is
5 believed to occur because adsorption on the quartz front filter acts to reduce the gas-phase
6 concentration downstream until gas phase (i.e., adsorbed phase equilibrium has been achieved in
7 the vicinity of the front quartz filter surface). Because Teflon filters have little affinity for
8 organic vapors, this equilibrium occurs almost instantaneously for Teflon filters, and the Teflon-
9 quartz back-up filter is exposed to the ambient concentration of organic vapors from the
10 beginning of the sampling period. It might be expected that the quantity of organic vapor
11 adsorbed on a quartz filters would depend on the organic composition and would vary by season
12 and location.
13
14 2.2.3.3 Use of Denuder Systems To Measure Semivolatile Compounds
15 Phase distribution of semivolatile organic species has been the subject of several studies
16 that have employed denuder technology (see Gundel et al., 1995; Gundel and Lane, 1999) to
17 directly determine the phase distributions while avoiding some of the positive and negative
18 sampling artifacts associated with using back-up quartz filters. In an ideal system with a denuder
19 that is 100% efficient, the gas phase would be collected in the denuder and the particle phase
20 would be the sum of the material collected on the filter and the adsorbent downstream. Denuder
21 collection efficiency depends on the denuder surface area (+), the diffusivity (+) and vapor
22 pressure (-) of the compound, the temperature (-) and flow rate (-) of the air stream, and the
23 presence of competing species (-), including water vapor (Cui et al., 1998; Kamens and Coe,
24 1997; Lane et al., 1988). (The + and - symbols in parentheses indicate qualitatively the effect
25 increasing each parameter would have on efficiency). In a system with a denuder collection
26 efficiency less than 100%, the collection efficiency must be known to accurately attribute
27 adsorbed organics from denuder breakthrough to the gas phase and adsorbed organics volatilized
28 from collected particles to the particle phase. In calculating the overall phase distributions of
29 SVOC PAH from a denuder system, the collection efficiency for each compound is needed.
30 The efficiency of silicone-grease-coated denuders for the collection of polynuclear aromatic
31 hydrocarbons was examined by Coutant et al. (1992), who examined the effects of uncertainties
March 2001 2-53 DRAFT-DO NOT QUOTE OR CITE
-------
1 in the diffusion coefficients, and in the collisional reaction efficiencies, on the overall phase
2 distributions of SVOC PAH calculated using denuder technology. In their study, they used a
3 single stage, silicone-grease-coated aluminum annular denuder, with a filter holder mounted
4 ahead of the denuder, and an XAD trap deployed downstream of the denuder. In a series of
5 laboratory experiments, they spiked the filter with a mixture of perdeuterated PAH, then swept
6 the system with ultra-high purity air for several hours, and then analyzed the filter and the XAD.
7 They found that the effects of these uncertainties, introduced by using a single compound as a
8 surrogate PAH (in their case, naphthalene) for validation of the denuder collection efficiency, are
9 less significant than normal variations because of sampling and analytical effects. Results on
10 field studies using their sampling system have not been published.
11 For measuring paniculate phase organic compounds, the denuder-based sampling system
12 represents an improvement over the filter/adsorbent collection method (Turpin et al., 1993).
13 Some researchers, however, have reported that denuder coatings themselves can introduce
14 contamination (Mukerjee et al., 1997), or the adsorbed species may be difficult to remove from
15 the coating (Eatough et al., 1993).
16 In a study conducted in southern California (Eatough et al., 1995), the Brigham Young
17 University Organic Sampling System (BOSS) (Eatough et al., 1993) was used for determining
18 POM composition, and a high-volume version (BIG BOSS) (flowrate 200 L/min) was utilized
19 for determining the particulate size distribution and the chemical composition of SVOC in fine
20 particles. The BOSS, a multi-channel diffusion denuder sampling system, consists of two
21 separate samplers (each operating at 35 L/min). The first sampler consists of a multi-parallel
22 plate diffusion denuder with charcoal-impregnated filter papers as the collection surfaces,
23 followed by a two-stage quartz filter pack, followed by a two-stage charcoal-impregnated filter
24 pack. The second sampler operating in parallel with the first consists of a two-stage quartz filter
25 pack, followed by the parallel plate denuder, followed by the two-stage charcoal-impregnated
26 filter pack. The filter samples collected by the BOSS sampler were analyzed by temperature-
27 programmed volatilization analysis. Eatough et al. (1995) also operated a two-stage quartz filter
28 pack alongside the BOSS sampler. The BIG BOSS system (Tang et al., 1994) consists of four
29 systems (each with a flowrate of 200 L/min). Particle size cuts of 2.5, 0.8, and 0.4 //m are
30 achieved by virtual impaction, and the sample subsequently flows through a denuder, then is
31 split, with the major flow (150 L/min) flowing through a quartz filter followed by an XAD-II
March 2001 2-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 bed. The minor flow is sampled through a quartz filter backed by a charcoal-impregnated filter
2 paper. The samples derived from the major flow (quartz filters and XAD-II traps) were extracted
3 with organic solvents and analyzed by gas chromatography and GC-mass spectroscopy. The
4 organic material lost from the particles was found to represent all classes of organic compounds.
5 Eatough et al. (1996) operated the BOSS sampler for a year at the IMPROVE site at
6 Canyonlands National Park, UT, alongside the IMPROVE monitor and alongside a separate
7 sampler consisting of a two-stage quartz filter pack. They found that concentrations of
8 particulate carbon determined from the quartz filter pack sampling system were low on average
9 by 39%, and this was attributed to volatilization losses of SVOC from the quartz filters.
10 In another study conducted with the BOSS in southern California, losses of 35% of the POM, on
11 average, were found and attributed to losses of the SVOC during sampling (Eatough et al.,
12 1995).
13 Ding et al. (1998a) developed a method for the determination of total n-nitroso compounds
14 in air samples, and used the method to examine organic compounds formed from NOX chemistry
15 in Provo, UT (Ding et al., 1998b). In their method, n-nitroso compounds are selectively
16 decomposed to yield nitric oxide, which is then detected using chemiluminescence. From the
17 samples from Provo, they found that the majority of the n-nitroso and nitrite organic compounds
18 that were present in fine particulate matter were semivolatile organic compounds that could be
19 evaporated from the particles during sampling. They found particulate n-nitroso compound
20 concentrations ranging between <1 and 3 nmoles/m3, and gas-phase n-nitroso compound
21 concentrations in the same range. Particulate organic nitrite concentrations were found in the
22 range of <1 to ~5 nmoles/m3, and gas-phase concentrations as high as 10 nmoles/m3 were found.
23 The PC-BOSS system of Eatough et al. (1999a) includes a virtual impactor upstream of the
24 denuder to improve the denuder collection efficiency by removing a majority of the gases from
25 the aerosol flow (i.e., gases and particles smaller than 0.1 //m are removed with the major flow of
26 the virtual impactor and the remaining aerosol enters the denuder). Particulate OC estimates are
27 corrected for particle losses of 46 to 48% in the inlet. The denuder consists of charcoal-
28 impregnated cellulose fiber filter material, and denuder collection efficiencies of greater than
29 98% are reported for organic gases that adsorb on quartz and charcoal-impregnated filters.
30 Turpin et al. (1993) developed a sampling system based upon a diffusion separator, which
31 corrects for the loss of semivolatile organic compounds during sampling by removal of most of
March 2001 2-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 the gas phase material from the particles in a diffusion separator sampling system. Unlike the
2 previously mentioned systems, wherein the particulate phase is measured directly, in the system
3 of Turpin et al., the gas-phase is measured directly. In the laminar flow system, ambient,
4 particle-laden air enters the sampler as an annular flow. Clean, particle-free air is pushed through
5 the core inlet of the separator. The clean air and ambient aerosol join downstream of the core
6 inlet section, and flow parallel to each other through the diffusion zone. Because of the much
7 higher diffusivities for gases compared to particles, the SVOC in the ambient air diffuse to the
8 clean, core flow. The aerosol exits the separator in the annular flow, and the core flow exiting
9 the separator now contains a known fraction of the ambient SVOC. Downstream of the diffusion
10 separator, the core exit flow goes into a PUF plug, where the SVOC is collected. The adsorbed
11 gas phase on the PUF plug is extracted with supercritical fluid CO2, and analyzed by gas
12 chromatography/mass-selective detection (GC/MSD). The gas-phase SVOC is thus determined.
13 Ultimately, to determine particulate phase SVOC concentrations, the total compound
14 concentration will also be measured, and the particulate phase obtained by difference. The
15 system was tested for the collection of PAH. The diffusional transport of gas-phase PAHs and
16 particle concentrations agreed well with theory. Breakthrough was problematic for low
17 molecular weight PAHs (MW < 160). Detection limits ranged from 20 to 50 pg of injected mass
18 for all PAHs.
19 Gundel et al. (1995) recently developed a technique for the direct determination of phase
20 distributions of semivolatile polcyclic aromatic hydrocarbons using annular denuder technology
21 instead of the different method. The method, called the integrated organic vapor/particle sampler
22 (IOVPS), uses a cyclone inlet with a D50 cutpoint of 2.5 //m at a sampling rate of 10 L/min. The
23 airstream then goes through two or three sandblasted glass annular denuders that are coated with
24 ground adsorbent resin material (XAD-4 was initially examined) that traps vapor-phase organics.
25 The airstream subsequently passes through a filter, followed by a backup denuder. The denuder
26 collection efficiency is high and compares well with predictions based on the diffusivity of the
27 compounds. The denuder can also be extracted to obtain gas-phase concentrations directly
28 (Gundel and Lane, 1999). Particle-phase PAHs are taken to be the sum of material on the filter
29 and XAD adsorbent downstream after correction for denuder collection efficiency. The IOVPS
30 was tested for sampling semivolatile PAH in laboratory indoor air, and environmental tobacco
31 smoke (ETS). After exposure, the denuders, filters, and sorbent traps were extracted with
March 2001 2-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 cyclohexane (Gundel et al., 1995) and analyzed for PAHs from naphthalene to chrysene using
2 dual-fluorescence detection (Mahanama et al., 1994). Recoveries from both denuders and filters
3 were approximately 70% for 30 samples. Detection limits (lower limits of detection, defined as
4 3 times the standard deviation of the blanks) for gas-phase SVOC PAHs ranged from 0.06 ng for
5 anthracene to 19 ng for 2-methylnaphthalene. The 95% confidence interval for reproduction of
6 an internal standard concentration was 6.5% of the mean value. Relative precision as determined
7 either from a propagation of errors analysis, or from the 95% confidence interval from replicate
8 analyses of standard reference material SRM 1649 (urban dust/organics) was 12% on average,
9 and ranged from 8% for naphthalene to 22% for fluorene. Sources of error included sampling
10 flow rate, internal standard concentration, and co-eluting peaks. Gundel and Lane (1999)
11 reported that roughly two-thirds of particulate PAH fluoranthene, pyrene, benz[a]anthracene, and
12 chrysene were found on the postfilter denuders, so that it is likely that considerable desorption
13 from the collected particles took place.
14 Solid adsorbent-based denuder systems have been investigated by other researchers, as
15 well. Bertoni et al. (1984) described the development of a charcoal-based denuder system, for
16 the collection of organic vapors. Risse et al. (1996) developed a diffusion denuder system to
17 sample aromatic hydrocarbons. In their system, denuder tubes with charcoal coating and
18 charcoal paper precede a filter pack for parti culate collection and an adsorption tube to capture
19 particle blow-off from the filter sample. Breakthrough curves for benzene, toluene, ortho-xylene,
20 and meta-xylene were developed for 60-, 90-, and 120-cm denuder tubes. The effects of relative
21 humidity on the adsorption capacities of the denuder system were examined, and it was found
22 that the capacity of the charcoal was not impacted significantly by increases in relative humidity.
23 The feasibility of outdoor air sampling with the system was demonstrated. Risse et al. (1996)
24 developed a diffusion denuder system for sampling aromatic hydrocarbons in which denuder
25 tubes were coated with charcoal.
26 Krieger and Kites (1992) designed a diffusion denuder system that uses capillary gas
27 chromatographic columns as the tubes for SVOC collection. The denuder was followed by a
28 filter to collect particles, which in turn was followed by a polyurethane foam (PUF) plug to
29 collect organic material volatilizing off the filter. Denuder samples were analyzed by liquid
30 solvent extraction (CH2C12) followed by GC-MS analysis. The PUF plugs and filters were
31 extracted with supercritical fluid extraction using supercritical N2O. Using this system, an indoor
March 2001 2-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 air sample was found to contain primarily chlorinated biphenyls, ranging from trichlorobiphenyls
2 (vapor pressures 10'3 - 10'4 Torr at 25 °C) to octachlorobiphenyls (10'6 - 10'7 Torr), which
3 demonstrated that the sampler collects compounds with a wide range of volatility. They also
4 found that on-line desorption is successful in maintaining good chromatographic peak shape and
5 resolution. The entire method, from sample collection to the end of the chromatographic
6 separation, took 2 h.
7 Organic acids in both the vapor and particulate phases may be important contributors to
8 ambient acidity, as well as representing an important fraction of organic particulate matter.
9 Lawrence and Koutrakis (1996a,b) used a modified Harvard/EPA annular denuder system
10 (HEADS) to sample both gas and particulate phase organic acids in Philadelphia, PA, in the
11 summer of 1992. The HEADS sampler inlet had a 2. l-//m cutpoint impactor (at 10 1pm),
12 followed by two denuder tubes, and finally a filter pack with a Teflon filter. The first denuder
13 tube was coated with KOH to trap gas phase organic acids. The second denuder tube was coated
14 with citric acid to remove ammonia and thus to avoid neutralizing particle phase acids collected
15 on the filter. The KOH-coated denuder tube was reported to collect gas phase formic and acetic
16 acids at better than 98.5% efficiency, and with precisions of 5% or better (Lawrence and
17 Koutrakis, 1994). It was noted that for future field measurements of particulate organic acids,
18 a Na2CO3-coated filter should be deployed downstream of the Teflon filter to trap organic acids
19 that may evaporate from the Teflon filter during sampling.
20
21 Role of the Collection Media
22 The role of the collection media was recently examined in a study conducted in Seattle
23 (Lewtas et al., 2001). In that study, the influence of denuder sampling methods and filter
24 collection media on the measurement of SVOC associated with PM2 5 was evaluated. Activated
25 carbon and XAD collection media were used in diffusion denuders and impregnated back-up
26 filters in two different samplers, the VAPS and the PC-BOSS. XAD-coated glass annular
27 denuders and charcoal-impregnated cellulose fiber (GIF) filter denuders also were used. GIF
28 filters also were compared to XAD-coated quartz filters as backup filter collection media.
29 Lewtas et al. (2001) found that the two denuder types resulted in equivalent measurement of
30 particulate organic carbon and particle mass. The carbon coated denuders in the BOSS sampler
31 were more efficient than the XAD coated denuders for the collection of more volatile carbon.
March 2001 2-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lewtas et al. (2001) concluded that the more volatile carbon that is collected in the carbon coated
2 BOSS denuder does not contribute substantially to the particle mass or to the SVOC measured as
3 OC on quartz filters. However, the more volatile carbon otherwise would be captured in carbon
4 impregnated filters placed behind quartz filters, so that, in the XAD denuder configuration, the
5 volatile carbon would result in a higher OC concentration and overestimation of the SVOC.
6 Some of the recent research in denuder technology also has focused on reduction in the size
7 of the denuder, optimization of the residence time in the denuder, understanding the effect of
8 diffusion denuders on the positive quartz filter artifact, identifying changes in chemical
9 composition that occur during sampling, determining the effects of changes in temperature and
10 relative humidity, and identifying possible losses by absorption in coatings.
11
12 Reducing the Size of Denuders
13 The typical denuder configuration is an annular diffusion denuder tube of significant length
14 (e.g., 26.5 cm for 10 L/min, Koutrakis et al., 1988a). A more compact design based on a
15 honeycomb configuration was shown to significantly increase the capacity (Koutrakis et al.,
16 1993). However, in intercomparisons with an annular denuder/filter pack system (Koutrakis
17 et al., 1988a), significant losses of ammonia and nitric acid were observed for the honeycomb
18 configuration, and attributed to the large inlet surface area and long sample residence time of the
19 honeycomb design, relative to the annular denuder system. Sioutas et al. (1996a) subsequently
20 designed a modified glass honeycomb denuder/filter pack sampler (HDS) with an inlet that
21 minimizes vapor losses on the inlet surfaces. The modified HDS has reduced inlet surfaces and
22 decreased residence time for sampled gases (NH3 and HNO3) compared to its predecessor
23 (Sioutas et al., 1994). Sioutas et al. (1996b) tested various inlet materials (glass, PFA, and
24 PTFE) in laboratory tests and found that a PTFE Teflon coated inlet minimized loss of sampled
25 gases (1 to 8% losses of HNO3 observed, and -4 to 2% losses of NH3 observed). The highest
26 inlet losses were observed for HNO3 lost to PFA surfaces (14 to 25%). The modified HDS was
27 tested in laboratory and field tests and found to agree within 10% with the annular denuder
28 system.
29
30
31
March 2001 2-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 Residence Time in the Denuder
2 The efficiency of a diffusion denuder sampler for the removal of gas phase material can be
3 improved by increasing the residence time of the sampled aerosol in the denuder. However, the
4 residence time can only be increased within certain limits. Because the diffusion denuder
5 reduces the concentration of gas-phase semivolatile organic material, semivolatile organic matter
6 present in the particles passing through the denuder will be in a thermodynamically unstable
7 environment and will tend to outgas SVOC during passage through the denuder. The residence
8 time of the aerosol in the denuder, therefore, should be short enough to prevent significant loss of
9 particulate phase SVOC to the denuder. Various studies have suggested that the residence time
10 in the denuder should be less than about 2 s (Gundel and Lane, 1999; Kamens and Coe, 1997;
11 Kamens et al., 1995). The residence times in the various denuder designs described by Gundel
12 and Lane (1999) are from 1.5 to 0.2 s. The equilibria and evaporation rates are not as well
13 understood for organic components as they are for NH4NO3 (Zhang and McMurry, 1987, 1992;
14 Hering and Cass, 1999).
15
16 Effect of Diffusion Denuders on the Positive Quartz Filter Artifact
17 To account for the volatilization losses of semivolatile organic compounds, Turpin et al.
18 (1994) recommended that a quartz filter be placed behind a Teflon filter in a parallel sampler.
19 Addition of a vapor trap (e.g., polyurethane foam plug) downstream of the filter also was
20 suggested as a method to collect semivolatile organic compounds. However, it was noted that
21 addition of some type of trap behind the Teflon filter collected both vapor phase organics as well
22 as "blow-off from the Teflon filter (i.e., material vaporized from particles collected on Teflon
23 filter [Van Vaeck et al., 1984]). Kim et al. (2000) used a quartz filter behind a Teflon filter
24 recently to account for positive organic artifacts in the South Coast Air Basin. They found that,
25 on an annual average basis, 30% of the PM25 organic carbon concentration resulted from positive
26 artifacts.
27 The adsorption of organic compounds by a second quartz filter has been shown to be
28 reduced, but not eliminated, in samples collected in the Los Angeles Basin, if a multi-channel
29 diffusion denuder with quartz filter material as the denuder collection surface preceded the quartz
30 filters (Fitz, 1990). This artifact can be further reduced by the use of activated charcoal as the
31 denuder surface and use of a particle concentrator to reduce the amount of gas phase organic
March 2001 2-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 compounds relative to condensed phase organic compounds (Cui et al., 1998, 1997; Eatough,
2 1999). Recent experiments (Cotham and Bidleman, 1992; Cui et al., 1998; Eatough et al., 1995,
3 1996) have shown that the quartz filter artifact can result both from the collection of gas phase
4 organic compounds and from the collection of semivolatile organic compounds lost from
5 particles during sampling. Thus, results available to date suggest that both a "positive" and a
6 "negative" artifact can be present in the determination of particulate phase organic compounds
7 using two tandem quartz filters.
8 The importance of the adsorption of organic vapors on filters or PM, relative to the
9 volatilization of organic compounds from PM collected on a filter, continues to be a topic of
10 active debate. The relative importance of positive and negative artifacts will be different for
11 denuded and undenuded filters; will depend on face velocity, sample loading, and the vapor
12 pressures of the compounds of interest; and may vary with season and location because of
13 variations in the composition of volatile and semivolatile organic material. Evidence exists for
14 substantial positive and negative artifacts in the collection of organic PM.
15 Undenuded quartz-quartz and Teflon-quartz back-up filters have been reported to collect
16 10 to 50% of the organic mass found on quartz front filters that remove parti culate matter with
17 essentially 100% efficiency (Turpin, et al., 2000). Larger percentages were found for samples
18 with shorter collection times and for cleaner locations. Kirchstetter et al. (2000) and Turpin et al.
19 (2000) argue that the quantity of organic material on a quartz back-up filter provides an estimate
20 of the positive artifact (i.e., adsorbed organic vapors), but provides no information about the
21 negative artifact (i.e., volatilized particulate organics). This argument is based on profiles of
22 thermal carbon analyses (i.e., plots of evolved carbon with temperature created during Evolved
23 Gas Analysis [EGA]) and the following argument. Material volatilized from the collected
24 particles will not add significantly to the loading on the quartz backup filter unless the ratio of the
25 mass of semivolatile vapor to the mass of semivolatile condensed phase material is low and the
26 rate of volatilization of the condensed phase semivolatile material is great enough to significantly
27 increase the concentration of the semivolatile vapor passing through the back-up filter (Zhang
28 and McMurry, 1987).
29 A net positive artifact for total particulate organic carbon was reported by Novakov et al.
30 (1997), whose filter-based aircraft measurements had carbon loadings that exceeded the total
31 aerosol mass. Novakov compared estimates of adsorption based on examination of EGA
March 2001 2-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 thermograms and estimates of adsorption obtained from quartz-quartz back-up filters and
2 concluded that, if anything, the quartz-quartz back-up filter underestimates the positive sampling
3 artifact. Also, both McDow and Huntzicker (1990) and Turpin et al. (1994) observed that
4 subtraction of the Teflon-quartz backup filter (an estimate of adsorbed organic gases) from the
5 quartz front filter loading removed the face velocity dependence of the particulate organic carbon
6 concentrations obtained at face velocities of 20, 40 and 80 cm/s. Kirchstetter et al. (2000)
7 reported that the organic carbon content of a denuded quartz filter collected in Berkeley, CA was
8 comparable to the carbon content of a concurrently-collected undenuded quartz filter after
9 subtraction of the matching Teflon-quartz backup (i.e., after correction for the positive artifact).
10 As a result, they concluded that volatile losses must not be important for this sample. (Some
11 denuder breakthrough was noted in this study.)
12 Evidence of a net negative artifact is provided by Lewtas et al. (2001), who emphasized
13 that if parti culate OC had been measured on a denuded quartz filter without an adsorbent
14 downstream, the negative bias would be large. Their data showed that the sum of a denuded
15 quartz filter and absorbent downstream (average = 9.1 //g/m3) was greater than a collocated
16 undenuded quartz filter (average = 7.7 //g/m3) in a PC-BOSS sampler after correction for losses
17 (46 to 48%) in the virtual impactor inlet. A net negative artifact for total parti culate OC has been
18 reported by Eatough and colleagues in a number of studies (e.g., Cui et al., 1998; Eatough, 1999).
19
20 Changes in Chemical Composition During Sampling
21 The use of sampling systems designed to correctly identify the atmospheric gas and
22 particulate phase distributions of collected organic material has been outlined above.
23 An additional sampling artifact that has received little consideration in the collection of
24 atmospheric sampling is the potential alteration of organic compounds as a result of the sampling
25 process. These alterations appear to result from the movement of ambient air containing
26 oxidants and other reactive compounds past the collected particles. The addition of NO2
27 (
-------
1 mid-volume sampler (Ding et al., 1998a,b). Similar results have been obtained for the exposure
2 of a deuterated amine on a filter to NOX (Pellizzari and Krost, 1984). When Tenax columns
3 spiked with deuterated styrene and cyclohexene were exposed to ppm concentrations of ozone or
4 halogens, oxygenated and halogenated compounds were shown to be formed (Pellizzari and
5 Krost, 1984). Similar oxidation of aldehydes and PAN during sampling has been observed
6 (Grosjean and Parmar, 1990). Collected PAH compounds can be oxygenated or nitrated on a
7 filter (Davis et al., 1987; Lindskog and Brorstrom-Lunden, 1987) but 1-nitropyrene has been
8 shown to be resistant to additional nitration (Grosjean, 1983). These various chemical
9 transformations of collected organic compounds can be eliminated by removal of the gas phase
10 oxidants, NOX, HNO3, etc., prior to collection of the particles (Ding, 1998a,b; Grosjean and
11 Parmar, 1990; Parmar and Grosjean, 1990; Pellizzari and Krost, 1984; Williams and Grosjean,
12 1990). The BOSS denuder should be effective in eliminating most of the chemical
13 transformation artifacts, because reactive gases are removed by the charcoal denuder that
14 precedes the particle collection filter.
15
16 Temperature and Relative Humidity Effects
17 The problems of sampling artifacts associated with SVOC adsorption and evaporation are
18 compounded by temperature and relative humidity effects (Pankow and Bidleman, 1991; Pankow
19 et al., 1993; Falconer et al., 1995; Goss and Eisenreich, 1997). Effects of temperature on the
20 partitioning of PAH were examined by Yamasaki et al. (1982), who found that the partition
21 coefficient (PAHvapor/PAHpart) was inversely related to temperature and could be described using
22 the Langmuir adsorption concept. The dissociation of ammonium nitrate aerosol is also a
23 function of temperature. Bunz et al. (1996) examined the dissociation and subsequent
24 redistribution of NH4NO3 within a bimodal distribution, using a nine-stage low-pressure Berner
25 impactor followed by analysis by ion chromatography and found a strong temperature
26 dependency on the redistribution. Bunz et al. (1996) found that at lower temperatures (below
27 10 °C), there was little change in the aerosol size distribution. At temperatures between 25 and
28 45 °C, however, the lifetime of NH4NO3 particles decreases by more than a factor of 10, and size
29 redistribution, as measured by average ending particle diameter, increased more for higher
30 temperatures than for lower temperatures.
March 2001 2-63 DRAFT-DO NOT QUOTE OR CITE
-------
1 The effects of relative humidity on the sorption of SVOC on particles are not well
2 understood. In a series of laboratory experiments, Goss and Eisenreich (1997) examined the
3 sorption of both nonpolar (hydrocarbons and chlorinated hydrocarbons) and polar (ethyl ether
4 and acetone) volatile organic compounds (VOC) onto combustion soot particles as a function of
5 temperature and relative humidity. The soot particles used in their experiments were collected
6 from oil furnaces and contained 60% (w/w) iron sulfate (water-soluble fraction) and 9% (w/w)
7 elemental and organic carbon. The carbon and sulfate contents of their paniculate matter are
8 comparable to the chemical composition of ambient fine particles. They found that, for all
9 compounds, the sorption of VOC onto soot particles decreased with increasing relative humidity
10 over the range of 10 to 95%. They also observed hysteresis in the relative humidity dependency,
11 with sorption coefficients at a given relative humidity higher when the RH is being increased
12 than when the RH is being decreased. The sorption coefficients were fit with an exponential
13 function to the RH so that the slope of the regression line would provide a measure of the
14 influence of relative humidity. Based on the magnitude of the slope, they concluded that the
15 RH-dependency of sorption was stronger for water-soluble organic compounds.
16 In another study by Jang and Kamens (1998), humidity effects on gas-particle partitioning
17 of SVOC were examined using outdoor environmental chambers and the experimentally
18 determined partitioning coefficients were compared to theoretical values. They examined the
19 partitioning of SVOC onto wood soot, diesel soot, and secondary aerosols and concluded that
20 "the humidity effect on partitioning was most significant for hydrophobic compounds adsorbing
21 onto polar aerosols." Although these two studies seem to be contradictory, on closer
22 examination, it is difficult to compare the two studies for several reasons. The experiments
23 conducted by Jang and Kamens (1998) were conducted in outdoor chambers at ambient
24 temperatures and humidities. Their model was for absorptive partitioning of SVOC on
25 liquid-like atmospheric particulate matter. In contrast, the results of Goss and Eisenreich (1997)
26 were obtained from a gas chromatographic system operated at 70 °C higher than ambient
27 conditions. The model of Goss and Eisenreich (1997) was for adsorptive partitioning of VOC on
28 solid-like atmospheric parti culate matter. In the study of Jang and Kamens (1998), calculated
29 theoretical values for water activity coefficients for diesel soot were based on an inorganic salt
30 content of 1 to 2%, whereas the combustion particles studied by Goss and Eisenreich (1997)
31 contained 60% water-soluble, inorganic salt content. Jang and Kamens (1998) obtained their
March 2001 2-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 diesel soot from their outdoor chamber, extracted it with organic solvent (mixtures of hexane and
2 methylene chloride), and measured the organic fraction. The resulting salt content of 2% of the
3 particulate matter studied in Jang and Kamens (1998) is enough to affect water uptake but
4 presumably not to affect the sorption partitioning of organics.
5
6 Impactor Coatings
1 Impactors are used as a means to achieve a size cutpoint and as particle collection surfaces.
8 Particles collected on impactors are exposed to smaller pressure drops than filter-collected
9 particles, making them less susceptible to volatile losses (Zhang and McMurry, 1987). However,
10 size resolution can be affected by bounce when samples are collected at low humidities (Stein
11 et al., 1994). There are other sources of error inherent in some of the currently acceptable
12 practices that could potentially affect particulate mass concentration measurements and that will
13 surely become even more important as more emphasis is placed on chemical speciation. Allen
14 et al. (1999a) reported that the practice of greasing impaction substrates may introduce an artifact
15 from the absorption of semivolatile species from the gas phase by the grease, which could
16 artificially increase the amount of PAHs and other organic compounds attributed to the aerosol.
17 Allen et al. (1999a) offer several criteria to ensure that this absorption artifact is negligible,
18 including selecting impaction oils in which analytes of interest are negligibly soluble and
19 ensuring that species do not have time to equilibrate between the vapor and oil phases (criterion
20 is met for nonvolatile species). They recommend using oiled impaction substrates only if the
21 absorption artifact is negligible as determined from these criteria. Application of greases and
22 impaction oils for preventing or reducing bounce when sampling with impactors is not suitable
23 for carbon analysis because the greases contain carbon (Vasilou et al., 1999).
24 Kavouras and Koutrakis (2000) investigated the use of polyurethane foam (PUF) as a
25 substrate for conventional inertial impactors. The PUF impactor substrate is not rigid like the
26 traditional impactor substrate so particle bounce and reentrainment artifacts are reduced
27 significantly. Kavouras and Koutrakis found that the PUF impaction substrate resulted in a much
28 smaller d50 at the same flow rate and Reynolds number. Moreover, the lower dso was obtained at
29 a lower pressure drop than with the conventional substrate, which could lead to a reduction of
30 artifact vaporization of semivolatile components.
31
March 2001 2-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.3.4 Particle-Bound Water
2 It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
3 compounds. However, for many measurements of suspended particle mass, it is desirable to
4 remove the particle-bound water before determining the mass. In other situations it may be
5 important to know how much of the suspended particle's mass or volume results from parti cle-
6 bound water. The water content of PM is significant and highly variable. Moreover, there is
7 significant hysteresis in the water adsorption-desorption pathways (Seinfeld and Pandis, 1998),
8 further complicating the mass measurement. Figures 2-8 and 2-9 show the change in diameter of
9 sulfate particles as a function of relative humidity. Figure 2-8 shows the difference between
10 deliquescence and crystallization points.
11 Pilinis et al. (1989) calculated the water content of atmospheric particulate matter above
12 and below the deliquescent point. They predicted that aerosol water content is strongly
13 dependent on composition, and concluded from their calculations that liquid water could
14 represent a significant mass fraction of aerosol concentration at relative humidities above 60%.
15 Since then, a few researchers have attempted to measure the water content of atmospheric
16 aerosol. Most techniques have focused on tracking the particle mass as the relative humidity is
17 changed, and are still in the development phase. There have been only a few demonstrations
18 using actual ambient aerosol, to date. Of interest, in particular, is the development of the
19 Tandem Differential Mobility Analyzer (TDMA) and its applications in investigations of the
20 effects of relative humidity on particle growth.
21 Lee et al. (1997) examined the influence of relative humidity on the size of atmospheric
22 aerosol using a TDMA coupled with a scanning mobility particle sizer (SMPS). They reported
23 that the use of the TDMA/SMPS system allowed for the abrupt size changes of aerosols at the
24 deliquescence point to be observed precisely. They also reported that, at relative humidities
25 between 81 and 89%, the water content of ammonium sulfate aerosols (by mass) ranged from
26 47 to 66%.
27 Andrews and Larson (1993) investigated the interactions of single aerosol particles coated
28 with an organic film with a humid environment. Using an electrodynamic balance, they
29 conducted laboratory experiments in which sodium chloride and carbon black particles were
30 coated with individual organic surfactants, intended to simulate the surface-active, organic films
31 that many atmospheric aerosol particles may exhibit, and their water sorption curves examined.
March 2001 2-66 DRAFT-DO NOT QUOTE OR CITE
-------
1 Their results showed that when ordinarily hydrophobic carbon black particles were coated with
2 an organic surfactant, they sorbed significant amounts of water (20 to 40% of the dry mass of the
3 particle).
4 Liang and Chan (1997) developed a fast technique using the electrodynamic balance to
5 measure the water activity of atmospheric aerosols. In their technique, the mass of a levitated
6 particle is determined as the particle either evaporates or grows in response to a step change in
7 the relative humidity. Their technique was demonstrated using laboratory experiments with
8 NaCl, (NH4)2SO4, NaNO3, and (NH4)2SO4/NH4NO3 solutions. They concluded that one of the
9 advantages of their fast method is the ability to measure the water activity of aerosols containing
10 volatile solutes such as ammonium chloride and some organics.
11 Mclnnes et al. (1996) measured aerosol mass concentration, ionic composition, and
12 associated water mass of marine aerosol over the remote Pacific Ocean. The mass of
13 particle-bound water was determined by taking the difference between the mass obtained at 48%
14 RH and at 19% RH, assuming the aerosol particles were dry at 19% RH. Based on a comparison
15 of the remote Pacific aerosol to aerosol collected at a site at the marine/continental interface of
16 the Washington coast, the amount of water associated with the aerosol was observed to be a
17 function of the ammonium to sulfate ratio. They found that the amount of water associated with
18 the submicrometer aerosol comprised 29% of the total aerosol mass collected at 47% RH and
19 9% of the total mass at 3 5% RH.
20 Ohta et al. (1998) characterized the chemical composition of atmospheric fine particles
21 (D50 = 2 //m) in Sapporo, Japan, and as part of their measurements, determined the water
22 content using the Karl Fischer method (Meyer and Boyd, 1959). After exposing a Teflon filter, a
23 portion of the filter was equilibrated at 30% RH for 24 h. Then the filter piece was placed in a
24 water evaporator heated at 150 °C, vaporizing the particle-bound water. The vapor evolved was
25 analyzed for water in an aqua-counter where it was titrated coulometrically in Karl Fischer
26 reagent solution (containing iodine, sulfur, and methanol). The accuracy of the aqua-counter is
27 ±1 mg. Using this technique, they determined that the water content of the particles ranged from
28 0.4 to 3.2% of the total particulate mass (at RH < 30%). This represents a smaller portion of
29 water compared to their previous reported values (Ohta and Okita, 1990) that were determined by
30 calculation at RH of 50%.
March 2001 2-67 DRAFT-DO NOT QUOTE OR CITE
-------
1 Speer et al. (1997) developed an aerosol liquid water content analyzer (LWCA), in which
2 aerosol samples are collected on PTFE filters, and then placed in a closed chamber in which the
3 relative humidity is closely controlled. The aerosol mass is monitored using a beta-gauge, first as
4 the relative humidity is increased from low RH to high RH, and then as the RH is decreased
5 again. They demonstrated the LWCA on laboratory-generated aerosol and on an ambient PM2 5
6 sample collected in Research Triangle Park, NC. The ambient aerosol sample was also analyzed
7 for chemical constituents. It is interesting to note that, although their laboratory-generated
8 (NH4)2SO4 aerosol demonstrated a sharp deliquescent point, their atmospheric aerosol, which
9 was essentially (NH4)2SO4, did not show a sharp deliquescent point.
10 Hygroscopic properties of aerosols have been studied from the viewpoint of their ability to
11 act as condensation nuclei. The hygroscopic properties of fresh and aged carbon and diesel soot
12 particles were examined by Weingartner et al. (1997) who found that fresh, submicron-size
13 particles tended to shrink with increasing relative humidity, because of a crystalline structural
14 change. Lammel and Novakov (1995) found, in laboratory studies, that the hygroscopicity of
15 soot particles could be increased by chemical modification, and that the cloud condensation
16 nucleation characteristics of diesel soot were similar to those of wood smoke aerosol.
17 The results of several of the above studies, in which aerosol water content as a function of
18 relative humidity was determined, are summarized in Figure 2-13. In this figure, the results of
19 Lee et al. (1997), Mclnnes et al. (1996), and Ohta et al. (1998) are included. Relative humidity
20 ranged from 9%, at which the aerosol water content was assumed to be zero (Mclnnes et al.,
21 1996), to 89%, at which the aerosol water content was determined to be 66% by mass (Lee et al.,
22 1997). Koutrakis et al. (1989) and Koutrakis and Kelly (1993) also have reported field
23 measurements of the equilibrium size of atmospheric sulfate particles as a function of relative
24 humidity and acidity.
25 The effects of relative humidity on particle growth were also examined in several studies.
26 Fang et al. (1991) investigated the effects of flow-induced relative humidity (RH) changes on
27 particle cut sizes for aqueous sulfuric acid particles in a multi-nozzle micro-orifice uniform
28 deposit impactor (MOUDI). Laboratory experiments were conducted in which polydisperse
29 sulfuric acid aerosols were generated and the RH was adjusted. The aerosols were analyzed by a
30 differential mobility analyzer. Fang et al. (1991) observed that for inlet RH less than 80%, the
31 cut sizes for the sulfuric acid aerosols were within 5% of that for nonhygroscopic particles except
March 2001 2-68 DRAFT-DO NOT QUOTE OR CITE
-------
.22
T3
in
i
0)
o
CD
Q.
0
Q
c/>
CO
CD
CD
o
CD
Q.
o£
90-
80-
70-
60-
50-
40-
30-
20-
10-
n-
• Mclnnes
• Lee et al
A Ohtaetc
B
•
•
A
•,,..,,.,
etal., 1996
., 1997
ll., 1998
g
0 10 20 30 40 50 60 70 80 90 100
Relative Humidity, %
Figure 2-13. Aerosol water content expressed as a mass percentage, as a function of
relative humidity.
Source: Mclnnes et al. (1996); Lee et al. (1997); Ohta et al. (1998).
1 at the stage for which the cut size was 0.047 //m, where the cut size was 10.7% larger than the
2 nonhygroscopic particle cut size. They concluded that flow-induced RH changes would have
3 only a modest effect on MOUDI cut sizes at RH < 80%.
4 Hitzenberger et al. (1997) collected atmospheric aerosol in the size range of 0.06 to 15 //m
5 in Vienna, Austria, using a nine-stage cascade impactor and measured the humidity-dependent
6 water uptake when the individual impaction foils were exposed to high RH. They observed
7 particle growth with varying growth patterns. Calculated extinction coefficients and single
8 scattering albedo increased with humidity.
9 Hygroscopic properties, along with mixing characteristics, of submicrometer particles
10 sampled in Los Angeles, CA, during the summer of 1987 SCAQS study and at the Grand
11 Canyon, AZ, during the 1990 Navajo Generating Station Visibility Study were reported by Zhang
March 2001
2-69
DRAFT-DO NOT QUOTE OR CITE
-------
1 et al. (1993). They used a tandem differential mobility analyzer (TDMA) (McMurry and
2 Stolzenburg, 1989) to measure the hygroscopic properties for particles in the 0.05- to 0.5-//m
3 range. In their experimental technique, monodisperse particles of a known size are selected from
4 the atmospheric aerosol with the first DMA. Then, the relative humidity of the monodisperse
5 aerosol is adjusted and the new particle size distribution is measured with the second DMA.
6 At both sites, they observed that monodisperse particles could be classified according to "more"
7 hygroscopic and "less" hygroscopic. Aerosol behavior observed at the two sites differed
8 markedly. Within the experimental uncertainty (±2%) the "less" hygroscopic particles sampled
9 in Los Angeles did not grow when the RH was increased to 90%, whereas at the Grand Canyon,
10 the growth of the "less" hygroscopic particles varied from day to day, but ranged from near 0 to
11 40% when the RH was increased to 90%. The growth of the "more" hygroscopic particles in
12 Los Angeles was dependent on particles size (15% at 0.05 //m to 60% at 0.5 // m), whereas at the
13 Grand Canyon, the "more" hygroscopic particles grew by about 50%, with the growth not
14 varying significantly with particle size. By comparison of the TDMA data to impactor data,
15 Zhang et al. (1993) surmised that the more hygroscopic particles contained more sulfates and
16 nitrates, while the less hygroscopic particles contained more carbon and crustal components.
17 Although most of the work to date on the hygroscopic properties of atmospheric aerosols
18 has focused on the inorganic fraction, the determination of the contribution of particle-bound
19 water to atmospheric particulate mass is greatly complicated by the presence of organics. The
20 effect of RH on adsorption of semivolatile organic compounds is discussed elsewhere in this
21 chapter. Saxena et al. (1995) observed that particulate organic compounds also can affect the
22 hygroscopic behavior of atmospheric particles. They idealized the organic component of aerosol
23 as containing a hydrophobic fraction (high-molecular weight alkanes, alkanoic acids, alkenoic
24 acids, aldehydes, and ketones) and a hydrophilic fraction (e.g., lower molecular weight
25 carboxylic acids, dicarboxylic acids, alcohols, aldehydes, etc.) that would be likely to absorb
26 water. They then analyzed data from a tandem differential mobility analyzer in conjunction with
27 particle composition observations from an urban site (Claremont, CA) and from a nonurban site
28 (Grand Canyon) to test the hypothesis that, by adding particulate organics to an inorganic aerosol,
29 the amount of water absorbed would be affected, and the effect could be positive or negative,
30 depending on the nature of the organics added. They further presumed that the particulate
31 organic matter in nonurban areas would be predominantly secondary and thus hydrophilic,
March 2001 2-70 DRAFT-DO NOT QUOTE OR CITE
-------
1 compared to the urban aerosol that was presumed to be derived from primary emissions and thus
2 hydrophobic in nature. Their observations were consistent with their hypothesis, in that at the
3 Grand Canyon, the presence of organics tended to increase the water uptake by aerosols, whereas
4 at the Los Angeles site, the presence of organics tended to decrease water uptake.
5 Nonequilibrium issues may be important for the TDMA, as well as for other methods of
6 measuring water content. Although approach to equilibrium when the RH is increased is
7 expected to be rapid for pure salts, it may be much slower for aerosols containing a complex mix
8 of components (Saxena et al., 1995). For example, if an aerosol contains an organic film or
9 coating, that film may impede the transport of water across the particle surface, thus increasing
10 the time required for equilibrium (Saxena et al., 1995). Insufficient time to achieve equilibrium
11 in the TDMA could result in underestimation of the water content.
12
13 2.2.4 U. S. Environmental Protection Agency Monitoring Programs
14 2.2.4.1 The Federal Reference Methods for Equilibrated Mass
15 Federal Reference Methods (FRM) have been specified for measuring PM10 (Code of
16 Federal Regulations, 1991a,b) and for measuring PM2 5 (Code of Federal Regulations, 1999a).
17 The FRM for PM10 has been discussed in previous PM AQCD's and will only be briefly
18 reviewed. The PM10 FRM defines performance specifications for samplers in which particles are
19 inertially separated with a penetration efficiency of 50% at an aerodynamic diameter of
20 10 ± 0.5 //m. The collection efficiency increases to ~ 100% for smaller particles and drops to
21 ~Q% for larger particles. Particles are collected on filters and mass concentrations are
22 determined gravimetrically. Instrument manufactures are required to demonstrate through field
23 tests a measurement precision for 24-h samples of ± 5 //g/m3 for PM10 concentrations below
24 80 //g/m3 and 7% above this value.
25 As opposed to the performance-based FRM standard for PM10, the new FRM for PM2 5
26 specifies certain details of the sampler design, as well as of sample handling and analysis,
27 whereas other aspects have performance specifications. The PM25 FRM sampler consists of a
28 PM10 inlet/impactor, a PM2 5 impactor with an oil-soaked impaction substrate to remove particles
29 larger than 2.5 //m, and a 47-//m polytetrafluoroethylene (PTFE) filter with a particle collection
30 efficiency greater than 99.7%. The sample duration is 24 h, during which the sample temperature
31 is not to exceed ambient temperatures by more than 5 °C. After collection, samples are
March 2001 2-71 DRAFT-DO NOT QUOTE OR CITE
-------
1 equilibrated for 24-h at temperatures in the range of 20 to 23 °C (± 2 °C) and at relative
2 humidities in the range of 30 to 40% (± 5%). The equilibration tends to reduce particle-bound
3 water and stabilizes the filter plus sample weight. Filters are weighed before and after sampling
4 under the same temperature and relative humidity conditions. For sampling conducted at
5 ambient relative humidity less than 30%, mass measurements at relative humidities down to 20%
6 are permissible (Code of Federal Regulations, 1999a).
7 The FRM also allows for Class I, II, and HI equivalent methods for PM2 5 (Code of Federal
8 Regulations, 1999b). Class I equivalent methods use samplers with relatively small deviations
9 from the sampler described in the FRM. Class II equivalent methods include "all other PM2 5
10 methods that are based upon 24-h integrated filter samplers that are subjected to subsequent
11 moisture equilibration and gravimetric mass analysis." Class HI equivalent methods include
12 non-filter-based methods such as beta attenuation, harmonic oscillating elements, or
13 nephelometry (McMurry, 2000).
14 The strength of the PM25 FRM is that specification of all details of the sampler design
15 ensures that measurements at all locations, if done properly, should be comparable. For example,
16 the FRM requires maintenance because the oil-soaked impaction substrate could otherwise
17 become loaded with coarse particles. Failure to do so could lead to coarse particle bounce, thus
18 artificially increasing the measured fine particle concentrations. Moreover, the specification of a
19 PM10 inlet requires the oil-soaked impaction substrate to collect particles between 2.5 and 10 //m.
20 The implication is that, during sampling periods of high coarse PM concentrations, the impaction
21 substrate could become overloaded, leading to particle bounce. If an inlet with a cutpoint
22 diameter smaller than 10 //m were specified, coarse particle bounce could potentially be reduced,
23 and perhaps the maintenance frequency could be reduced (McMurry, 2000).
24 Since the implementation of the PM10 standard in 1987 (Federal Register, 1987)
25 considerable information has accumulated on the factors that affect the quality of the data
26 gathered from the EPA reference method for PM10. These include inlet losses of coarse fraction
27 particles (e.g., Anand et al., 1992); biases in concentrations due to differences between samplers
28 in large particle cutpoints that are within the EPA's specified acceptable tolerances (Ranade
29 et al., 1990); and particle bounce tolerances and reentrainment leading to as much as 30% errors
30 (Wang and John, 1988). The sampling issues associated with cutpoint tolerances are predictable,
31 and the particle bounce and reentrainment problems have since been dealt with voluntarily by the
March 2001 2-72 DRAFT-DO NOT QUOTE OR CITE
-------
1 manufacturers by recommending operational procedures including oiling of impact surfaces and
2 regular cleaning. The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996)
3 concluded that the PM10 sampling systems can be designed such that concentration measurements
4 are precise to ±10%. For PM25, cutpoint tolerances are not expected to affect the mass
5 concentration as much as for PM10, since the 2.5-//m cutpoint generally occurs near a minimum
6 in the mass distribution (e.g., Figure 2-5).
7 The PM2 5 mass concentration will be affected, on the other hand, by other sampling issues
8 mentioned but not discussed extensively in the previous 1996 PM AQCD (U.S. Environmental
9 Protection Agency, 1996). These issues have been discussed earlier in this chapter and include
10 gas/particle and particle/substrate interactions for sulfates and nitrates (e.g., Appel et al., 1984),
11 volatilization losses of nitrates (Zhang and McMurry, 1992), semivolatile organic compound
12 (SVOC) artifacts (e.g., Eatough et al., 1993), and relative humidity effects (e.g., Keeler et al.,
13 1988).
14 Several studies now have been reported, in which the FRM was collocated with other PM2 5
15 samplers in intercomparison studies. During the Aerosol Research and Inhalation Epidemiology
16 Study (ARIES) several PM2 5 samplers were collocated at a mixed industrial-residential site near
17 Atlanta, GA (Van Loy et al., 2000). These samplers included a standard PM25 FRM, a TEOM
18 with Nafion drier, a particulate composition monitor (PCM) (Atmospheric Research and
19 Analysis, Gary, NC), a high-volume carbon sampler operated by the Desert Research Institute, a
20 HEADS sampler, and a dichotomous sampler for coarse PM. The PCM sampler has three
21 channels, all of which have PM10 cyclone inlets. The first two channels both have two denuders
22 preceding a 2.5-//m WINS impact and filter packs. The first denuder is coated with sodium
23 carbonate to remove acid gases, and the second is coated with citric acid to remove ammonia.
24 The third channel has a carbon coated parallel-plate denuder preceding the WINS impactor.
25 Measurements of 24-h mass from the FRM, PCM, and TEOM samplers, as well as reconstructed
26 PM2 5 mass (RPM) were compared for a 12-mo period. The slopes for the TEOM-FRM,
27 PCM-FRM, and RPM-FRM correlations were 1.01, 0.94, and 0.91, respectively, whereas the
28 y-intercepts for each were 0.68, 0.04, and 0.98. Particulate sulfate measurements on the FRM
29 Teflon filter, the PCM Teflon filter, and PCM Nylon filter were nearly identical. Nitrate results
30 from the three filters were much less consistent, with the FRM collecting substantially less nitrate
31 than that collected on either the denuded nylon filter or a denuder followed by a Teflon-nylon
March 2001 2-73 DRAFT-DO NOT QUOTE OR CITE
-------
1 filter sandwich. Particulate ammonia measurements were also compared, and showed more
2 scatter than the sulfate measurements, but less than the nitrate measurements.
3 An intercomparison of both PM10 and PM2 5 mass measurements was conducted during the
4 1998 Baltimore PM Study (Williams et al., 2000). PM monitors were collocated at a residential
5 indoor, residential outdoor, and ambient monitoring site within Baltimore County, MD. PM
6 samplers included TEOMs, PM2 5 FRMs, cyclone-based inlets manufactured by University
7 Research Glassware (URG), and Versatile Air Pollution Samplers (VAPS). Personal
8 Environmental Monitors (PEMs; MSP, Inc.) also were included but will not be discussed in this
9 section. The VAPS sampler is a dichotomous sampler operating at 33 L/min (one coarse particle
10 channel at 3 L/min, and two fine particle channels at 15 L/min, each). In the configuration
11 employed during this study, one fine particle channel was operated with a Teflon filter, backed
12 by a nylon filter and preceded by a sodium carbonate coated annular denuder; the second fine
13 particle channel has a quartz filter preceded by a citric acid-coated annular denuder; and the
14 coarse particle channel had a polycarbonate filter followed by a Zefluor filter for flow
15 distribution. Differences in PM25 mass concentrations between the samplers, although not large,
16 were attributed to potential particle nitrate losses, denuder losses, and losses of SVOC for some
17 samplers. Differences between coarse particulate mass concentrations, on the other hand, varied
18 widely between the instruments.
19 In another intercomparison study, Tolocka et al. (2000) examined the magnitude of
20 potential sampling artifacts associated with the use of the FRM by collocating FRMs alongside
21 other chemical speciation samplers at four U.S. cities. The locations included a high nitrate and
22 carbon, low sulfate site (Rubidoux, CA); high crustal, moderate carbon and nitrate site
23 (Phoenix); high sulfate, moderate carbon, and low nitrate (Philadelphia); and low PM2 5 mass
24 (Research Triangle Park, NC). The use of Teflon and heat-treated quartz filters also was
25 examined in this study. The Teflon filters collected less nitrate than the heat-treated quartz
26 filters. Filters in samplers using denuders to remove organic gases collected less organic PM
27 than filters in samplers without denuders.
28
29 2.2.4.2 Speciation Monitoring
30 In addition to FRM sampling to determine compliance with PM standards, EPA requires
31 states to conduct chemical speciation sampling primarily to determine source categories and
March 2001 2-74 DRAFT-DO NOT QUOTE OR CITE
-------
1 trends (Code of Federal Regulations, 1999c). A PM2 5 chemical speciation network is being
2 deployed that will consist of 54 core National Ambient Monitoring Stations (NAMS) and
3 250 State and Local Air Monitoring Stations (SLAMS). The overall goal of the speciation
4 program is "to provide ambient data that support the Nation's air quality program objectives."
5 (U.S. Environmental Protection Agency, 1999). The NAMS speciation sites will provide routine
6 chemical speciation data that will be used to develop annual and seasonal aerosol
7 characterization, air quality trends analysis, and emission control strategies. The SLAMS
8 speciation sites will further support the NAMS network and provide information for
9 development of State Implementation Plans (SIPs). At both types of sites, aerosol samples will
10 be collected for analysis of trace elements, ions (sulfate, nitrate, ammonium, sodium, and
11 potassium), and total carbon. The NAMS speciation sites will operate on a 1 in 3 day schedule,
12 with 10 of these sites augmented for everyday operation. The SLAMS speciation sites will
13 generally operate on a 1 in 6 day basis; however, many sites may be operated on a 1 in 3 day
14 basis in locations where increased data collection is needed. The current samplers include three
15 filters: (1) Teflon for equilibrated mass and elemental analysis (EDXRF), (2) a nitric acid
16 denuded Nylon filter for ion analysis (ion chromatography), and (3) a quartz fiber filter for
17 elemental and organic carbon (but without any correction for positive or negative artifacts caused
18 by adsorption of organic gases or the quartz filters or evaporation of semivolatile organic
19 compounds from the collected particles) (thermal optical analysis via NIOSH 5040 method).
20 There are several samplers that are suitable for use in the NAMS/SLAMS network. These
21 samples include an inlet cutpoint comparable to the WINS, FRM; proven denuder technology for
22 ions; and sampler face velocity and sample volume similar to that of the FRM with 46.2-mm
23 diameter filters. Information and reports on EPA's speciation monitoring program may be found
24 on EPA's Technology Transfer Network at http://www.epa.gov/ttn/amtic/pmspec.html.
25
26 Measurements for Source Category Apportionment
27 Chemical analyses from the speciation network will be used for source category
28 apportionment via receptor modeling of PM. There are two major approaches to receptor
29 modeling: the chemical mass balance (CMB) receptor modeling approach, and statistically based
30 approaches. The CMB approach requires chemical characterization of all relevant sources.
31 Similar analyses should be used for characterization of receptor samples. One of the advantages
March 2001 2-75 DRAFT-DO NOT QUOTE OR CITE
-------
1 of using the CMB approach for receptor modeling is that it can be applied to a single sample, or
2 to a limited number of samples. CMB also uses chemical analyses that are performed routinely
3 on speciation samples, such as EDXRF and ionic species. A considerable amount of receptor
4 modeling work has been conducted with CMB using elemental analyses coupled with OC/EC
5 and some ionic species (e.g., Watson et al., 1994; Hidy and Venkataraman, 1996; McLaren et al.,
6 1996; Vega et al., 1997). Recent developments in receptor modeling include using organic
7 analyses for tracers of specific sources (Benner et al., 1995), very detailed organic analyses for
8 source fingerprinting (Rogge et al., 1991, 1993b,c,d,e, 1994, 1997a,b, 1998), and chemical mass
9 balance receptor modeling (Schauer et al., 1996). Further detail on the organic analyses for these
10 studies is beyond the scope of this chapter.
11 Statistical models based upon factor analysis or principal component analysis (PCA) do not
12 require detailed source characterization information but have the drawback of requiring a large
13 data set of receptor sample analyses. These statistically based models have an additional benefit
14 in that they also can use other parameters such as meteorology. For a detailed review of factor
15 analysis and PCA, see Henry et al. (1984). In PCA, many intercorrelated variables within a large
16 data set are sorted into a smaller number of independent components, or factors, that account for
17 the variability in the data set. Veltkamp et al. (1996) performed a PCA for a study conducted at
18 Niwot Ridge, CO. Organic constituents of atmospheric aerosols were measured, along with
19 physical and meteorological data. Organic compounds were thermally desorbed from the aerosol
20 particles at 250 °C in a pure helium atmosphere, separated by gas chromatography, and identified
21 by mass spectrometry. A principle component analysis was conducted using 31 variables that
22 included 18 particulate organic compounds, 11 gas-phase species (e.g., NO, NO2, FDSTO3, HONO,
23 PAN, H2O2, etc.), wind direction, and time of day. Several factors were identified that
24 distinguished various sources. These included gas-phase internal combustion products;
25 particulate phase, oxygenated biogenic hydrocarbons; high molecular weight n-alkanes;
26 particulate phase anthropogenic products; and particulate phase biogenic aldehydes.
27 Pinto et al. (1998) also used a combination of PM25 chemical speciation and ambient
28 monitoring data in a receptor modeling calculation to determine the relative sources of particulate
29 pollution in an industrial area in the northern Bohemia region of the Czech Republic. During
30 that study, a severe air pollution episode occurred in 1993 during which smoke and SO2
31 concentrations were 1800 and 1600 //g/m3, respectively.
March 2001 2-76 DRAFT-DO NOT QUOTE OR CITE
-------
1 In addition to chemical speciation for factor analysis and source apportionment, Norris
2 et al. (1999) showed that meteorological indices could prove useful in identifying sources of
3 particulate matter that are responsible for observed health effects (specifically asthma) associated
4 with exposure to particulate matter. They examined meteorology associated with elevated
5 pollution events in Spokane and Seattle, WA, and identified a "stagnation index" that was
6 associated with low wind speeds and increases in concentrations of combustion-related
7 pollutants. Their factor analysis also identified a meteorological index (low relative humidity
8 and high temperatures) that was associated with increases in soil-derived particulate matter, as
9 well as a third factor (low temperatures and high relative humidity) that was associated with
10 increasing concentrations of particulate sulfate and nitrate species (Norris, 1998).
11 Ondov (1996) examined the feasibility of using sensitive isotopic and elemental tracer
12 materials to determine the contributions of petroleum-fueled sources of PM10 in the San Joaquin
13 Valley, CA. Costs of these experiments are affected not only by the tracer materials cost, but
14 also by the sensitivities of the analytical methods for each, as well as the background levels.
15 Suarez et al. (1996) used iridium tracer to tag emissions from diesel-burning sanitation trucks in
16 Baltimore and determined the size distribution of soot from the trucks.
17
18 Elemental Analyses
19 X-ray emission, stimulated either by X rays (X-ray fluorescence, XRF) or by proton beams
20 (Proton Induced X-ray Emission, PIXE) are standard techniques for nondestructive analysis of
21 certain elements. Both were discussed in the previous 1996 PM AQCD. Some newer techniques
22 with some advantages have become available in recent years.
23
24 Energy Dispersive X-ray Fluorescence (EDXRF). EDXRF by Method IO-4.4 is the
25 method of choice for analysis of trace elements for the NAMS speciation program. EDXRF can
26 accommodate small sample sizes and requires little sample preparation or operator time after the
27 samples are placed into the analyzer. It also leaves the sample intact after analysis so further
28 analysis is possible. The previous 1996 PM AQCD included a detailed discussion of EDXRF, so
29 that will not be repeated here.
30
March 2001 2-77 DRAFT-DO NOT QUOTE OR CITE
-------
1 Instrumental Neutron Activation Analysis (INAA). INAA was mentioned only briefly in
2 the 1996 PM AQCD and is expanded on here. INAA has been used to examine the chemical
3 composition of atmospheric aerosols in several studies, either as the only method of analysis, or
4 in addition to XRF (e.g., Yatin et al., 1994; Gallorini, 1995). INAA has higher sensitivity for
5 many trace species, and it is particularly useful in analyzing for many trace metals. Landsberger
6 and Wu (1993) analyzed air samples collected near Lake Ontario for Sb, As, Cd, In, I, Mo, Si,
7 and V, using INAA. They demonstrated that using INAA in conjunction with epithermal
8 neutrons and Compton suppression produces very precise values with relatively low detection
9 limits.
10 Enriched rare-earth isotopes have been analyzed via INAA and used to trace sources of
11 particulate matter from a coal-fired power plant (Ondov et al., 1992), from various sources in the
12 San Joaquin Valley (Ondov, 1996), from intentially tagged (iridium) diesel emissions from
13 sanitation trucks (Suarez et al., 1996; Wu et al., 1998), and from iridium-tagged emissions from
14 school buses (Wu et al., 1998).
15 An intercomparison was conducted in which 18 pairs of filters were sent to participants in
16 the Coordinated Research Program (CRP) on Applied Research on Waste Using Nuclear Related
17 Analytical Techniques (Landsberger et al., 1997). As part of that study, participants used PIXE,
18 INAA, XRF, or AAS to analyze the samples. Many of the results for XRF and PIXE in the
19 coarse fraction were observed to be biased low compared to INAA. The authors speculated that
20 there is a systematic error because of self-attenuation of the X rays resulting from the particle
21 size effect.
22 In source apportionment studies, it is possible to use a combination of XRF and INAA to
23 develop a relatively complete set of elemental measurements. Between these two analytical
24 techniques, good sensitivity is possible for many elements, including most of the toxic metals of
25 interest. The previous 1996 PM AQCD compared several methods for measuring elements.
26 In general, XRF provides better sensitivity for some metals (e.g., Ni, Pb, Cu, and Fe), whereas
27 INAA provides better sensitivity for others (Sb, As, Cr, Co, Se, and Cd). Both methods provide
28 similar detection limits for still other elements such as V, Zn, and Mn.
29
30 Atomic Absorption Spectrophotometry (AAS). AAS was used to characterize the
31 atmospheric deposition of trace elements Zn, Ni, Cr, Cd, Pb, and Hg, to the Rouge River
March 2001 2-78 DRAFT-DO NOT QUOTE OR CITE
-------
1 watershed by particulate deposition (Pirrone and Keeler, 1996). The modeled deposition rates
2 were compared to annual emissions of trace elements that were estimated from the emissions
3 inventory for coal and oil combustion utilities, iron and steel manufacturing, metal production,
4 cement manufacturing, and solid waste and sewage sludge incinerators. They found generally
5 good agreement between the trend observed in atmospheric inputs to the river (dry + wet
6 deposition) and annual emissions of trace elements, with R2s varying from «0.84 to 0.98. Both
7 atmospheric inputs and emissions were found to have followed downward trends for Pb. For the
8 period of 1987 to 1992, steady increases were observed for Cd (major sources are municipal solid
9 waste incineration, coal combustion, sludge incineration, and iron and steel manufacturing),
10 Cr and Ni (major sources are iron and steel production and coal combustion), and Hg (major
11 sources are coal, the contribution from which had decreased from 53 to 45%, and municipal,
12 solid, and medical waste incineration, the contribution from which has increased).
13
14 Inductively Coupled Plasma-Mass Spectroscopy (ICP-MS). Keeler and Pirrone (1996)
15 also used ICP-MS to determine trace elements Cd, Mn, V, As, Se, and Pb in atmospheric
16 parti culate fine (PM2 5) and total suspended parti culate samples collected in two Detroit sites.
17 The results were then similarly used in a deposition model to estimate the dry deposition flux of
18 trace elements to Lake Erie.
19
20 Scanning Electron Microscopy (SEM). Mamane et al. (2000) investigated the use of
21 computer-controlled scanning electron microscopy (CCSEM) as a way of supplementing XRF
22 analysis and providing automated analysis of particle size, chemistry, and particle classification.
23 An ambient coarse particulate sample from Baltimore was collected on a polycarbonate filter for
24 this analysis. CCSEM analyses were conducted for 2819 particles in 78 randomly selected fields
25 of view during an unattended 8-h run. Mamane et al. confirmed the stability of the CCSEM
26 instrument over several hours of operation. The physical properties of the sample such as
27 particle diameter, mass loading per field, and particle number per field, were well represented by
28 analyzing approximately 360 particles, with little additional information gained by analyzing
29 more particles. Teflon filters are not well suited for SEM analyses. Analysis of fine PM is
30 expected to pose analytical challenges not addressed in the present study (Mamane et al., 2000).
March 2001 2-79 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nelson et al. (2000) applied Raman chemical imaging and SEM (Raman/SEM) to study the
2 size, morphology, elemental and molecular composition, and molecular structure of fine
3 particulate matter. In their study, filter compatibility was examined, and Raman/SEM chemical
4 imaging was conducted for several standard materials as well as for ambient PM2 5 samples.
5 Polycarbonate was determined to be a suitable substrate for both SEM and Raman chemical
6 imaging analysis.
7
8 Elemental and Organic Carbon in Particulate Matter. Total carbon in aerosol particles
9 (TC) can be expressed as the sum of organic carbon (OC), elemental carbon (EC), and carbonate
10 carbon (CC), with the contribution of CC to TC usually on the order of 5% or less, for particulate
11 samples collected in urban areas (Appel, 1993). The 1996 PM AQCD (U.S. Environmental
12 Protection Agency, 1996) listed several filter-based, thermal methods for measuring OC and EC,
13 and described the thermal/optical reflectance (TOR) method, which was noted, along with
14 thermal manganese oxidation, to be one of the most commonly applied methods in the United
15 States at the time. In thermal separation methods, thermally evolved OC and EC are oxidized to
16 CO2 and quantified either by nondispersive infrared detection or electrochemically, or the CO2
17 can be reduced to CH4 and quantified via flame ionization detection (FID). The various methods
18 give similar results for TC, but not for EC or OC.
19 In a methods comparison study (Countess, 1990), it was shown that it is necessary to
20 minimize or correct for pyrolytically generated EC ("char"), and that CC found in wood smoke
21 and automobile exhaust samples may interfere with some of the thermal methods. Recently,
22 Lavanchy et al. (1999) reported on a study in which the operation of a catalytic oxidation system
23 was modified in an attempt to minimize pyrolysis of OC and at the same time minimize the
24 contribution of CaCO3. The system uses two ovens, one at 340 °C and one at 650 °C. The filter
25 sample is placed in a moveable sample boat. In order to minimize charring, the sample is first
26 flash heated in the 650 °C oven for 1 min. Then it is inserted into the 340 °C stage of the two-
27 stage oven. In both steps OC is oxidized to CO2 in the presence of O2. The second step requires
28 42 min. The filter then is moved into the second-stage oven. During this third step, EC is
29 oxidized to CO2 at 650 °C for 32 min. This temperature is reported to be sufficient to completely
30 oxidize EC, but with only about 1% of the CaCO3 being vaporized (Lavanchy et al., 1999;
31 Petzold et al., 1997). To test for charring, they challenged their system with atmospheric samples
March 2001 2-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 for which duplicates were analyzed via the German reference method for measuring OC and EC
2 in atmospheric samples (Petzold and Niessner, 1995), in which a solvent extraction is used to
3 remove organics before combustion. Lavanchy et al. (1999) reported a high correlation
4 (R2 = 0.97) between their thermal oxidation method and the German method VDI. The slope of
5 the EC:EC VDI line was 0.92, and the intercept was -0.37 //g cm"2. They also reported detection
6 limits of 1.3 //g for EC and 1.8 //g for OC.
7 Pyrolytic char is corrected for in thermal-optical analysis. In thermal-optical carbon
8 analysis (Birch and Gary, 1996; Chow et al., 1993), punches from a quartz sampling filter are
9 inserted into the carbon analyzer and heated in a helium atmosphere to volatilize organic carbon.
10 Then, the temperature is reduced, and oxygen is added to the carrier gas, so that desorbed
11 compounds are then oxidized to CO2, reduced to methane, and measured in a flame ionization
12 detector. In order to account for the portion of the OC that is pyrolyzed, a He-Ne laser monitors
13 the sample reflectance (or transmittance). As the pyrolysis occurs, the sample gets darker, and
14 the reflectance decreases. Then, as elemental carbon is removed, the filter lightens, and the
15 reflectance increases until all carbon has been removed from the filter. The split between organic
16 and elemental carbon is considered to be the point at which the reflectance regains its
17 prepyrolysis value, with material removed prior to this point being considered organic, and that
18 after, elemental.
19 The thermal/optical transmission method (TOT) is similar to the TOR with the primary
20 difference being that light transmission rather than reflectance is monitored on the filter
21 throughout the analysis. The TOT method of Birch and Gary (1996) consists of a two-stage
22 process, with the first stage being conducted in a pure helium atmosphere, and the second stage
23 conducted in a 10% oxygen-helium mix. The temperature is ramped to about 820 °C in the
24 helium phase, during which organic and carbonate carbon are volatilized from the filter. In the
25 second stage, the oven temperature is reduced, and then raised to about 860 °C. During the
26 second stage, pyrolysis correction and EC measurement are made. Figure 2-14, an example of a
27 thermogram, shows temperature, transmittance, and FID response traces. Peaks are evident that
28 correspond to OC, CC, EC, and pyrolitic carbon (PC). As can be seen in this figure, the high
29 temperature in the first stage allows for decomposition of CC. The ability to quantify PC is
30 particularly important in high OC/EC regions (like wood-smoke-impacted airsheds), allowing for
March 2001 2-81 DRAFT-DO NOT QUOTE OR CITE
-------
OC - EC split
Time/min
Figure 2-14. This thermogram, for a sample containing rock dust (a carbonate source) and
diesel exhaust, shows three traces that correspond to temperature, filter
transmittance, and FID detector response. Peaks correspond to organic (OC),
carbonate (CC), pyrolytic (PC), and elemental (EC) carbon. The final peak is
a methane calibration peak.
Source: Birch and Gary (1996).
1 the volatilization of any remaining complex organic compounds so they are not apportioned to
2 the EC phase.
3 The National Institute for Occupational Safety and Health (NIOSH) Method 5040 for
4 monitoring elemental carbon as a marker for particulate diesel exhaust is based upon a TOT
5 method analyzer (Birch and Cary, 1996), while the OC/EC method specified for the IMPROVE
6 network is the TOR method (Chow et al., 2000). Chow et al. (2000) compared the OC, EC, and
March 2001
2-82
DRAFT-DO NOT QUOTE OR CITE
-------
1 TC measurements from NIOSH and IMPROVE methods. The two methods differ in that
2 temperature and atmospheric controls that are used to accomplish carbon speciation, in addition
3 to the use of light transmission in the case of the NIOSH method, as compared to light
4 reflectance in the IMPROVE method, of the filter is measured during the analysis. The
5 IMPROVE thermal protocol specifies organic carbon fractions at 120, 250, 450, and 550 °C in a
6 nonoxidizing atmosphere (He), and elemental organic fractions at 550, 700, and 800 °C in an
7 oxidizing atmosphere. The NIOSH method differs in its thermal protocol, which has organic
8 carbon fractions at 250, 500, 650, and 850 °C in a nonoxidizing atmosphere (also He), and
9 elemental carbon fractions at 650, 750, and 850 °C in an oxidizing atmosphere. The high
10 temperature before addition of oxygen in the NIOSH method is to quantify particulate carbonate,
11 which evolves between 650 and 830 °C (Birch and Gary, 1996). The two methods also differ in
12 the specified residence times at each temperature setpoint. The residence times at each setpoint
13 are typically longer for the IMPROVE analysis compared to the NIOSH analysis.
14 Chow et al. (2000) analyzed 60 quartz filter samples that represented a wide variety of
15 aerosol compositions and concentrations. The TC measurements from each protocol were in
16 good agreement, with no statistically significant differences. A statistically significant difference
17 was observed in the fraction of TC that is attributed to EC, as determined by the IMPROVE and
18 NIOSH thermal evolution protocols, with the IMPROVE EC measurements typically higher than
19 the NIOSH EC measurements. This difference was attributed to the 850 °C temperature step in
20 the oxidizing atmosphere in the NIOSH protocol. Chow et al. compared the OC for each method
21 and found that the two methods showed good agreement when the 850 °C nonoxidizing
22 temperature step in the NIOSH method was not included in determination of OC. There was also
23 a difference between the reflectance and transmittance detection methods in the pyrolysis
24 adjustment, although this difference was most noticeable for very black filters for which neither
25 reflectance nor transmittance was able to accurately detect further blackening by pyrolysis.
26 Because OC and EC are operationally defined parameters, Chow et al. pointed out that it is
27 importance to retain ancillary information when reporting EC and OC by these analytical
28 methods, so that comparisons can be made among measurements taken at different sites using
29 these two methods.
30 Further refinement of thermal techniques has resulted in the evolved gas analysis (EGA)
31 method, described by Grosjean et al. (1994). This technique involves combustion of paniculate
March 2001 2-83 DRAFT-DO NOT QUOTE OR CITE
-------
1 matter samples in an oxidizing environment while the temperature is raised from 100 to 600 °C.
2 The amount of evolved CO2 contains information about the volatility of the organic aerosol
3 compounds. Grosjean et al. (1994) present thermograms both for specific organic compounds
4 (e.g., adipic acid) and for specific sources (e.g., vehicular traffic). They suggest that EGA may
5 be useful for source apportionment applications. Kirchstetter et al. (2000) and Novakov et al
6 (1997) have used EGA to provide insights regarding organic sampling artifacts.
7 Black carbon (BC) also is used, in addition to the thermal and thermal/optical methods, for
8 determining EC as a measure of soot (Penner and Novakov, 1996). Both EC and BC define a
9 similar fraction of aerosol, but EC is determined based on thermal properties, whereas BC is
10 based on light-absorption properties. Optical methods for determining BC tend to suffer from
11 calibration problems (Hitzenberger et al., 1996). Lavanchy et al. (1999) compared their EC
12 concentrations as determined from their catalytic thermal oxidation method to BC concentrations
13 determined using an aethalometer operated at the same site, and found that the instrumental
14 calibration factor provided by the manufacturer was on the order of two times the calibration
15 factor they determined (9.3 ± 0.4 m2g4). It is possible to calculate a theoretical specific
16 absorption coefficient (Ba) from Mie theory given a known size distribution and refractive index,
17 and typically BC aerosols have values of Ba between 3 and 17 n^g"1 (Hitzenberger et al. [1996]
18 and references therein). The Ba is defined as absorption per mass concentration and can be
19 calculated given the sample filter area, the total deposited mass, and absorption signals for both
20 the loaded and unloaded filters. Often, when no direct measurements are available, values of Ba
21 on the order of 10 m2g4 have been used (Hitzenberger et al. (1996), and references therein).
22 European countries are trying to set air pollution standards that target diesel vehicles, one of the
23 principal sources of BC in urban areas (Hitzenberger et al. (1996), and references therein) and so
24 it is essential that accurate values for Ba are available. Hitzenberger et al. (1996) investigated the
25 feasibility of using an integrating sphere photometer as an adequate measurement system for the
26 BC content and the absorption coefficient. Based on samples collected during a 10-day period in
27 May 1994, they determined that the usually assumed value of 10 n^g"1 was also applicable to
28 aerosol BC occurring in Vienna.
29 Hitzenberger et al. (1999) recently reported on a study in which the integrating sphere
30 method was compared to an aethalometer (Hansen et al., 1984), the thermal method of Cachier
31 et al. (1989), and the thermal/optical method of Birch and Gary (1996). The absorption
March 2001 2-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 coefficients that were obtained from both the integrating sphere and the aethalometer were
2 comparable. The BC mass concentrations obtained from the aethalometer were 23% of those
3 obtained from the integrating sphere. Compared to the thermal method, the integrating sphere
4 overestimated the BC mass concentrations by 21%. Compared to the thermal/optical method, the
5 integrating sphere was within 5% of the 1:1 line. However, the data were not so well correlated.
6 In 1986, the Carbonaceous Species Methods Comparison Study (CSMCS) was conducted
7 in Los Angeles, during which a number of methods for the measurement of this species were
8 intercompared. The CSMCS was mentioned in the 1996 PM AQCD (U.S. Environmental
9 Protection Agency, 1996). Hansen and McMurry (1990) specifically compared two very
10 dissimilar methods for aerosol elemental carbon—collection of impactor samples backed by a
11 quartz fiber afterfilter, followed by EC analysis by oxidation in helium over a MnO2 catalyst, and
12 real-time measurements using an aethalometer (an optical absorption technique)—and found
13 good agreement between these two, very different methods. The CSMCS interlaboratory
14 precision for total carbon was 4.2% (Turpin et al., 2000). However, because the split between
15 OC and EC is operationally defined, there was substantial interlaboratory variability in OC and
16 EC (e.g., 34% for EC [Turpin et al., 1990]). The implications for data analysis are twofold:
17 (1) the analysis method used must be reported with particulate carbon data, and (2) comparative
18 analyses should not be conducted with data analyzed by more than one carbon analysis method
19 unless the mutual compatibility of the methods has been demonstrated. Carbon analysis methods
20 currently are being compared as a part of the Atlanta Supersite.
21 Turpin et al. (1990) reported on an in situ, time-resolved analyzer for particulate organic
22 and elemental carbon that could operate on a time cycle as short as 90 min. The analyzer
23 consists of a filter-based sampling section and a thermal-optical carbon detector. Adsorbed
24 organic material is thermally desorbed from the filter at 650 °C and oxidized at 1000 °C over a
25 MnO2 catalyst bed. The evolved CO2 is converted to methane over a nickel catalyst, and the
26 methane is measured in a flame ionization detector. Then the elemental carbon is oxidized in a
27 98% He-2% O2 atmosphere, at 350 °C. Correction is made for pyrolytic conversion of some of
28 the organic particulate matter. The instrument was operated with a 2-h time resolution during the
29 Southern California Air Quality Study (SCAQS) in 1987 (Turpin and Huntzicker, 1991), as well
30 as during the Carbonaceous Species Methods Comparison Study (CSMCS) in 1986.
31
March 2001 2-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 EC/OC Summary. The state of the art for soot measurements continues to develop, and,
2 although advances are being made, the definitions of EC and BC continue to be operational and
3 determined by the method employed. Similarly, the distinction between OC and EC also is
4 defined operationally. Therefore, reports of EC/OC measurements should include mention of the
5 method with which the species were determined. Finally, if possible, all ancillary data should be
6 retained, to allow later comparison to other methods.
7
8 Ions. Aerosol ions refers to the water-soluble portion of suspended PM. Ion
9 chromatography (1C) is widely used for analyzing ionic species. 1C is the method of choice for
10 the measurement of sulfate, nitrate, ammonium, sodium, and potassium ions for the NAMS
11 program. Aerosol strong acidity, H+, is determined by titration of a water solution of PM
12 collected following a series of annular denuders to remove acid and basic gases with back-up
13 filters to collect NH3 and HNO3 that might volatilize from the PM during collection. The 1996
14 PM AQCD (U. S. Environmental Protection Agency, 1996) discussed measurement of ions by 1C
15 (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
16 discussed here.
17
18 2.2.5 Continuous Monitoring
19 The EPA expects that more than 200 local agency monitoring sites throughout the States
20 will operate continuous PM monitors. All currently available continuous measurements of
21 suspended particle mass share the problem of dealing with semivolatile PM components. So as
22 not to include particle-bound water as part of the mass, the particle-bound water must be
23 removed by heating or dehumidification. However, heating also causes loss of ammonium
24 nitrate and semivolatile organic components. A variety of potential candidates for continuous
25 measurement of mass or chemical components will be discussed in this section.
26
27 2.2.5.1 Tapered Element Oscillating Microbalance
28 The advantages of continuous PM monitoring, and the designation of the Tapered Element
29 Oscillating Microbalance (TEOM) as an equivalent method for PM10, have led to the deployment
30 of the TEOM at a number of air monitoring sites. The TEOM also is being used to measure
31 PM2 5. The TEOM differs philosophically from the federal reference methods for particulate
March 2001 2-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 mass in that it does not require equilibration of the samples at a specified temperature and
2 relative humidity. Moreover, the TEOM samples at a constant temperature, typically heated to
3 some temperature higher than the ambient temperature (Meyer et al., 1995; Meyer and
4 Rupprecht, 1996), whereas the FRM samples at the ambient temperature. Thus, the TEOM does
5 not provide data equivalent to the FRM because of losses of volatile species. Volatilization
6 losses in the TEOM sampler can be reduced by operating the instrument heated to 30 °C rather
7 than the 50 °C specified, during the cooler times of the year, and by using Nafion dryers on the
8 inlet.
9 This philosophical difference in operation and implications for fine particle measurements
10 were examined by researchers at CSIRO Atmospheric Research in Australia (Ayers et al., 1999).
11 That group compared 24-h mean PM2 5 mass concentrations as determined by a TEOM and by
12 two manual, gravimetric samplers (a low-volume filter sampler and a MOUDI sampler) in four
13 Australian cities, on 15 days in the winter half-year. The TEOM was operated at 50 °C at one
14 location and at 35 °C at the other three locations. A systematically low TEOM response in
15 comparison to the integrated gravimetric methods was observed. In a comprehensive study,
16 Allen et al. (1997) reported results in which TEOM data collected at 10 urban sites in the United
17 States and Mexico were compared with 24-h integrated mass concentrations, for both PM10 and
18 PM25. They collected a large data set that included both winter and summer seasons. Allen etal.
19 (1997) concluded that, especially for urban areas, a significant fraction of PM10 could be
20 semivolatile compounds that could be lost from the heated filter in the TEOM, thus leading to a
21 systematic difference between the TEOM and the EPA FRM for PM10. They suggested that this
22 difference is likely to be larger for PM2 5 than it is for PM10 (Allen et al., 1997).
23 In a similar study conducted in Vancouver, British Columbia, the effect of equilibration
24 temperature on PM10 concentrations from the TEOM was examined. Two collocated TEOM
25 monitors, operated at 30 and 50 °C, respectively, were operated in the Lower Fraser Valley in
26 British Columbia for a period of approximately 17 mo (Mignacca and Stubbs, 1999). A third
27 TEOM operating at 40 °C was operated for 2 mo during this period. They found that, on
28 average, the 1-h PM10 from the TEOM operating at 30 °C was consistently greater than that from
29 the TEOM operated at 50 °C. For the period during which the third TEOM was operated (at
30 40 °C), the PM10 from that instrument was between those values for the other two instruments.
31 They also found that the differences in masses were proportional to the PM10 loading, and more
March 2001 2-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 strongly correlated to the PM10 from the TEOM operated at the lower temperature. They
2 recommended that the TEOM monitors be operated at 40 °C as opposed to operating at 50 °C in
3 summer and 30 °C in winter, in order to avoid introducing a methodological seasonal bias.
4 In most parts of Canada, and including the Greater Vancouver Regional District, TEOMs for
5 both PM10 and PM25 are operated with this revised protocol.
6 A new sample equilibration system (SES) was developed to allow conditioning of the
7 sample stream to lower humidity and temperature, to reduce losses of semivolatile species from
8 the TEOM (Meyer et al., 2000). The SES utilizes a Nafion dryer designed for low particle loss,
9 and humidity sensors. The dryer fits between the flow splitter that follows the size-selective inlet
10 and the sensor unit. A dry purge gas flow over the exterior of the Nafion tubing allows for self-
11 regeneration. A TEOM with PM2 5 inlet and equipped with an SES was operated at 30 °C
12 alongside another TEOM operating at 50 °C without the SES in Albany, NY, over a 6-day period
13 during a summertime high-temperature, high-relative-humidity episode. The SES maintained the
14 sample air relative humidity under 30% and the TEOM with the SES generally measured more
15 mass than the other TEOM. The TEOM with SES also was operated alongside an FRM-type
16 sampler for the period of June 6 through September 25, 1999. The correlation between the FRM
17 and TEOM/SES showed a slope of 1.0293 and R2 of 0.9352, whereas the correlation between the
18 FRM and the TEOM without SES and operating at 50 °C showed a slope of 0.8612 and R2 of
19 0.8209. The SES can be installed on existing TEOM monitors.
20 Patashnick et al. (2000) developed a differential TEOM system that is based on a pair of
21 TEOM sensors, each of which is preceded by its own electrostatic precipitator (ESP), and
22 downstream from a common size selective inlet. By alternately switching the ESPs on and off,
23 and out of phase with each other, the two sensors measure "effective mass" that includes both the
24 nonvolatile component and the volatile component sampled by the TEOM, less the volatile
25 component that vaporized during the sampling interval. On the sensor side with the ESP turned
26 on, there is no particle collection on that filter, so that only volatilization of previously collected
27 particles continues. This would allow for correcting the effective mass as measured from the
28 first sensor, by subtracting out the volatilization artifact, and leaving the nonvolatile and volatile
29 components of the particulate matter. This system has yet to be well characterized for other
30 biases or interferences such as reactions on the filters, particle collection efficiency of the ESPs,
31 and particle and semivolatile material losses.
March 2001 2-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.5.2 Real-Time Total Ambient Mass Sampler
2 A Real-Time Total Ambient Mass Sampler, RAMS, based on diffusion denuder and
3 TEOM monitor technology has been developed, validated, and field tested (Eatough et al.,
4 1999a; Obeidi and Eatough, 1999; Pang et al., 2001) for the real-time determination of total fine
5 particulate mass, including semivolatile species. The RAMS measures total mass of collected
6 particles, including semivolatile species with a TEOM monitor using a "sandwich filter". The
7 sandwich contains a Teflon coated particle collection filter followed by a charcoal impregnated
8 filter (GIF) to collect any semivolatile species lost from the particles during sampling. Because
9 the instrument measures total mass collected by the sandwich filter, all gas phase compounds that
10 can be adsorbed by a GIF must be removed from the sampling stream prior to the TEOM
11 monitor. Laboratory and field validation data indicate that the precision of fine PM mass
12 determination is better than 10%. The RAMS uses a Nafion dryer to remove particle bound
13 water from the suspended particles and a particle concentrator to reduce the amount of gas phase
14 organics that must be removed by the denuder. An example of data from the RAMS, the TEOM,
15 and the PC-BOSS is shown in Figure 2-15. This figure also shows the PM2 5 mass from the
16 TEOM as being negative for the hours of 16 to 19. This likely results from the loss of volatile
17 materials from the heated filter.
18
19 2.2.5.3 Continuous Ambient Mass Monitor
20 Koutrakis and colleagues (Koutrakis et al., 1996; Wang, 1997) have developed the
21 Continuous Ambient Mass Monitor (CAMM), a technique for the continuous measurement of
22 ambient particulate matter mass concentration, based on the measurement of pressure drop
23 increase with particle loading across a membrane filter. Recently, Sioutas et al. (1999) examined
24 the increase in pressure drop with increasing particle loading on Nuclepore filters. They tested
25 filters with two pore diameters (2 and 5 //m) and filter face velocities ranging from 4 to 52 cm/s,
26 and examined the effects of relative humidity in the range of 10 to 50%. They found that, for
27 hygroscopic ammonium sulfate particles, the change in pressure drop per unit time and
28 concentration was a strong function of relative humidity, decreasing with increasing relative
29 humidity. These results suggest that particulate concentration measurements like the method of
30 Koutrakis et al. (1996) that use the pressure drop method may be subject to additional
31 uncertainties if used in an environment where the ambient relative humidity cannot be controlled
March 2001 2-89 DRAFT-DO NOT QUOTE OR CITE
-------
PC-BOSS (Nonvolatile Material)
PC-BOSS (Lost From Particles)
TEOM
at35C
—B— RAMS
at35C
FRM PM25
24 h average
100
Riverside, CA
•no.
CM
s±£_ 5~
13 14 15 16 17 18 19 20 21 22 23 0 1 2 3 4 5 6 7 8 9
Time of Day
Figure 2-15. Comparison of mass measurements with collocated RAMS (real-time data),
PC-BOSS (1-h data), FRM PM2 5 sampler (24-h data), and a conventional
TEOM monitor (real-time data). The semivolatile fine particulate matter is
sampled with the RAMS and PC-BOSS, but not with the TEOM monitor or
the FRM PM2 5 single filter sampler. The PC-BOSS provides information on
both the nonvolatile component (NV) and the semivolatile organic component
(SVOC).
Source: Eatoughetal. (1999a).
1 accurately. The current version of the CAMM (Wang, 1997) uses a particle concentrator, a
2 Nafion dryer, and frequent changes of the position on the filter tape where the pressure drop
3 measurement is made to avoid artifacts due to semivolatile components.
4 The CAMMS was recently operated alongside a gravimetric PM method (the Harvard
5 Impactor, or HI) in seven U.S. cities selected for their distinctly different ambient particulate
March 2001
2-90
DRAFT-DO NOT QUOTE OR CITE
-------
1 compositions and densities. The correlation between the two methods was high, with an overall
2 r2 of 0.90, and average CAMM/HI ratio of 1.07 (Babich et al., 2000).
3
4 2.2.5.4 Light Scattering
5 The evaporation of ammonium nitrate aerosol in a heated nephelometer was examined by
6 Bergin et al. (1997). This is of interest, because nephelometers are frequently operated with the
7 sampled airstream heated to a low reference relative humidity of 40%, in order to measure the
8 light scattering because of the dry aerosol rather than that caused by particle-bound water.
9 Bergin et al. conducted laboratory experiments at low relative humidity («10%) and as a function
10 of temperature (300 to 320 K), mean residence time in the nephelometer, and initial particle size
11 distribution. The evaporation of ammonium nitrate aerosol was also modeled, for comparison,
12 and was found to accurately describe the decrease in aerosol scattering coefficient as a function
13 of aerosol physical properties, and nephelometer operating conditions. Bergin et al. (1997)
14 determined an upper limit estimate of the decrease in the aerosol light scattering coefficient at
15 450 nm from evaporation for typical field conditions. The model estimates for their worst-case
16 scenario suggest that the decrease in the aerosol scattering coefficient could be roughly 40%.
17 Under most conditions, however, they estimate that the decrease in aerosol scattering coefficient
18 generally is expected to be less than 20%.
19 Morawska et al. (1996) examined the correlations between PM10, visibility, and submicron
20 concentration data in Brisbane, and concluded that the different principles of operation for each
21 instrument and the different aerosol characteristics measured by each technique make it difficult
22 to observe any relationships. Morawska et al. (1998b) reported on a long-term monitoring
23 program that included the criteria pollutants as well as light scattering, number/size distributions,
24 number concentrations, and elemental analysis via inductively coupled plasma mass
25 spectrometry. Particle size classification was conducted using a TSI scanning mobility particle
26 sizer (SMPS) for the size range of 0.016 to 0.7 //m, and a TSI aerodynamic particle sizer (APS)
27 for the size range of 0.7 to 30 //m. They reported correlation coefficients between the light-
28 scattering coefficient and PM10, SMPS concentration, and APS concentration of 0.58, 0.38, and
29 0.37, respectively. They also reported a correlation coefficient between PM10 and the SMPS
30 concentration of 0.25. A lower correlation between PM10 mass and the SMPS concentration is
March 2001 2-91 DRAFT-DO NOT QUOTE OR CITE
-------
1 consistent with the notion that PM10 mass measurements would provide less information about
2 smaller particles in the 0.016 to 0.7 //m range.
3
4 2.2.5.5 Beta-Gauge Techniques
5 The use of absorption of beta radiation as a indicator of particle mass has been used
6 effectively to measure the mass of equilibrated particulate matter collected on Teflon filters
7 (Jaklevic et al., 1981; Courtney et al., 1982). The technique also has been used to provide near
8 real-time measurements with time intervals on the order of an hour (Wedding and Weigand,
9 1993). However, real-time beta gauge monitors experience the same problems as other
10 continuous or near real-time parti culate matter mass monitoring techniques. Particle-bound
11 water must be removed to reduce the sensitivity of the indicated mass to relative humidity.
12 However, the simplest technique, mild heating, will remove a portion of the ammonium nitrate
13 and the semivolatile organic compounds as well as the particle-bound water.
14 An intercomparison study of two beta gauges at three sites indicated that the Wedding beta
15 gauge and the Sierra Anderson SA 1200 PM10 samplers were highly correlated, r > 0.97 (Tsai and
16 Cheng, 1996). The Wedding beta gauge was not sensitive to relative humidity but was
17 approximately 7% lower. This suggests that the mild heating in the beta gauge causes losses
18 comparable to those caused by equilibration, although the differences could result from slight
19 differences in the upper cut points. The Kimoto beta gauge, however, which was operated at
20 ambient temperature, was sensitive to relative humidity, yielding significantly higher mass
21 concentrations relative to the Sierra Anderson SA 1200 for RH > 80% than for RH < 80%, even
22 though the correlation with the SA 1200 was reasonable, r = 0.94 for RH > 80% and 0.83 for
23 RH < 80%.
24
25 2.2.5.6 Measurements of Individual Particles
26 Recently, several researchers have developed instruments for real-time in situ analysis of
27 single particles (e.g., Noble and Prather, 1996; Gard et al., 1997; Johnson and Wexler, 1995;
28 Silva and Prather, 1997; Thomson and Murphy, 1994). Although the technique varies from one
29 laboratory to another, the underlying principle is to fragment each particle into ions using either a
30 high-power laser or a heated surface and to then use a time-of-flight mass spectrometer (TOFMS)
31 to measure the ion fragments in a vacuum. Each particle is analyzed in a suspended state in the
March 2001 2-92 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
air stream (i.e., without collection), avoiding sampling artifacts associated with impactors and
filters. By measuring both positive and negative ions from the same particle, information can be
obtained about the chemical composition, not just the elemental composition, of individual
particles of known aerodynamic diameter. This information is especially useful in determining
sources of particles. An example of the type of information that can be determined is shown in
Figure 2-16.
o
O
JD
O
tr
ro
0.
0
CD
f- Organic
//////- Marine
Soil
0.2
\ ''''r
0.3 0.4 0.5 0.6 0.70.80.91.0
2.0
3.0
4.0
Aerodynamic Diameter (pm)
Figure 2-16. Size distribution of particles divided by chemical classification into organic,
marine, and crustal.
1 Because particles are analyzed individually, biases in particle sampling (the efficiency of
2 particle transmission into the sensor chamber as a function of size; particle size measurement,
3 and detection of particles prior to fragmentation) represent a major challenge for these
4 instruments. Moreover, the mass spectrometer has a relatively large variability in ion yields (i.e.,
5 identical samples would yield relatively large differences in MS signals [Thomson and Murphy,
March 2001
2-93
DRAFT-DO NOT QUOTE OR CITE
-------
1 1994]); therefore, quantitation is inherently difficult (Murphy and Thomson, 1997). Quantitation
2 will be even more challenging for complex organic mixtures because of the following two
3 reasons: (1) a large number of fragments are generated from each molecule, and (2) ion peaks
4 for organics can be influenced or obscured by inorganic ions (Middlebrook et al., 1998).
5 Nonetheless, scientists have been successful in using these techniques to identify the presence of
6 organics in atmospheric particles and laboratory-generated particles (i.e., as contaminants in
7 laboratory-generated sulfuric acid droplets) as well as the identification of specific compound
8 classes such as PAHs in combustion emissions (Castaldi and Senkan, 1998; Hinz et al., 1994;
9 Middlebrook et al., 1998; Murphy and Thomson, 1997; Neubauer et al., 1998; Noble and Prather,
10 1998; Reilly et al., 1998; Silva and Prather, 1997).
11 Until recently, single particle ATOFMS systems have only been able to characterize
12 particles that are larger than approximately 0.2 to 0.3 //m in diameter. Wexler and colleagues
13 (Carson et al., 1997; Ge et al., 1998) have developed a single particle, TOFMS instrument that is
14 able to size, count, and provide chemical composition on individual particles ranging in size
15 from 10 nm to 2 //m.
16 Noble and Prather (1996) used ATOFMS to provide compositionally resolved particle-size
17 distributions. Their instrument is capable of analyzing, at typical ambient concentrations, size
18 and chemical composition of 50 to 100 particles/min, and up to 600/min at high particle
19 concentrations. Data storage requirements are met using a Pentium 90 mHz personal computer.
20
21 2.2.5.7 Automated Fine Particulate Nitrate
22 An integrated collection and vaporization cell was developed by Stolzenburg and Hering
23 (2000) that provides automated, 10-min resolution monitoring of fine particulate nitrate. In this
24 system, particles are collected by humidified impaction process and analyzed in place by flash
25 vaporization and chemiluminescent detection of the evolved nitrogen oxides. In field tests in
26 which the system was collocated with two FRM samplers, the automated nitrate sampler results
27 followed the results from the FRM, but were offset lower. The system also was collocated with a
28 Harvard EPA annular denuder sampler (HEADS), as well as a SASS speciation sampler
29 (MetOne Instruments). In all these tests, the automated sampler was well correlated to other
30 samplers, with slopes near 1 (ranging from 0.95 for the FRM to 1.06 for the HEADS) and
31 correlation coefficients ranging from 0.94 to 0.996.
March 2001 2-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.5.8 Semi-continuous Carbon Analysis
2 Testing and refinement of models that simulate aerosol concentrations from gas and
3 particle emissions require air quality measurements of approximately 1 -h time resolution to
4 reflect the dynamics of atmospheric transport, dispersion, transformation and removal. Below
5 we describe instruments that have been used to collect and analyze atmospheric organic PM with
6 better than 2-h time resolution. These instruments were all present at the Atlanta Supersite
7 experiment during the summer of 1999, and an intercomparison of results is underway.
8 An "in situ carbon analyzer" measured total particulate organic and elemental carbon (i.e.,
9 (j,g of carbon/m3) with 1 to 2 h resolution in Glendora and Claremont, CA, during 1986 and 1987
10 (Turpin and Huntzicker, 1991; Turpin and Huntzicker, 1995), and in Atlanta, GA, during 1999
11 (Supersite experiment, unpublished). By using elemental carbon as a tracer for primary,
12 combustion-generated organic carbon, these authors estimated the contributions of primary
13 sources (i.e., material emitted in particulate form) and secondary sources (i.e., particulate
14 material formed in the atmosphere) to the total atmospheric particulate organic carbon
15 concentrations in these locations. This in situ carbon analyzer collects particulate matter on a
16 quartz fiber filter mounted in a thermal-optical transmittance carbon analyzer (Turpin et al.,
17 1990). The material on a quartz fiber filter behind a Teflon filter in the second sampling port
18 provides an estimate of the positive sampling artifact (i.e., gas adsorption on the quartz sampling
19 filter).
20 An automated carbon analyzer with 15-min to 1-h time resolution is now commercially
21 available (Rupprecht et al., 1995) and has been operated in several locations, including the
22 Atlanta Supersite. It collects samples on a 0.1-//m impactor downstream of an inlet with a
23 2.5-jum cutpoint. Use of an impactor eliminates gas adsorption that must be addressed when
24 filter collection is used. However, this collection system may experience substantial particle
25 bounce, and a sizable fraction of EC is in particles < 0.2 //m. In the analysis, step carbonaceous
26 compounds are removed by heating in filtered ambient air. Carbonaceous material removed
27 below 340 °C is reported as organic carbon, material removed between 340 and 750 °C is
28 reported as elemental carbon. Turpin et al. (2000) comment that it would be more appropriate to
29 report carbon values obtained by this method as "low-" and "high-temperature" carbon, because
30 some organics are known to evolve at temperatures greater than 340 °C (e.g., organics from
31 woodsmoke).
March 2001 2-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 An aethalometer is an automated, time-resolved instrument (i.e., 5- to 15-min sample
2 duration) that measures the light attenuation of aerosol particles collected on a filter tape (Hansen
3 etal., 1984). It is also commercially available. The concentration of elemental carbon is derived
4 from the light absorption measured on a filter using an estimate of the specific absorption (m2/g)
5 of elemental carbon on the filter; the specific absorption value is derived from laboratory and
6 atmospheric tests and is specified by the manufacturer. The specific absorption value could be
7 expected to vary with location, season, and source mix. Comparisons in atmospheric
8 experiments at some locations with EC values measured by thermal methods confirm that the
9 aethalometer provides a statistically meaningful estimate of EC concentration (Allen et al.,
10 1999b; Liousse et al., 1993). For instance, Allen et al. (1999b) found the following statistical
11 relationship for Uniontown, PA, during summer 1990: Black Carbon (aethaometer) = 0.95*EC
12 (thermal) - 0.2 (r2 = 0.925, n not specified but appears to be >50, EC range from 0 to 9 //g/m3).
13 The most recent semi-continuous carbon method is currently being tested by Dr. Susanne
14 Hering (unpublished). This is a flash volatilization method in which particles are impacted on a
15 surface and flash volatilized. Higher collection efficiencies are obtained for smaller particles by
16 growing the particles by humidification prior to impaction. This device was first demonstrated at
17 the Atlanta Supersite.
18
19 2.2.5.9 Determination of Aerosol Surface Area in Real Time
20 Aerosol surface area is an important aerosol property for health effects research. However,
21 methods for on-line measurement of surface area are not widely available. Woo et al. (2000)
22 used three continuous aerosol sensors to determine aerosol surface area. They used a
23 condensation particle counter (CPC, TSI, Inc., Model 3020), an aerosol mass concentration
24 monitor (MCM, TSI, Inc., Model 8520), and an electrical aerosol detector for measuring particle
25 charge concentration (EAD, TSI, Inc., Model 3070). The three sensor signals were inverted to
26 obtain the aerosol size distribution using a lognormal size distribution model by minimizing the
27 difference between the measured signals and the theoretical values based upon a size distribution
28 model, the instrument calibration, and its theoretical responses. The lognormal function was then
29 integrated to calculate the total surface area concentration. Woo et al. demonstrated that this
30 method can give near real-time measurements of aerosol surface area.
31
March 2001 2-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.6 Data Quality
2 Although much of the recent work in instrument development for airborne particulate
3 matter measurement has focused on addressing sampling artifacts associated with loss or gain of
4 semivolatile species, or development of real-time measurements of aerosol concentrations, other
5 recent efforts have examined issues associated with improving the quality of data being collected.
6
7 2.2.6.1 Errors in Gravimetric Analyses
8 Issues of precision and accuracy associated with gravimetric analyses have been examined
9 in several studies. It is well known that weighing of particle sampler filters is subject to
10 fluctuations of environmental conditions in the weighing. Gravimetric analysis issues reviewed
11 by Allen et al. (1999c) include proper temperature and humidity controls, use of a high quality
12 microbalance, 100% replicate weighings, control of static charge, aging of new filters, weighing
13 of a sufficient number of laboratory blank filters, and accounting for buoyancy errors caused by
14 variability in barometric pressure. Lawless and Rodes (1999) investigated the magnitude of
15 uncertainties attributed to fluctuations in some of these parameters (humidity, temperature,
16 drafts, vibration, and electrostatic charges) and recommended methods for improving their
17 control. They noted that the role of humidity control in the laboratory did not seem to be as
18 critical of a factor as the humidity under which the sample was collected. Koistinen et al. (1999)
19 give an excellant discussion of the procedures developed to overcome problems associated with
20 gravimetric measurements of PM25 mass in the EXPOLIS Study. They discuss factors such as
21 corrections for buoyancy, elimination of static charge, and increases in the mass of blank filters
22 with time.
23 Mass concentration measurements of coarse particulate matter are inherently less precise
24 than the corresponding PM2 5 or PM10 measurements (Allen et al., 1999c). Coarse particulate
25 mass concentrations are determined either by difference between collocated PM10 and PM2 5
26 samplers or more directly by use of a dichotomous sampler. The difference method suffers from
27 errors because of the use of two independent measurements. The dichotomous sampler also has
28 potential errors caused by postexposure loss of particles from unoiled filters and uncertainties in
29 the coarse mass channel enrichment factor. Allen et al. (1999c) summarized several sampling
30 issues to consider in measuring coarse particulate mass, including the use of identical
31 instrumentation (except cutpoints) such as filter media, filter face velocity, and ambient-filter
March 2001 2-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 temperature differences; common flow measurement devices; use of higher sampler flow rates
2 (10 L/min minimum for 24-h sample is recommended); avoiding excessive filter loading; and
3 full characterization of the cutpoint characteristics of the FRM without the PM25 WINS inlet.
4
5 2.2.6.2 Quality Assurance
6 EPA has undertaken extensive studies to evaluate the quality of the 1999 PM2 5 FRM data
7 including precision and bias statistics, data reporting statistics, etc. Following a peer review,
8 EPA will issue a report that documents the quality assurance (QA) activities that were
9 undertaken for the PM25 environmental data operations for the calendar year January 1, 1999, to
10 December 31, 1999, which was the first year of implementation the PM2 5 monitoring program.
11 The QA report will evaluate the adherence to the quality assurance requirements described in
12 40 CFR Part 58 Appendix A (Code of Federal Regulations, 1999c) and evaluate the data quality
13 indicators of precision, accuracy/bias, completeness, comparability, and detectability. The report
14 also will provide conclusions and recommendations for future improvements.
15
16
17 2.3 SUMMARY
18 Atmospheric particles originate from a variety of sources and possess a range of
19 morphological, chemical, physical, and thermodynamic properties. The composition and
20 behavior of airborne particles are linked with those of the surrounding gas. Aerosol may be
21 defined as a suspension of solid or liquid particles in air and includes both the particles and all
22 vapor or gas phase components of air. However, the term aerosol often is used to refer to the
23 suspended particles only. Particulate is an adjective and should only be used as a modifier, as in
24 particulate matter.
25 A complete description of the atmospheric aerosol would include an accounting of the
26 chemical composition, morphology, and size of each particle and the relative abundance of each
27 particle type as a function of particle size. Recent developments in single particle analysis
28 techniques are bringing such a description closer to reality.
29 The diameter of a spherical particle may be determined geometrically, from optical or
30 electron microscopy; by light scattering and Mie theory; or by its behavior, such as its electrical
31 mobility or its aerodynamic behavior. However, the various types of diameters may be different,
March 2001 2-98 DRAFT-DO NOT QUOTE OR CITE
-------
1 and atmospheric particles often are not spherical. Therefore, particle diameters are described by
2 an "equivalent" diameter. Aerodynamic diameter (i.e., the diameter of a unit density sphere that
3 would have the same terminal settling velocity as the real particle) is the most widely used
4 equivalent diameter. Therefore, in this document, particle diameters, unless otherwise indicated,
5 refer to the aerodynamic diameter.
6 Atmospheric size distributions show that most atmospheric particles are quite small, below
7 0.1 //m, whereas most of the particle volume (and therefore most of the mass) is found in
8 particles greater than 0.1 //m. An important feature of the mass or volume size distributions of
9 atmospheric aerosols is their multimodal nature. Volume distributions, measured in ambient air
10 in the United States, almost always are found to be bimodal, with a minimum between 1.0 and
11 3.0 /u,m. The distribution of particles that are mostly larger than the minimum is termed the
12 coarse mode. The distribution of particles that are mostly smaller than the minimum is termed
13 the fine mode. Fine-mode particles include both the accumulation mode and the nuclei mode.
14 Accumulation-mode particles are that portion of the fine particle fraction with diameters above
15 about 0.1 //m. The nuclei mode, that portion of the fine particle fraction with diameters below
16 about 0.1 //m, can be observed as a separate mode in mass or volume distributions only in clean
17 or remote areas or near sources of new particle formation by nucleation. Toxicologists and
18 epidemiologists use ultrafme to refer to particles in the nuclei-mode size range. Aerosol
19 physicists and material scientists tend to use nanoparticles to refer to particles generated in the
20 laboratory in this size range.
21 The aerosol community uses four different approaches or conventions in the classification
22 of particles by size: (1) modes, based on the observed size distributions and formation
23 mechanisms; (2) cut point, usually based on the 50% cut point of the specific sampling device
24 (i.e., the particle size at which 50% of the particles enter and 50% of the particles are rejected);
25 (3) dosimetry or occupational sizes, based on the entrance into various compartments of the
26 respiratory system; and (4) legally specified, regulatory sizes for air quality standards. Over the
27 years, the terms fine and coarse, as applied to particle sizes, have lost the original precise
28 meaning of fine mode and coarse mode. In any given article, therefore, the meaning of fine and
29 coarse, unless defined, must be inferred from the author's usage. In particular, PM25 and
30 fine-mode particles are not equivalent. In this document, the term "mode" is used with fine and
March 2001 2-99 DRAFT-DO NOT QUOTE OR CITE
-------
1 coarse when it is desired to specify the distribution of fine-mode particles or coarse-mode
2 particles as shown in Figures 2-4 and 2-5.
3 Several processes influence the formation and growth of particles. New particles may be
4 formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
5 material condenses onto existing particles. Particles may also grow by coagulation as two
6 particles combine to form one. Gas phase material condenses preferentially on smaller particles
7 and the rate constant for coagulation of two particles decreases as the particle size increases.
8 Therefore, nuclei mode particles grow into the accumulation mode but accumulation mode
9 particles do not grow into the coarse mode.
10 The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
11 ions; particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
12 material. Atmospheric PM contains a large number of elements in various compounds and
13 concentrations and hundreds to thousands of specific organic compounds. Particulate material
14 can be primary or secondary. PM is called primary if it is in the same chemical form in which it
15 was emitted into the atmosphere. PM is called secondary if it is formed by chemical reactions in
16 the atmosphere. Primary coarse particles are usually formed by mechanical processes. Primary
17 fine particles are emitted from sources, either directly as particles or as vapors that rapidly
18 condense to form particles.
19 Most of the sulfate and nitrate and a portion of the organic compounds in atmospheric
20 particles are secondary (i.e., they are formed by chemical reactions in the atmosphere).
21 Secondary aerosol formation depends on numerous factors including the concentrations of
22 precursors; the concentrations of other gaseous reactive species such as ozone, hydroxyl radical,
23 peroxy radicals, or hydrogen peroxide; atmospheric conditions, including solar radiation and
24 relative humidity; and the interactions of precursors and preexisting particles within cloud or fog
25 droplets or on or in the liquid film on solid particles. As a result, it is considerably more difficult
26 to relate ambient concentrations of secondary species to sources of precursor emissions than it is
27 to identify the sources of primary particles.
28 The lifetimes of particles vary with particle size. Coarse particles can settle rapidly from
29 the atmosphere within minutes or hours, and normally travel only short distances. However,
30 when mixed high into the atmosphere, as in dust storms, the smaller-sized, coarse-mode particles
31 may have longer lives and travel greater distances. Accumulation-mode fine particles are kept
March 2001 2-100 DRAFT-DO NOT QUOTE OR CITE
-------
1 suspended by normal air motions and have very low deposition rates to surfaces. They can be
2 transported thousands of kilometers and remain in the atmosphere for a number of days.
3 Accumulation-mode particles are removed from the atmosphere primarily by cloud processes.
4 Coarse mode particles of less than ~ 10-//m diameter as well as accumulation-mode and
5 nuclei-mode (or ultrafme) particles all have the ability to penetrate deep into the lungs and be
6 removed by deposition in the lungs. Dry deposition rates are expressed in terms of a deposition
7 velocity that varies with the particle size, reaching a minimum between 0.1 and 1.0 //m
8 aerodynamic diameter.
9 The role of particles in reducing visibility and affecting radiative balance through scattering
10 and absorption of light is evident as are the effects of particles in soiling and damaging materials.
11 EPA addresses visibility effects through regional haze regulations. The direct effects of particles
12 in scattering and absorbing light and the indirect effects of particles on clouds are being
13 addressed in climate change programs in several government agencies.
14 The role of PM in acid deposition has not always been recognized. Acid deposition and
15 PM are intimately related, however, first, because particles contribute significantly to the
16 acidification of rain and, second, because the gas-phase species that lead to dry deposition of
17 acidity are also precursors of particles. Therefore, reductions in SO2 and NOX emissions will
18 decrease both acid deposition and PM concentrations. Sulfuric acid, ammonium nitrate, and
19 organic particles also are deposited on surfaces by dry deposition. The utilization of ammonium
20 by plants leads to the production of acidity. Therefore, dry deposition of particles also can
21 contribute to the ecological damages caused by acid deposition.
22 The decision by the EPA to revise the PM standards by adding daily and yearly standards
23 for PM2 5 has led to a renewed interest in the measurement of atmospheric particles and also to a
24 better understanding of the problems in developing precise and accurate measurements of
25 particles. Unfortunately, it is very difficult to measure and characterize particles suspended in
26 the atmosphere.
27 PM monitoring is needed to develop information to guide implementation of standard (i.e.,
28 by identifying sources of particles; to determine whether or not a standard has been attained; and
29 to determinate health, ecological, and radiative effects). Federal Reference Methods (FRM)
30 specify techniques for measuring PM10 and PM25. Particles are collected on filters and mass
31 concentrations are determined gravimetrically. The PM2 5 FRM sampler consists of a PM10
March 2001 2-101 DRAFT-DO NOT QUOTE OR CITE
-------
1 inlet/impactor, a PM2 5 impactor with an oil-covered impaction substrate to remove particles
2 larger than 2.5 //m, and a 47-mm polytetrafluoroethylene (PTFE) filter with a particle collection
3 efficiency greater than 99.7%. Both techniques provide relatively precise (±10 %) methods for
4 determining the mass of material remaining on a Teflon filter after equilibration. The goal of a
5 PM indicator might be to accurately measure the material that exists as a particle in the
6 atmosphere. However, numerous uncertainties exist as to the relationship between the mass and
7 composition of material remaining on the filter, as measured by the FRMs, and the mass and
8 composition of material that exists in the atmosphere as suspended PM. It is currently not
9 possible to accurately characterize the material that exists as a particle in the atmosphere. There
10 is no reference standard for particles suspended in the atmosphere; there is no accepted way to
11 remove particle-bound water without losing some of the semivolatile components of PM, such as
12 ammonium nitrate and semivolatile organic compounds and particle-bound water. It also is
13 difficult to cleanly separate fine-mode and coarse-mode PM. As a result, EPA defines accuracy
14 for PM measurements in terms of agreement of a candidate sampler with a reference sampler.
15 Therefore, intercomparisons of samplers become very important in determining how well various
16 samplers agree and how various design choices influence what is actually measured.
17 Fine-mode and coarse-mode particles differ not only in size and morphology (e.g., smooth
18 droplets versus rough solid particles), but also in formation mechanisms; sources; and chemical,
19 physical, and biological properties. It is desirable to separate fine-mode PM and coarse-mode
20 PM as cleanly as possible in order to properly allocate health effects to either fine-mode PM or
21 coarse-mode PM and to correctly determine sources by factor analysis or chemical mass balance.
22 In areas with high concentrations of wind-blown soil, the current practice of separating fine- and
23 coarse-mode particles at 2.5-//m AD may not provide the best separation of exposure,
24 epidemiologic, and source apportionment studies. A cut near 1 //m would provide a good
25 indicator of fine-mode PM if the air stream could be dehumidified to remove particle-bound
26 water without evaporating semivolatile components.
27 Current filtration-based mass measurements lead to significant evaporative losses, during
28 and possibly after collection, of a variety of semivolatile components (i.e., species that exist in
29 the atmosphere in dynamic equilibrium between the condensed phase and gas phase). Important
30 examples include ammonium nitrate, semivolatile organic compounds, and particle-bound water.
31 Loss of these components may significantly impact the quality of the measurement, and can lead
March 2001 2-102 DRAFT-DO NOT QUOTE OR CITE
-------
1 to both positive and negative sampling artifacts. Negative artifacts, resulting from loss of
2 ammonium nitrate and semivolatile organic compounds, may occur during sampling because of
3 changes in temperature, relative humidity, or composition of the aerosol, or because of the
4 pressure drop across the filter. Negative artifacts also may occur during handling and storage
5 because of evaporation. Positive artifacts occur when gas-phase compounds (H2O, HNO3, SO2,
6 and organic compounds) absorb onto or react with filter media or collected PM, or when some
7 particle-bound water is not removed.
8 The loss of particulate nitrate may be determined by comparing nitrate collected on a
9 Teflon filter to that collected on a nylon filter (which absorbs nitrate) preceded by a denuder to
10 remove nitric acid. In two studies in southern California, the PM2 5 mass lost because of
11 volatilization of ammonium nitrate was found to represent 10 to 20% of the total PM2 5 mass and
12 almost a third of the nitrate. Semivolatile organic compounds (SVOCs) similarly can be lost
13 from Teflon filters because of volatilization during or after collection. Such losses can cause the
14 PM25 mass to be significantly underestimated. Positive sampling artifacts also can occur as the
15 result of the adsorption of organic gases onto the filter materials. There is a larger positive
16 artifact caused by adsorption of organic vapor onto quartz fiber filters than onto Teflon filters.
17 Denuder-based sampling systems also have been developed for measuring parti culate phase
18 organic compounds. This technique is an improvement over the filter/adsorbent collection
19 method. In most denuder systems, a denuder that removes gas-phase absorbable organic gases is
20 followed by a filter pack. The first filter collects particles. It is followed by a charcoal-
21 impregnated glass-fiber filter that absorbs semivolatile material that evaporates from particles on
22 the front filter. The FRM for PM25 will suffer loss of parti culate nitrates and SVOC, similar to
23 the losses experienced with other single filter collection systems.
24 It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
25 compounds as part of parti culate matter mass. However, it is usually desirable to remove the
26 particle-bound water before determining the mass. In some situations it may be important to
27 know how much of the suspended particle's mass or volume results from particle-bound water.
28 Calculations and measurements indicate that aerosol water content is strongly dependent on
29 composition, but that liquid water can represent a significant mass fraction of the aerosol
30 concentration at relative humidities above 60%. Relative humidity may affect particle size,
31 growth, and other properties. Accumulation-mode particles are usually hygroscopic. The more
March 2001 2-103 DRAFT-DO NOT QUOTE OR CITE
-------
1 hygroscopic particles tend to contain more sulfates, nitrates, and secondary organic compounds,
2 while the less hygroscopic particles tend to contain more elemental carbon, primary organic
3 compounds, and crustal components. Fresh, submicron-size soot particles may to shrink with
4 increasing relative humidity because of a structural change. The effects of relative humidity on
5 the sorption of SVOC on particles are not well understood. The amount of water sorbed to an
6 atmospheric aerosol may be affected by the presence of an organic film on the particle, which
7 may impede the transport of water across the surface.
8 In addition to FRM sampling of equilibrated mass to determine compliance with PM
9 standards, EPA requires states to conduct speciation sampling primarily to determine source
10 categories and trends. The current speciation samplers include three filters: (1) Teflon for
11 equilibrated mass and elemental analysis, (2) a Nylon filter with a nitric acid denuder to collect
12 nitrate, and (3) a quartz fiber filter for elemental and organic carbon (but without any correction
13 for positive or negative artifacts because of adsorption of volatile organic compounds on the
14 quartz filters or evaporation of semivolatile organic compounds from the collected particles).
15 The IMPROVE network of samplers provides four 24-h integrated filter samples: (1) a PM10
16 filter and (2) a PM25 Teflon filter for gravimetric determination of mass and for analysis of heavy
17 elements by X-ray fluorescence; (3) a Nylon filter, preceded by a nitric acid denuder, for artifact-
18 free determination of nitrate and measurement of other ionic species by ion chromatography; and
19 (4) a quartz filter for measurement of elemental carbon (EC) and organic carbon (OC) by thermal
20 optical analysis. The EC/OC measurement method utilized in the IMPROVE network is based
21 on optical correction of pyrolytic char using optical reflectance, whereas the EC/OC method
22 specified in the NIOSH method 5040 (for diesel soot) is based on optical transmission for
23 correction for pyrolytic char. These methods also differ in their temperature profiles. The two
24 methods agree on total carbon but differ in the split of total carbon into EC and OC.
25 The EPA expects that more than 200 local agency monitoring sites throughout the states
26 will operate continuous PM monitors. However, EPA has not yet provided any guidance
27 regarding appropriate continuous monitoring techniques. All currently available techniques for
28 continuous measurements of suspended particle mass, such as the integrating nephelometer, the
29 beta-absorption monitor, and the Tapered Element Oscillating Microbalance (TEOM) share the
30 problem of dealing with semivolatile PM components (i.e., so as not to include particle-bound
31 water as part of the mass, the particle-bound water must be removed by heating or
March 2001 2-104 DRAFT-DO NOT QUOTE OR CITE
-------
1 dehumidification). However, heating also causes ammonium nitrate and semivolatile organic
2 compounds to evaporate. The TEOM monitor operates at a constant, but higher than ambient,
3 temperature to remove particle-bound water. However, the FRM is required to operate at no
4 more than 5 °C above the ambient temperature. This philosophical difference in operation leads
5 to differences between the TEOM and integrated mass concentrations for both PM10 and PM2 5.
6 Several candidates for continuous PM mass measurements are currently being field tested.
7 The Real-Time Total Ambient Mass Sampler (RAMS) measures the total mass of collected
8 particles, including semivolatile species with a TEOM monitor using a "sandwich filter". The
9 sandwich contains a Teflon-coated particle-collection filter followed by a charcoal-impregnated
10 filter to collect any semivolatile species lost from the particles during sampling. The RAMS uses
11 a Nafion dryer to remove particle-bound water from the suspended particles and a particle
12 concentrator to reduce the quantity of gas phase organic compounds that must be removed by the
13 denuder. The Continuous Ambient Mass Monitor (CAMM) estimates ambient particulate matter
14 mass by measurement of the increase in the pressure drop across a membrane filter caused by
15 particle loading. It also uses a Nafion dryer to remove particle-bound water. In addition to
16 continuous mass measurement, a number of techniques for continuous measurement of sulfate,
17 nitrate, or elements are being tested.
18 Aerosol time-of-flight mass spectroscopy (ATOFMS) provides a new technique for
19 real-time measurement of correlated size and composition profiles of individual atmospheric
20 aerosol particles. Measurements are made in situ by combining a dual-laser aerodynamic particle
21 sizing system to size and track individual particles through the instrument and laser
22 desorption/ionization time-of-flight mass spectrometry to obtain correlated single particle
23 composition data. ATOFMS technology is able to size, count, and provide chemical composition
24 on individual particles ranging in size from 10 nm to 2 //m. However, there is still controversy
25 over the calibration of such techniques.
March 2001 2-105 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Allen, G.; Sioutas, C.; Koutrakis, P.; Reiss, R.; Lurmann, F. W.; Roberts, P. T. (1997) Evaluation of the TEOM
3 method for measurement of ambient paniculate mass in urban areas. J. Air Waste Manage. Assoc.
4 47:682-689.
5 Allen, J. O.; Paschkewitz, J. S.; Plummer, E. F.; Lafleur, A. L.; Sarofim, A. F.; Smith, K. A. (1999a) Absorption of
6 semi-volatile compounds in oiled impaction substrates: measurement of pyrene absorption. Aerosol Sci.
7 Technol. 30: 16-29.
8 Allen, G. A.; Lawrence, J.; Koutrakis, P. (1999b) Field validation of a semi-continuous method for aerosol black
9 carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
10 southwestern PA. Atmos. Environ. 33: 817-823.
11 Allen, G. A.; Oh, J. A.; Koutrakis, P. (1999c) Techniques for high quality ambient coarse paniculate mass
12 measurements. J. Air Waste Manage. Assoc. 49: 133-141.
13 Altshuller, A. P. (1983) Review: natural volatile organic substances and their effect on air quality in the United
14 States. Atmos. Environ. 17: 2131-2165.
15 American Conference of Governmental Industrial Hygienists. (1994) 1994-1995 threshold limit values for chemical
16 substances and physical agents and biological exposure indices. Cincinnati, OH: American Conference of
17 Governmental Industrial Hygienists.
18 Anand, N. K.; McFarland, A. R.; Kihm, K. D.; Wong, F. S. (1992) Optimization of aerosol penetration through
19 transport lines. Aerosol Sci. Technol. 16: 105-112.
20 Andrews, E.; Larson, S. M. (1993) Effect of surfactant layers on the size changes of aerosol particles as a function
21 of relative humidity. Environ. Sci. Technol. 27: 857-865.
22 Ansari, A. S.; Pandis, S. N. (1998) Response of inorganic PM to precursor concentrations. Environ. Sci. Technol.
23 32:2706-2714.
24 Appel, B. R. (1993) Atmospheric sample analysis and sampling artifacts. In: Willeke, K.; Baron, P. A., eds. Aerosol
25 measurement: principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 233-259.
26 Appel, B. R.; Tokiwa, Y.; Haik, M. (1981) Sampling of nitrates in ambient air. Atmos. Environ. 15: 283-289.
27 Appel, B. R.; Tokiwa, Y.; Haik, M.; Kothny, E. L. (1984) Artifact paniculate sulfate and nitrate formation on filter
28 media. Atmos. Environ. 18: 409-416.
29 Aumont, B.; Madronich, S.; Bey, I.; Tyndall, G. S. (2000) Contribution of secondary VOC to the composition of
30 aqueous atmospheric particles: a modeling approach. J. Atmos. Chem. 35: 59-75.
31 Ayers, G. P.; Keywood, M. D.; Gras, J. L. (1999) TEOM vs. manual gravimetric methods for determination of
32 PM25 aerosol mass concentrations. Atmos. Environ. 33: 3717-3721.
33 Babich, P.; Davey, M.; Allen, G.; Koutrakis, P. (2000) Method comparisons for paniculate nitrate, elemental
34 carbon, and PM2 5 mass in seven U.S. cities. J. Air Waste Manage. Assoc. 50: 1095-1105.
35 Benner, B. A., Jr.; Wise, S. A.; Currie, L. A.; Klouda, G. A.; Klinedinst, D. B.; Zweidinger, R. B.; Stevens, R. K.;
36 Lewis, C. W. (1995) Distinguishing the contributions of residential wood combustion and mobile source
37 emissions using relative concentrations of dimethylphenanthrene isomers. Environ. Sci. Technol.
38 29:2382-2389.
39 Bergin, M. H.; Ogren, J. A.; Schwartz, S. E.; Mclnnes, L. M. (1997) Evaporation of ammonium nitrate aerosol in a
40 heated nephelometer: implications for field measurements. Environ. Sci. Technol. 31: 2878-2883.
41 Bertoni, G.; Febo, A.; Perrino, C.; Possanzini, M. (1984) Annular active diffuse sampler: a new device for the
42 collection of organic vapours. Ann. Chim. (Rome) 74: 97-104.
43 Binkowski, F. S.; Ching, J. K. S. (1995) Modeling fine paniculate mass and visibility using the EPA regional
44 paniculate model. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research
45 and Development, National Exposure Research Laboratory; report no. EPA/600/A-95/143. Available from:
46 NTIS, Springfield, VA; PB96-139100.
47 Binkowski, F. S.; Shankar, U. (1995) The regional paniculate matter model. I. Model description and preliminary
48 results. J. Geophys. Res. [Atmos.] 100: 26,191-26,209.
49 Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of paniculate diesel
50 exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
51 Brook, J. R.; Darin, T. F. (1999) Contribution of nitrate and carbonaceous species to PM2 5 observed in Canadian
52 cities. J. Air Waste Manage. Assoc. 49: 193-199.
March 2001 2-106 DRAFT-DO NOT QUOTE OR CITE
-------
1 Brorstrom, E.; Grennfelt, P.; Lindskog, A. (1983) The effect of nitrogen dioxide and ozone on the decomposition of
2 particle-associated polycyclic aromatic hydrocarbons during sampling from the atmosphere. Atmos. Environ.
3 17:601-605.
4 Bunz, H.; Koyro, M; Mohler, O.; Saathoff, H. (1996) Temperature dependence of the redistribution mechanisms
5 for volatile components of atmospheric aerosol particles. J. Aerosol Sci. 27(suppl. 1): S653-S654.
6 Carson, P. G.; Johnston, M. V.; Wexler, A. S. (1997) Laser desorption/ionization of ultra fine aerosol particles.
7 Rapid Commun. Mass Spectrom. 11: 993-996.
8 Cachier, H.; Bremond, M.-P.; Buat-Menard, P. (1989) Determination of atmospheric soot carbon with a simple
9 thermal method. Tellus 41B: 379-390.
10 Castaldi, M. J.; Senkan, S. M. (1998) Real-time, ultrasensitive monitoring of air toxics by laser photoionization
11 time-of-flight mass spectroscopy. J. Air Waste Manage. Assoc. 48: 77-81.
12 Castro, L. M.; Pio, C. A.; Harrison, R. M.; Smith, D. J. T. (1999) Carbonaceous aerosol in urban and rural
13 European atmospheres: estimation of secondary organic carbon concentrations. Atmos. Environ.
14 33:2771-2781.
15 Chen, B. T. (1993) Instrument calibration. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles,
16 techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 493-520.
17 Chow, J. C. (1995) Measurement methods to determine compliance with ambient air quality standards for
18 suspended particles. J. Air Waste Manage. Assoc. 45: 320-382.
19 Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
20 thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
21 studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere;
22 April 1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
23 Chow, J. C.; Fujita, E. M.; Watson, J. G.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994) Evaluation of filter-based
24 aerosol measurements during the 1987 Southern California Air Quality Study. Environ. Monit. Assess.
25 30:49-80.
26 Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Countess, R. J. (1996) Sources and chemistry of PM10 aerosol in
27 Santa Barbara County, CA. Atmos. Environ. 30: 1489-1499.
28 Chow, J. C.; Watson, J. G.; Crow, D.; Lowenthal, D. H.; Merrifield, T. (2000) Comparison of IMPROVE and
29 NIOSH carbon measurements. Aerosol Sci. Technol.: accepted.
30 Christoffersen, T. S.; Hjorth, J.; Horie, O.; Jensen, N. R.; Kotzias, D.; Molander, L. L.; Neeb, P.; Ruppert, L.;
31 Winterhalter, R.; Virkkula, A.; Wirtz, K.; larsen, B. R. (1998) cis-Pinic acid, a possible precursor for organic
32 aerosol formation from ozonolysis of "alpha"-pinene. Atmos. Environ. 32: 1657-1661.
33 Claiborn, C. S.; Finn, D.; Larson, T. V.; Koenig, J. Q. (2000) Windblown dust contributes to high PM2 5
34 concentrations. J. Air Waste Manage. Assoc. 50: 1440-1445.
3 5 Clark, W. E.; Whitby, K. T. (1975) Measurements of aerosols produced by the photochemical oxidation of SO2 in
36 air. J. Colloid Interface Sci. 51: 477-490.
37 Clarke, A. D. (1992) Atmospheric nuclei in the remote free-troposphere. J. Atmos. Chem. 14: 479-488.
38 Code of Federal Regulations. (1991a) Appendix J to Part 50~reference method for the determination of paniculate
39 matter as PM10 in the atmosphere. C. F. R. 40: "section" 50.
40 Code of Federal Regulations. (1991b) Ambient air monitoring reference and equivalent methods. C. F. R.
41 40: sect. 53.
42 Code of Federal Regulations. (1999a) Appendix L to Part 50~reference method for the determination of fine
43 paniculate matter as PM2 5 in the atmosphere. C. F. R. 40: "section"50.
44 Code of Federal Regulations. (1999b) Part 53~ambient air monitoring reference and equivalent methods. C. F. R.
45 40: "sections"53, 58.
46 Code of Federal Regulations. (1999c) Part 58~ambient air quality surveillance for paniculate matter: final rule.
47 C. F. R. 40: "section"58.
48 Gotham, W. E.; Bidleman, T. F. (1992) Laboratory investigations of the partitioning of organochlorine compounds
49 between the gas phase and atmospheric aerosols on glass fiber filters. Environ. Sci. Technol. 26: 469-478.
50 Countess, R. J. (1990) Interlaboratory analyses of carbonaceous aerosol samples. Aerosol Sci. Technol.
51 12:114-121.
52 Courtney, W. J.; Shaw, R. W.; Dzubay, T. G. (1982) Precision and accuracy of a "beta" gauge for aerosol mass
53 determinations. Environ. Sci. Technol. 16: 236-239.
54 Coutant, R. W.; Callahan, P. J.; Chuang, J. C.; Lewis, R. G. (1992) Efficiency of silicone-grease-coated denuders
55 for collection of polynuclear aromatic hydrocarbons. Atmos. Environ. Part A 26: 2831-2834.
March 2001 2-107 DRAFT-DO NOT QUOTE OR CITE
-------
1 Covert, D. S.; Kapustin, V. N.; Quinn, P. K.; Bates, T. S. (1992) New particle formation in the marine boundary
2 layer. J. Geophys. Res. [Atmos.] 97: 20,581-20,589.
3 Cui, W.; Machir, J.; Lewis , L.; Eatough, D. J.; Eatough, N. L. (1997) Fine paniculate organic material at Meadview
4 during the project MOHAVE summer intensive study. J. Air Waste Manage. Assoc. 47: 357-369.
5 Cui, W.; Eatough, D. J.; Eatough, N. L. (1998) Fine paniculate organic material in the Los Angeles Basin -1:
6 assessment of the high-volume Brigham Young University organic sampling system, BIG BOSS. J. Air
7 Waste Manage. Assoc. 48: 1024-1037.
8 Davis, C. S.; Fellin, P.; Otson, R. (1987) A review of sampling methods for polyaromatic hydrocarbons in air.
9 JAPCA 37: 1397-1408.
10 Ding, Y.; Lee, M. L.; Eatough, D. J. (1998a) The determination of total nitrite and N-nitroso compounds in
11 atmospheric samples. Int. J. Environ. Anal. Chem. 69: 243-255.
12 Ding, Y.; Cui, W.; Lee, M. L.; Eatough, D. J. (1998b) Fine paniculate N-nitroso and nitrite organic compounds in
13 the atmosphere. Appl. Occup. Environ. Hyg. 13: 432-438.
14 Eatough, D. J. (1999) BOSS, the Brigham Young University Organic Sampling System: determination of paniculate
15 carbonaceous material using diffision denuder sampling technology. In: Lane, D., ed. Gas/phase partition
16 measurements of atmospheric organic compounds. New York, NY: Gordon and Breach Science Publishers;
17 in press.
18 Eatough, D. J.; Wadsworth, A.; Eatough, D. A.; Crawford, J. W.; Hansen, L. D.; Lewis, E. A. (1993)
19 A multiple-system, multi-channel diffusion denuder sampler for the determination of fine-particulate organic
20 material in the atmosphere. Atmos. Environ. Part A 27: 1213-1219.
21 Eatough, D. J.; Tang, H.; Cui, W.; Machir, J. (1995) Determination of the size distribution and chemical
22 composition of fine paniculate semi-volatile organic material in urban environments using diffusion denuder
23 technology. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution
24 and human mortality and morbidity, part II; Janaury 1994; Irvine, CA. Inhalation Toxicol. 7: 691-710.
25 Eatough, D. J.; Eatough, D. A.; Lewis, L.; Lewis, E. A. (1996) Fine paniculate chemical composition and light
26 extinction at Canyonlands National Park using organic paniculate material concentrations obtained with a
27 multisystem, multichannel diffusion denuder sampler. J. Geophys. Res. [Atmos.] 101: 19,515-19,531.
28 Eatough, D. J.; Obeidi, F.; Pang, Y.; Ding, Y.; Eatough, N. L.; Wilson, W. E. (1999a) Integrated and real-time
29 diffusion denuder sampler for PM2 5. Atmos. Environ. 33: 2835-2844.
3 0 Eatough, D. J.; Pang, Y.; Eatough, N. L. (1999b) Determination of PM2 5 sulfate and nitrate with a PC-BOS S
31 designed for routine sampling for semi-volatile paniculate matter. J. Air Waste Manage. Assoc. 49: 69-75.
32 Falconer, R. L.; Bidleman, T. F.; Gotham, W. E. (1995) Preferential sorption of non- and
33 mono-ortho-polychlorinatedbiphenyls to urban aerosols. Environ. Sci. Technol. 29: 1666-1673.
34 Fang, C. P.; McMurry, P. H.; Marple, V. A.; Rubow, K. L. (1991) Effect of flow-induced relative humidity changes
35 on size cuts for sulfuric acid droplets in the microorifice uniform deposit impactor (MOUDI). Aerosol Sci.
36 Technol. 14: 266-277.
37 Faust, B. C. (1994) Photochemistry of clouds, fogs, and aerosols. Environ. Sci. Technol. 28: 217A-222A.
38 Federal Register. (1987) Revisions to the national ambient air quality standards for paniculate matter. F. R. (July 1)
39 52:24,634-24,669.
40 Federal Register. (1997) National ambient air quality standards for paniculate matter; final rule. F. R. (July 18)
41 62: 38,652-38,752.
42 Federal Register. (1998) National ambient air quality standards for paniculate matter and revised requirements for
43 designation of reference and equivalent methods for PM2 5 and ambient air quality surveillance for paniculate
44 matter; correction. F. R. (February 17) 63: 7710-7718.
45 Finn, D.; Rumburg, B.; Claiborn, C.; Bamesberger, W.; Norris, G.; Koenig, J.; Larson, T. (2000) Sampling artifacts
46 from the use of Teflon filters and denuder tubes during the measurement of airborne carbonaceous paniculate
47 matter. J. Air Waste Manage. Assoc.: submitted.
48 Fitz, D. R. (1990) Reduction of the positive organic artifact on quartz filters. Aerosol. Sci. Technol. 12: 142-148.
49 Forstner, J. H. L.; Flagan, R. C.; Seinfeld, J. H. (1997) Molecular speciation of secondary organic aerosol from
50 photooxidation of the higher alkenes: 1-octene and 1-decene. Atmos. Environ. 31: 1953-1964.
51 Forstner, J. H. L.; Flagan, R. C.; Seinfeld, J. H. (1997) Secondary organic aerosol from the photooxidation of
52 aromatic hydrocarbons: molecular composition. Environ. Sci. Technol. 31: 1345-1358.
53 Friedlander, S. K. (1970) The characterization of aerosols distributed with respect to size and chemical composition.
54 J. Aerosol Sci. 1:295-307.
March 2001 2-108 DRAFT-DO NOT QUOTE OR CITE
-------
1 Friedlander, S. K. (1977) Smoke, dust and haze: fundamentals of aerosol behavior. New York, NY: John Wiley &
2 Sons.
3 Friedlander, S. K.; Yeh, E. K. (1998) The submicron atmospheric aerosol as a carrier of reactive chemical species:
4 case of peroxides. Appl. Occup. Environ. Hyg. 13: 416-420.
5 Fuchs, N. A. (1964) The mechanics of aerosols. New York, NY: Pergamon Press.
6 Fuchs, N. A. (1989) The mechanics of aerosols. New York, NY: Dover Publications.
7 Gallorini, M. (1995) Trace element monitoring in atmospheric pollution processes by neutron activation analysis.
8 Microchem. J. 51: 127-137.
9 Card, E.; Mayer, J. E.; Morrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. (1997) Real-time analysis of
10 individual atmospheric aerosol particles: design and performance of a portable ATOFMS. Anal. Chem.
11 69:4083-4091.
12 Ge, Z.; Wexler, A. S.; Johnston, M. V. (1998) Laser desorption/ionization of single ultrafine multicomponent
13 aerosols. Environ. Sci. Technol. 32: 3218-3223.
14 Glasius, M.; Lahaniati, M.; Calogirou, A.; Di Bella, D.; Jensen, N. R.; Hjorth, J.; Kotzias, D.; Larsen, B. R. (2000)
15 Carboxylic acids in secondary aerosols from oxidation of cyclic monoterpenes by ozone. Environ. Sci.
16 Technol. 34: 1001-1010.
17 Goss, K.-U.; Eisenreich, S. J. (1997) Sorption of volatile organic compounds to particles from a combustion source
18 at different temperatures and relative humidities. Atmos. Environ. 31: 2827-2834.
19 Graedel, T. E.; Goldberg, K. I. (1983) Kinetic studies of raindrop chemistry: 1. inorganic and organic processes.
20 J. Geophys. Res. C: Oceans Atmos. 88: 10,865-10,882.
21 Gray, H. A.; Cass, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. (1984) Elemental and organic carbon
22 particle concentrations: a long-term perspective. Sci. Total Environ. 36: 17-25.
23 Griffin, R. J.; Cocker, D. R., Ill; Seinfeld, J. H. (1999) Incremental aerosol reactivity: application to aromatic and
24 biogenic hydrocarbons. Environ. Sci. Technol. 33: 2403-2408.
25 Grosjean, D. (1983) Polycyclic aromatic hydrocarbons in Los Angeles air from samples collected on Teflon, glass
26 and quartz filters. Atmos. Environ. 17: 2565-2573.
27 Grosjean, D. (1992) In situ organic aerosol formation during a smog episode: estimated production and chemical
28 functionality. Atmos. Environ. Part A 26: 953-963.
29 Grosjean, D.; Parmar, S. S. (1990) Interferences from aldehydes and peroxyacetyl nitrate when sampling urban air
30 organic acids on alkaline traps. Environ. Sci. Technol. 24: 1021-1026.
31 Grosjean, D.; Seinfeld, J. H. (1989) Parameterization of the formation potential of secondary organic aerosols.
32 Atmos. Environ. 23: 1733-1747.
33 Grosjean, D.; Williams, E. L., II; Grosjean, E.; Novakov, T. (1994) Evolved gas analysis of secondary organic
34 aerosols. Aerosol Sci. Technol. 21: 306-324.
3 5 Gundel, L. A.; Lane, D. A. (1999) Sorbent-coated diffusion denuders for direct measurement of gas/particle
36 partitioning by semi-volatile organic compounds. In: Lane, D. A., ed. Gas and particle phase measurements
37 of atmospheric organic compounds. Amsterdam, The Netherlands: Gordon and Breach Science Publishers.
3 8 (Advances in environmental, industrial and process control technologies: v. 2).
39 Gundel, L. A.; Lee, V. C.; Mahanama, K. R. R.; Stevens, R. K.; Daisey, J. M. (1995) Direct determination of the
40 phase distributions of semi-volatile polycyclic aromatic hydrocarbons using annular denuders. Atmos.
41 Environ. 29: 1719-1733.
42 Haagen-Smit, A. J. (1952) Chemistry and physiology of Los Angeles smog. Ind. Eng. Chem. 44: 1342-1346.
43 Hansen, A. D. A.; McMurry, P. H. (1990) An intercomparison of measurements of aerosol elemental carbon during
44 the 1986 carbonaceous species method comparison study. J. Air Waste Manage. Assoc. 40: 894-895.
45 Hansen, A. D. A.; Rosen, H.; Novakov, T. (1984) The aethalometer - an instrument for the real-time measurement
46 of optical absorption by aerosol particles. Sci. Total Environ. 36: 191-196.
47 Henry, R. C.; Lewis, C. W.; Hopke, P. K.; Williamson, H. J. (1984) Review of receptor model fundamentals.
48 Atmos. Environ. 18: 1507-1515.
49 Hering, S.; Cass, G. (1999) The magnitude of bias in the measurement of PM25 arising from volatilization of
50 paniculate nitrate from Teflon filters. J. Air Waste Manage. Assoc. 49: 725-733.
51 Hidy, G. M.; Venkataraman, C. (1996) The chemical mass balance method for estimating atmospheric particle
52 sources in southern California. Chem. Eng. Commun. 151: 187-209.
53 Hildemann, L. M.; Rogge, W. F.; Cass, G. R.; Mazurek, M. A.; Simoneit, B. R. T. (1996) Contribution of primary
54 aerosol emissions from vegetation-derived sources to fine particle concentrations in Los Angeles.
55 J. Geophys. Res. [Atmos.] 101: 19,541-19,549.
March 2001 2-109 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hinds, W. C. (1999) Aerosol technology: properties, behavior, and measurement of airborne particles. 2nd ed.
2 New York, NY: John Wiley & Sons.
3 Hinz, K.-P.; Kaufmann, R.; Spengler, B. (1994) Laser-induced mass analysis of single particles in airborne state.
4 Anal. Chem. 66: 2071-2076.
5 Hitzenberger, R.; Dusek, U.; Berner, A. (1996) Black carbon measurements using an integrating sphere. J. Geophys.
6 Res. [Atmos.] 101: 19,601-19,606.
7 Hitzenberger, R.; Berner, A.; Dusek, U.; Alabashi, R. (1997) Humidity-dependent growth of size-segregated aerosol
8 samples. Aerosol Sci. Technol. 27: 116-130.
9 Hitzenberger, R.; Jennings, S. G.; Larson, S. M; Dillner, A.; Cachier, H.; Galambos, Z.; Roue, A.; Spain, T. G.
10 (1999) Intercomparison of measurement methods for black carbon aerosols. Atmos. Environ. 33:2823-2833.
11 Hoffmann, T.; Odum, J. R.; Bowman, F.; Collins, D.; Klockow, D.; Flagan, R. C.; Seinfeld, J. H. (1997) Formation
12 of organic aerosols from the oxidation of biogenic hydrocarbons. J. Atmos. Chem. 26: 189-222.
13 Izumi, K.; Fukuyama, T. (1990) Photochemical aerosol formation from aromatic hydrocarbons in the presence of
14 NOx. Atmos. Environ. Part A 24: 1433-1441.
15 Jaklevic, J. M.; Gatti, R. C.; Goulding, F. S.; Loo, B. W. (1981) A p-gauge method applied to aerosol samples.
16 Environ. Sci. Technol. 15: 680-686.
17 Jang, M.; Kamens, R. M. (1998) A thermodynamic approach for modeling partitioning of semivolatile organic
18 compounds on atmospheric paniculate matter: humidity effects. Environ. Sci. Technol. 32: 1237-1243.
19 Jang, M.; Kamens, R. M. (1999) Newly characterized products and composition of secondary aerosols from the
20 reaction of"alpha"-pinene with ozone. Atmos. Environ. 33: 459-474.
21 Jang, M.; Kamens, R. M.; Leach, K. B.; Strommen, M. R. (1997) A thermodynamic approach using group
22 contribution methods to model the partitioning of semivolatile organic compounds on atmospheric paniculate
23 matter. Environ. Sci. Technol. 31: 2805-2811.
24 Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
25 and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning.
26 Atmos. Environ. 30: 3825-3835.
27 John, W.; Wall, S. M.; Ondo, J. L. (1988) A new method for nitric acid and nitrate aerosol measurement using the
28 dichotomous sampler. Atmos. Environ. 22: 1627-1635.
29 Johnson, M V.; Wexler, A. S. (1995) MS of individual aerosol particles. Anal. Chem. 67: 721A-726A.
30 Kamens, R.; Odum, J.; Fan, Z.-H. (1995) Some observations on times to equilibrium for semivolatile polycyclic
31 aromatic hydrocarbons. Environ. Sci. Technol. 29: 43-50.
32 Kamens, R. M.; Coe, D. L. (1997) A large gas-phase stripping device to investigate rates of PAH evaporation from
33 airborne diesel soot particles. Environ. Sci. Technol. 31: 1830-1833.
34 Kamens, R.; Jang, M.; Chien, C.-J.; Leach, K. (1999) Aerosol formation from the reaction of "alpha"-pinene and
35 ozone using a gas-phase kinetics-aerosol partitioning model. Environ. Sci. Technol. 33: 1430-1438.
36 Kao, A. S.; Friedlander, S. K. (1995) Frequency distributions of PM10 chemical components and their sources.
37 Environ. Sci. Technol. 29: 19-28.
3 8 Kavouras, I. G.; Koutrakis, P. (2000) Use of polyurethane foam as the impaction substrate/collection medium in
39 conventional inertial impactors. Aerosol Sci. Technol.: accepted.
40 Kavouras, I. G.; Mihalopoulos, N.; Stephanou, E. G. (1998) Formation of atmospheric particles from organic acids
41 produced by forests. Nature (London) 395: 683-686.
42 Kavouras, I. G.; Mihalopoulus, N.; Stephanou, E. G. (1999a) Formation and gas/particle partitioning of
43 monoterpene photooxidation products over forests. Geophys. Res. Lett. 26: 55-58.
44 Kavouras, I. G.; Mihalopoulus, N.; Stephanou, E. G. (1999b) Secondary organic aerosol formation vs primary
45 organic aerosol emission: in situ evidence for the chemical coupling between monoterpene acidic
46 photooxidation products and new particle formation over forests. Environ. Sci. Technol. 33: 1028-1037.
47 Keeler, G. J.; Pirrone, N. (1996) Atmospheric transport and deposition of trace elements to Lake Erie from urban
48 areas. Water Sci. Technol. 33: 259-265.
49 Keeler, G. J.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Pierson, W. R. (1988) Effect of ambient humidity
50 on dichotomous sampler coarse/fine ratios. Atmos. Environ. 22: 1715-1720.
51 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
52 segregated mass concentration data and ultrafine particle number concentrations in urban areas. Atmos.
53 Environ. 33:2907-2913.
54 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (2000) Size distribution and sources of
55 aerosol in Launceston, Australia, during winter 1997. J. Air Waste Manage. Assoc. 50: 418-427.
March 2001 2-110 DRAFT-DO NOT QUOTE OR CITE
-------
1 Khalil, M. A. K.; Rasmussen, R. A. (1992) Forest hydrocarbon emissions: relationships between fluxes and ambient
2 concentrations. J. Air Waste Manage. Assoc. 42: 810-813.
3 Kim, Y. J.; Boatman, J. F.; Gunter, R. L.; Wellman, D. L.; Wilkison, S. W. (1993) Vertical distribution of
4 atmospheric aerosol size distribution over south-central New Mexico. Atmos. Environ. Part A
5 27: 1351-1362.
6 Kim, B. M.; Lester, J.; Tisopulos, L.; Zeldin, M. D. (1999) Nitrate artifacts during PM2 5 sampling in the South
7 Coast Air Basin of California. J. Air Waste Manage. Assoc. 49: 142-153.
8 Kim, B. M.; Cassmassi, J.; Hogo, H.; Zeldin, M. D. (2000) Positive organic carbon artifacts on filter medium during
9 PM2 5 sampling in the South Coast Air Basin. Aerosol. Sci. Technol.: submitted.
10 Kirchstetter, T. W.; Corrigan, C. E.; Novakov, T. (2000) Laboratory and field investigation of the adsorption of
11 gaseous organic compounds onto quartz filters. Atmos. Environ.: Accepted.
12 Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
13 Klouda, G. A.; Klinedinst, D. B.; Steel, E. B.; Benner, B. A., Jr.; Parish, H. J. (1996) Exploring a method to
14 produce an urban dust particle filter standard. J. Aerosol Sci. 27(suppl. 1): S351-S352.
15 Koch, S.; Winterhalter, R.; Uherek, E.; Kolloff, A.; Neeb, P.; Moortgat, G. K. (2000) Formation of new particles in
16 the gas-phase ozonolysis of monoterpenes. Atmos. Environ. 34: 4031-4042.
17 Koistinen, K. J.; Kousa, A.; Tenhola, V.; Hanninen, O.; Jantunen, M. J.; Oglesby, L.; Kuenzli, N.; Georgoulis, L.
18 (1999) Fine particle (PM2 5) measurement methodology, quality assurance procedures and pilot results of the
19 EXPOLIS study. J. Air Waste Manage. Assoc. 49: 1212-1220.
20 Koutrakis, P.; Kelly, B. P. (1993) Equilibrium size of atmospheric aerosol sulfates as a function of particle acidity
21 and ambient relative humidity. J. Geophys. Res. [Atmos.] 98: 7141-7147.
22 Koutrakis, P.; Wolfson, J. M.; Slater, J. L.; Brauer, M.; Spengler, J. D.; Stevens, R. K.; Stone, C. L. (1988a)
23 Evaluation of an annular denuder/filter pack system to collect acidic aerosols and gases. Environ. Sci.
24 Technol. 22: 1463-1468.
25 Koutrakis, P.; Wolfson, J. M.; Spengler, J. D. (1988b) An improved method for measuring aerosol strong acidity:
26 results from a nine-month study in St. Louis, Missouri and Kingston, Tennessee. Atmos. Environ.
27 22: 157-162.
28 Koutrakis, P.; Wolfson, J. M.; Spengler, J. D.; Stern, B.; Franklin, C. A. (1989) Equilibrium size of atmospheric
29 aerosol sulfates as a function of the relative humidity. J. Geophys. Res. [Atmos.] 94: 6442-6448.
30 Koutrakis, P.; Wang, P.-Y.; Wolfson, J. M.; Soutas, C., inventors. (1996) Method and apparatus to measure
31 paniculate matter in gas. U.S. patent 5,571,945. November 5. Available: www.patents.ibm.com [1999,
32 November 24].
33 Krieger, M. S.; Hites, R. A. (1992) Diffusion denuder for the collection of semivolatile organic compounds.
34 Environ. Sci. Technol. 26: 1551-1555.
3 5 Lammel, G.; Novakov, T. (1995) Water nucleation properties of carbon black and diesel soot particles. Atmos.
36 Environ. 29: 813-823.
37 Landsberger, S.; Wu, D. (1993) Improvement of analytical sensitivities for the determination of antimony, arsenic,
3 8 cadmium, indium, iodine, molybdenum, silicon and uranium in airborne paniculate matter by epithermal
39 neutron activation analysis. J. Radioanal. Nucl. Chem. 167: 219-225.
40 Landsberger, S.; Wu, D.; Vermette, S. J.; Cizek, W. (1997) Intercomparison of IAEA airborne paniculate matter
41 reference material. J. Radioanal. Nucl. Chem. 215: 117-127.
42 Lane, D. A.; Johnson, N. D.; Barton, S. C.; Thomas, G. H. S.; Schroeder, W. H. (1988) Development and
43 evaluation of a novel gas and particle sampler for semivolatile chlorinated organic compounds in ambient air.
44 Environ. Sci. Technol. 22: 941-947.
45 Lavanchy, V. M. H.; Gaggeler, H. W.; Nyeki, S.; Baltensperger, U. (1999) Elemental carbon (EC) and black carbon
46 (BC) measurements with a thermal method and an aethalometer at the high-alpine research station
47 Jungfraujoch. Atmos. Environ. 33: 2759-2769.
48 Lawless, P. A.; Rodes, C. E.; Evans, G.; Sheldon, L.; Creason, J. (2000) Aerosol concentrations during the 1999
49 Fresno Exposure Studies as functions of size, season and meteorology. Aerosol Sci. Technol.: accepted.
50 Lawrence, J. E.; Koutrakis, P. (1994) Measurement of atmospheric formic and acetic acids: methods evaluation and
51 results from field studies. Environ. Sci. Technol. 28: 957-964.
52 Lawrence, J. E.; Koutrakis, P. (1996a) Measurement and speciation of gas and paniculate phase organic acidity in
53 an urban environment: 1. analytical. J. Geophys. Res. [Atmos.] 101: 9159-9169.
54 Lawrence, J.; Koutrakis, P. (1996b) Measurement and speciation of gas and paniculate phase organic acidity in an
55 urban environment: 2. speciation. J. Geophys. Res. [Atmos.] 101: 9171-9184.
March 2001 2-111 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lawson, D. R. (1990) The Southern California Air Quality Study. J. Air Waste Manage. Assoc. 40: 156-165.
2 Leach, K. B.; Kamens, R. M; Strommen, M. R.; Jang, M. (1999) Partitioning of semivolatile organic compounds in
3 the presence of a secondary organic aerosol in a controlled atmosphere. J. Atmos. Chem. 33: 241-264.
4 Lee, W.-M. G.; Shih, P.-M.; Wang, C.-S. (1997) The influence of relative humidity on the size of atmospheric
5 aerosol. J. Environ. Sci. Health Part A: Environ. Sci. Eng. Toxic Hazard. Subst. Control 32: 1085-1097.
6 Lewtas, J.; Pang, Y.; Booth, D.; Reimer, S.; Eatough, D. J.; Gundel, L. A. (2001) Comparison of sampling methods
7 for semi-volatile organic carbon (SVOC) associated with PM2 5. Aerosol. Sci. Technol.: submitted.
8 Li, W.; Montassier, N.; Hopke, P. K. (1992) A system to measure the hygroscopicity of aerosol particles. Aerosol
9 Sci. Technol. 17: 25-35.
10 Liang, Z.; Chan, C. K. (1997) A fast technique for measuring water activity of atmospheric aerosols. Aerosol Sci.
11 Technol. 26: 255-268.
12 Liang, C.; Pankow, J. F.; Odum, J. R.; Seinfeld, J. H. (1997) Gas/particle partitioning of semivolatile organic
13 compounds to model inorganic, organic, and ambient smog aerosols. Environ. Sci. Technol. 31: 3086-3092.
14 Lindskog, A.; Brorstrom-Lunden, E.; Alfheim, I; Hagen, I. (1987) Chemical transformation of PAH on airborne
15 particles by exposure to NO2 during sampling: a comparison between two filter media. Sci. Total Environ.
16 61:51-57.
17 Liousse, C.; Cachier, H.; Jennings, S. G. (1993) Optical and thermal measurements of black carbon aerosol content
18 in different environments: variation of the specific attenuation cross-section, sigma. Atmos. Environ.
19 27A: 1203-1211.
20 Lundgren, D. A.; Burton, R. M. (1995) Effect of particle size distribution on the cut point between fine and coarse
21 ambient mass fractions. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air
22 pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 131-148.
23 Lundgren, D. A.; Hausknecht, B. J.; Burton, R. M. (1984) Large particle size distribution in five U.S. cities and the
24 effect on a new ambient paniculate matter standard (). Aerosol Sci. Technol. 7: 467-473.
25 Mahanama, K. R. R.; Gundel, L. A.; Daisey, J. M. (1994) Selective fluorescence detection of polycyclic aromatic
26 hydrocarbons in environmental tobacco smoke and other airborne particles. Int. J. Environ. Anal. Chem.
27 56:289-309.
28 Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
29 concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
30 Mamane, Y.; Willis, R.; Conner, T. (2000) Evaluation of computer-controlled scanning electron microscopy applied
31 to an ambient urban aerosol sample. Aerosol Sci. Technol.: accepted.
32 Mazurek, M.; Masonjones, M. C.; Masonjones, H. D.; Salmon, L. G.; Cass, G. R.; Hallock, K. A.; Leach, M.
33 (1997) Visibility-reducing organic aerosols in the vicinity of Grand Canyon National Park: properties
34 observed by high resolution gas chromatography. J. Geophys. Res. [Atmos.] 102: 3779-3793.
3 5 McDow, S. R.; Huntzicker, J. J. (1990) Vapor adsorption artifact in the sampling of organic aerosol: face velocity
36 effects. Atmos. Environ. Part A 24: 2563-2571.
37 Mclnnes, L. M.; Quinn, P. K.; Covert, D. S.; Anderson, T. L. (1996) Gravimetric analysis, ionic composition, and
38 associated water mass of the marine aerosol. Atmos. Environ. 30: 869-884.
39 McLaren, R.; Gertler, A. W.; Wittorff, D. N.; Belzer, W.; Dann, T.; Singleton, D. L. (1996) Real-world
40 measurements of exhaust and evaporative emissions in the Cassair tunnel predicted by chemical mass balance
41 modeling. Environ. Sci. Technol. 30: 3001-3009.
42 McMurry, P. H. (2000) A review of atmospheric aerosol measurements. Atmos. Environ. 34: 1959-1999.
43 McMurry, P. H.; Friedlander, S. K. (1979) New particle formation in the presence of an aerosol. Atmos. Environ.
44 13: 1635-1651.
45 McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
46 aerosols. Atmos. Environ. 23: 497-507.
47 Meyer, A. S., Jr.; Boyd, C. M. (1959) Determination of water by titration with coulometrically generated Karl
4 8 Fischer reagent. Anal. Chem. 31:215-219.
49 Meyer, M. B.; Rupprecht, E. G. (1996) Paniculate matter sampling methods: the importance of standardization.
50 J. Aerosol. Sci. 27(suppl. 1): S349-S350.
51 Meyer, M. B.; Rupprecht, E.; Patashnick, H. (1995) Considerations for the sampling and measurement of ambient
52 paniculate matter. In: Paniculate matter: health and regulatory issues: proceedings of an international
53 specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association;
54 pp. 204-210. (A&WMA publication VIP-49).
March 2001 2-112 DRAFT-DO NOT QUOTE OR CITE
-------
1 Meyer, M. B.; Patashnick, H.; Ambs, J. L.; Rupprecht, E. (2000) Development of a sample equilibration system for
2 the TEOM continuous PM monitor. J. Air Waste Manage. Assoc. 50: 1345-1349.
3 Middlebrook, A. M.; Murphy, D. M.; Thomson, D. S. (1998) Observations of organic material in individual marine
4 particles at Cape Grim during the first aerosol characterization experiment (ACE 1). J. Geophys. Res.
5 103: 16475-16483.
6 Mignacca, D.; Stubbs, K. (1999) Effects of equilibration temperature on PM10 concentrations from the TEOM
7 method in the Lower Fraser Valley. J. Air Waste Manage. Assoc. 49: 1250-1254.
8 Miller, F. J.; Gardner, D. E.; Graham, J. A.; Lee, R. E., Jr.; Wilson, W. E.; Bachmann, J. D. (1979) Size
9 considerations for establishing a standard for inhalable particles. J. Air Pollut. Control Assoc. 29: 610-615.
10 Morawska, L.; Thomas, S. B.; Bofinger, N. D.; Neale, D.; Wainwright, D. (1996) Characterisation of airborne
11 paniculate pollution. In: Smith, A., ed. 2000 Then What?: proceedings of the 13th international clean air and
12 environment conference; September; Adelaide, Australia. Eastwood, New South Wales, Australia: Clean Air
13 Society of Australia and New Zealand Inc.; pp. 285-290.
14 Morawska, L.; Bofinger, N. D.; Kocis, L.; Nwankwoala, A. (1998a) Submicrometer and supermicrometer particles
15 from diesel vehicle emissions. Environ. Sci. Technol. 32: 2033-2042.
16 Morawska, L.; Thomas, S.; Bofinger, N.; Wainwright, D.; Neale, D. (1998b) Comprehensive characterization of
17 aerosols in a subtropical urban atmosphere: particle size distribution and correlation with gaseous pollutants.
18 Atmos. Environ. 32: 2467-2478.
19 Mukerjee, S.; Ellenson, W. D.; Lewis, R. G.; Stevens, R. K.; Somerville, M. C.; Shadwick, D. S.; Willis, R. D.
20 (1997) An environmental scoping study in the lower Rio Grande Valley of Texas—Ill, residential
21 microenvironmental monitoring for air, house dust, and soil. Environ. Int. 23: 657-673.
22 Murphy, D. M.; Thomson, D. S. (1997) Chemical composition of single aerosol particles at Idaho Hill: negative ion
23 measurements. J. Geophys. Res. 102: 6353-6368.
24 National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
25 National Academy Press. 3v.
26 Nelson, M. P.; Zugates, C. T.; Treado, P. J.; Casucio, G. S.; Exline, D. L.; Schlaegle, S. F. (2000) Combining
27 Raman chemical imaging scanning electron microscopy (SEM) to characterize ambient fine paniculate.
28 Aerosol Sci. Technol.: accepted.
29 Neubauer, K. R.; Johnson, M. V.; Wexler, A. S. (1998) Humidity effects of the mass spectra of single aerosol
30 particles. Atmos. Environ. 32: 2521-2529.
31 Noble, C. A.; Prather, K. A. (1996) Real-time measurement of correlated size and composition profiles of individual
32 atmospheric aerosol particles. Environ. Sci. Technol. 30: 2667-2680.
33 Noble, C. A.; Prather, K. A. (1998) Single particle characterization of albuterol metered dose inhaler aerosol in near
34 real-time. Aerosol Sci. Technol. 29: 294-306.
35 Norris, G. (1998) Air pollution and the exacerbation of asthma in an arid, western, US city [dissertation]. Seattle,
36 WA: University of Washington. Available from: University Microfilms International, Ann Arbor, MI;
37 publication no. 9836228.
38 Norris, G.; Young-Pong, S. N.; Koenig, J. Q.; Larson, T. V; Sheppard, L.; Stout, J. W. (1999) An association
39 between fine particles and asthma emergency department visits for children in Seattle. Environ. Health
40 Perspect. 107: 489-493.
41 Obeidi, F. D.; Eatough, D. J. (1999) Continuous measurement of semi-volatile fine paniculate mass inProvo, UT.
42 J. Aerosol Sci.: submitted.
43 Odum, J. R.; Hoffmann, T.; Bowman, F.; Collins, D.; Flagan, R. C.; Seinfeld, J. H. (1996) Gas/particle partitioning
44 and secondary organic aerosol yields. Environ. Sci. Technol. 30: 2580-2585.
45 Odum, J. R.; Jungkamp, T. P. W.; Griffin, R. J.; Forstner, H. J. L.; Flagan, R. C.; Seinfeld, J. H. (1997a) Aromatics,
46 reformulated gasoline, and atmospheric organic aerosol formation. Environ. Sci. Technol. 31: 1890-1897.
47 Odum, J. R.; Jungkamp, T. P. W.; Griffin, R. J.; Flagan, R. C.; Seinfeld, J. H. (1997b) The atmospheric
48 aerosol-forming potential of whole gasoline vapor. Science (Washington, DC) 276: 96-99.
49 Ohta, S.; Okita, T. (1990) A chemical characterization of atmospheric aerosol in Sapporo. Atmos. Environ. Part A
50 24A: 815-822.
51 Ohta, S.; Hori, M.; Yamagata, S.; Murao, N. (1998) Chemical characterization of atmospheric fine particles in
52 Sapporo with determination of water content. Atmos. Environ. 32: 1021-1025.
53 Ondov, J. M. (1996) Paniculate tracers for source attribution: potential for application to California's San Joaquin
54 Valley. J. Aerosol Sci. 27(suppl. 1): S687-S688.
March 2001 2-113 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ondov, J. M; Kelly, W. R.; Holland, J. Z.; Lin, Z. C.; Wight, S. A. (1992) Tracing fly ash emitted from a coal-fired
2 power plant with enriched rare-earth isotopes: an urban scale test. Atmos. Environ. Part B 26B: 453-462.
3 Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1991) Aerosol formation in the photooxidation of
4 isoprene and "beta"-pinene. Atmos. Environ. Part A 25: 997-1008.
5 Pandis, S. N.; Harley, R. A.; Cass, G. R.; Seinfeld, J. H. (1992) Secondary organic aerosol formation and transport.
6 Atmos. Environ. Part A 26: 2269-2282.
7 Pang, Y.; Ren, Y.; Obeidi, F.; Hastings, R.; Eatough, D. J.; Wilson, W. E. (2001) Semi-volatile species inPM25:
8 comparison of integrated and continuous samplers for PM2 5 research or monitoring. J. Air Waste Manage.
9 Assoc. 51: 174-185.
10 Pankow, J. F.; Bidleman, T. F. (1991) Effects of temperature, TSP and per cent non-exchangeable material in
11 determining the gas-particle partitioning of organic compounds. Atmos. Environ. Part A 25: 2241-2249.
12 Pankow, J. F.; Storey, J. M. E.; Yamasaki, H. (1993) Effects of relative humidity on gas/particle partitioning of
13 semivolatile organic compounds to urban paniculate matter. Environ. Sci. Technol. 27: 2220-2226.
14 Parmar, S. S.; Grosjean, D. (1990) Laboratory tests of KI and alkaline annular denuders. Atmos. Environ. Part A
15 24:2695-2698.
16 Patashnick, H.; Rupprecht, G.; Ambs, J. L.; Meyer, M. B. (2000) Development of a reference standard for
17 paniculate matter mass in ambient air. Aerosol Sci. Technol.: accepted.
18 Pellizzari, E. D.; Krost, K. J. (1984) Chemical transformations during ambient air sampling for organic vapors.
19 Anal. Chem. 56: 1813-1819.
20 Penner, J. E.; Novakov, T. (1996) Carbonaceous particles in the atmosphere: a historical perspective to the fifth
21 international conference on carbonaceous particles in the atmosphere. J. Geophys. Res. [Atmos.]
22 101: 19,373-19,378.
23 Petzold, A.; Kopp, C.; Niessner, R. (1997) The dependence of the specific attenuation cross-section on black carbon
24 mass fraction and particle size. Atmos. Environ. 31: 661-672.
25 Pilinis, C.; Seinfeld, J. H.; Grosjean, D. (1989) Water content of atmospheric aerosols. Atmos. Environ.
26 23: 1601-1606.
27 Pinto, J. P.; Pleasant, M.; Kellogg, R.; Torok, S. Z.; Stiles, D.; Willis, R. D.; Koutrakis, P.; Allen, G. A. (1995)
28 Chemical characterization of the ambient aerosol in Philadelphia. Research Triangle Park, NC: U.S.
29 Environmental Protection Agency, National Center for Environmental Assessment.
30 Pinto, J. P.; Stevens, R. K.; Willis, R. D.; Kellogg, R.; Mamane, Y.; Novak, J.; Santroch, J.; Benes, I.; Lenicek, J.;
31 Bures, V. (1998) Czech air quality monitoring and receptor modeling study. Environ. Sci. Technol.
32 32: 843-854.
33 Pio, C.; Alves, C.; Duarte, A. (2001) Organic components of aerosols in a forested area of central Greece. Atmos.
34 Environ. 35: 389-401.
3 5 Pirrone, N.; Keeler, G. J. (1996) The Rouge River watershed pollution by trace elements: atmospheric depositions
36 and emission sources. Water Sci. Technol. 33: 267-275.
37 Poore, M. W. (2000) Oxalic acid in PM2 s paniculate matter in California. J. Air Waste Manage. Assoc.
38 50: 1874-1875.
39 Pun, B. K.; Seigneur, C.; Grosjean, D.; Saxena, P. (2000) Gas-phase formation of water-soluble organic compounds
40 in the atmosphere: a retrosynthetic analysis. J. Atmos. Chem. 35: 199-223.
41 Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. (1998) Real-time characterization of the organic
42 composition and size of individual diesel engine smoke particles. Environ. Sci. Technol. 32: 2672-2679.
43 Reist, P. C. (1984) Introduction to aerosol science. New York, NY: Macmillan Publishing Company.
44 Reist, P. C. (1993) Introduction to aerosol science. New York, NY: McGraw-Hill.
45 Risse, U.; Flammenkamp, E.; Kettrup, A. (1996) Determination of aromatic hydrocarbons in air using diffusion
46 denuders. Fresenius J. Anal. Chem. 356: 390-395.
47 Roberts, R. A.; Corkill, J. (1998) Grass seed field smoke and its impact on respiratory health. J. Environ. Health 60
48 (June): 10-16.
49 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine organic
50 aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
51 Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993a) Quantification of urban
52 organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos. Environ. Part
53 A 27: 1309-1330.
March 2001 2-114 DRAFT-DO NOT QUOTE OR CITE
-------
1 Rogge, W. F.; Hildemann, L. M; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
2 aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci.
3 Technol. 27:636-651.
4 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
5 aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks. Environ.
6 Sci. Technol. 27: 1892-1904.
7 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993d) Sources of fine organic
8 aerosol. 4. Paniculate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol.
9 27:2700-2711.
10 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993e) Sources of fine organic
11 aerosol. 5. Natural gas home appliances. Environ. Sci. Technol. 27: 2736-2744.
12 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1994) Sources of fine organic
13 aerosol. 6. Cigarette smoke in the urban atmosphere. Environ. Sci. Technol. 28: 1375-1388.
14 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997a) Sources of fine organic
15 aerosol. 7. Hot asphalt roofing tar pot fumes. Environ. Sci. Technol. 31: 2726-2730.
16 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997b) Sources of fine organic
17 aerosol. 8. Boilers burning no. 2 distillate fuel oil. Environ. Sci. Technol. 31: 2731-2737.
18 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1998) Sources of fine organic
19 aerosol. 9. Pine, oak, and synthetic log combustion in residential fireplaces. Environ. Sci. Technol. 32: 13-22.
20 Rupprecht, G.; Patashnick, H.; Beeson, d. E.; Green, R. N.; Meyer, M. B. (1995) A new automated monitor for the
21 measurement of paniculate carbon in the atmosphere. In: Paniculate matter: health and regulatory issues:
22 proceedings of an international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
23 Management Association; pp. 309-319. (A&WMA publication VIP-49).
24 Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
25 literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem. 24: 57-109.
26 Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. H. (1995) Organics alter hygroscopic behavior of
27 atmospheric particles. J. Geophys. Res. [Atmos.] 100: 18,755-18,770.
28 Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R. (1996) Source apportionment of
29 airborne paniculate matter using organic compounds as tracers. Atmos. Environ. 30: 3837-3855.
30 Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change. New
31 York, NY: John Wiley & Sons, Inc.
32 Shah, J. J.; Johnson, R. L.; Heyerdahl, E. K.; Huntzicker, J. J. (1986) Carbonaceous aerosol at urban and rural sites
33 in the United States. J. Air Pollut. Control Assoc. 36: 254-257.
34 Silva, P. J.; Prather, K. A. (1997) On-line characterization of individual particles from automobile emissions.
3 5 Environ. Sci. Technol. 31: 3074-3080.
36 Sioutas, C.; Koutrakis, P.; Wolfson, J. M. (1994) Particle losses in glass honeycomb denuder samplers. Aerosol Sci.
37 Technol. 21: 137-148.
38 Sioutas, C.; Koutrakis, P.; Wolfson, J. M.; Azaroff, L. S.; Mulik, J. D. (1996a) Field evaluation of a glass
39 honeycomb denuder/filter pack system to collect atmospheric gases and particles. In: Gay, B. W.; Fuerst,
40 R. G.; Jayanty, R. K. M., eds. Measurement of toxic and related air pollutants: proceedings of a conference;
41 May, 1994; Durham, NC. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
42 Research and Development; report no. EPA/600/R-94/136; pp. 426-435.
43 Sioutas, C.; Wang, P. Y.; Ferguson, S. T.; Koutrakis, P.; Mulik, J. D. (1996b) Laboratory and field evaluation of an
44 improved glass honeycomb denuder/filter pack sampler. Atmos. Environ. 30: 885-895.
45 Sioutas, C.; Koutrakis, P.; Wang, P.-Y.; Babich, P.; Wolfson, J. M. (1999) Experimental investigation of pressure
46 drop with particle loading in Nuclepore filters. Aerosol Sci. Technol. 30: 71-83.
47 Solomon, P. A.; Salmon, L. G.; Fall, T.; Cass, G. R. (1992) Spatial and temporal distribution of atmospheric nitric
48 acid and particulate nitrate concentrations in the Los Angeles area. Environ. Sci. Technol. 26: 1596-1601.
49 Speer, R. E.; Barnes, H. M.; Brown, R. (1997) An instrument for measuring the liquid water content of aerosols.
50 Aerosol Sci. Technol. 27: 50-61.
51 Stein, S. W.; Turpin, B. J.; Cai, X.; Huang, P.-F.; McMurry, P. H. (1994) Measurements of relative
52 humidity-dependent bounce and density for atmospheric particles using the DMA-impactor technique.
53 Atmos. Environ. 28: 1739-1746.
54 Stolzenburg, M. R.; Hering, S. V. (2000) Method for the automated measurement of fine particle nitrate in the
55 atmosphere. Environ. Sci. Technol. 34: 907-914.
March 2001 2-115 DRAFT-DO NOT QUOTE OR CITE
-------
1 Strommen, M. R.; Kamens, R. M. (1997) Development and application of a dual-impedance radial diffusion model
2 to simulate the partitioning of semivolatile organic compounds in combustion aerosols. Environ. Sci.
3 Technol. 31:2983-2990.
4 Suarez, A. E.; Caffrey, P. F.; Ondov, J. M.; Thaung, K. C. (1996) Size distribution of Ir-tagged soot from diesel
5 sanitation trucks. J. Aerosol Sci. 27(suppl. 1): S697-S698.
6 Tang, I. N. (1980) Deliquescence properties and particle size change of hygroscopic aerosols. In: Willeke, K., ed.
7 Generation of aerosols and facilities for exposure experiments. Ann Arbor, MI: Ann Arbor Science
8 Publishers, Inc.; pp. 153-167.
9 Tang, H.; Lewis, E. A.; Eatough, D. J.; Burton, R. M.; Farber, R. J. (1994) Determination of the particle size
10 distribution and chemical composition of semi-volatile organic compounds in atmospheric fine particles with
11 a diffusion denuder sampling system. Atmos. Environ. 28: 939-947.
12 Thomson, D. S.; Murphy, D. M. (1994) Analyzing single aerosol particles in real time. Chemtech 24: 30-35.
13 Tolocka, M. P.; Solomon, P. A.; Mitchell, W.; Norris, G. A.; Gemmill, D. B.; Wiener, R. W.; Vanderpool, R. W.;
14 Homolya, J. B.; Rice, J. (2000) East vs. west in the US: chemical characteristics of PM25 during the winter of
15 1999. Aerosol Sci. Technol.: accepted.
16 Tsai, C.-J.; Cheng, Y.-H. (1996) Comparison of two ambient beta gauge PM10 samplers. J. Air Waste Manage.
17 Assoc. 46: 142-147.
18 Tsai, C.-J.; Huang, H.-Y. (1995) Atmospheric aerosol sampling by an annular denuder system and a high-volume
19 PM10 sampler. Environ. Int. 21: 283-291.
20 Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin:
21 a descriptive analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
22 Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
23 primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
24 Turpin, B. J.; Gary, R. A.; Huntzicker, J. J. (1990) An in situ, time-resolved analyzer for aerosol organic and
25 elemental carbon. Aerosol Sci. Technol. 12: 161-171.
26 Turpin, B. J.; Liu, S.-P.; Podolske, K. S.; Gomes, M. S. P.; Eisenreich, S. J.; McMurry, P. H. (1993) Design and
27 evaluation of a novel diffusion separator for measuring gas/particle distributions of semivolatile organic
28 compounds. Environ. Sci. Technol. 27: 2441-2449.
29 Turpin, B. J.; Huntzicker, J. J.; Hering, S. V. (1994) Investigation of organic aerosol sampling artifacts in the Los
30 Angeles basin. Atmos. Environ. 28: 3061-3071.
31 U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
32 Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
33 Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
34 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
35 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
36 U.S. Environmental Protection Agency. (1999) Paniculate matter (PM2 5) speciation guidance. Final draft. Edition 1.
37 Research Triangle Park, NC: Office of Air Quality Planning and Standards. Available:
38 www.epa.gov/ttn/amtic/files/ambient/pm25/spec/specfinl.pdf [2000, October 11].
39 Van Loy, M.; Bahadori, T.; Wyzga, R.; Hartsell, B.; Edgerton, E. (2000) The Aerosol Research and Inhalation
40 Epidemiology Study (ARIES): PM2 5 mass and aerosol component concentrations and sampler
41 intercomparisons. J. Air Waste Manage. Assoc. 50: 1446-1458.
42 Van Vaeck, L.; Van Cauwenberghe, K.; Janssens, J. (1984) The gas-particle distribution of organic aerosol
43 constituents: measurements of the volatilisation artefact in hi-vol cascade impactor sampling. Atmos.
44 Environ. 18: 417-430.
45 Vasilou, J. G.; Sorenson, D.; McMurry, P. H. (1999) Sampling at controlled relative humidity with a cascade
46 impactor. Atmos. Environ. 30: 1049-1056.
47 Vega, E.; Garcia, I.; Apam, D.; Ruiz, M. E.; Barbiaux, M. (1997) Application of a chemical mass balance receptor
48 model to respirable paniculate matter in Mexico City. J. Air Waste Manage. Assoc. 47: 524-529.
49 Veltkamp, P. R.; Hansen, K. J.; Barkley, R. M.; Sievers, R. E. (1996) Principal component analysis of summertime
50 organic aerosols at Niwot Ridge, Colorado. J. Geophys. Res. [Atmos.] 101: 19,495-19,504.
51 Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
52 and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci.
53 Technol. 28: 555-562.
54 Vincent, J. H. (1995) Aerosol science for industrial hygienists. Oxford, United Kingdom: Pergamon.
March 2001 2-116 DRAFT-DO NOT QUOTE OR CITE
-------
1 Vincent, J. H., ed. (1999) Particle size-selective sampling of paniculate air contaminants. Cincinnati, OH: American
2 Conference of Governmental Industrial Hygienists (ACGIH).
3 Vincent, J. H.; Jones, A. D.; Johnston, A. M.; McMillan, C.; Bolton, R. E.; Cowie, H. (1987) Accumulation of
4 inhaled mineral dust in the lung and associated lymph nodes: implications for exposure and dose in
5 occupational lung disease. Ann. Occup. Hyg. 31: 375-393.
6 Wang, P. Y. (1997) Continuous aerosol mass measurement by flow obstruction. Boston, MA: Harvard University,
7 School of Public Health.
8 Wang, H.-C.; John, W. (1988) Characteristics of the Berner impactorfor sampling inorganic ions. Aerosol Sci.
9 Technol. 8: 157-172.
10 Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M.; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994)
11 Chemical mass balance source apportionment of PM10 during the Southern California Air Quality Study.
12 Aerosol Sci. Technol. 21: 1-36.
13 Wedding, J. B.; Weigand, M. A. (1993) An automatic particle sampler with beta gauging. Air Waste 43: 475-479.
14 Weingartner, E.; Burtscher, H.; Baltensperger, U. (1997) Hygroscopic properties of carbon and diesel soot particles.
15 Atmos. Environ. 31: 2311-2327.
16 Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
17 Environ. 33:2301-2312.
18 West, J. J.; Ansari, A. S.; Pandis, S. N. (1999) Marginal PM2.5: nonlinear aerosol mass response to sulfate
19 reductions in the eastern United States. J. Air Waste Manage. Assoc. 49: 1415-1424.
20 Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
21 Whitby, K. T.; Charlson, R. E.; Wilson, W. E.; Stevens, R. K. (1974) The size of suspended particle matter in air.
22 Science (Washington, DC) 183: 1098-1099.
23 Whitby, K. T.; Sverdrup, G. M. (1980) California aerosols: their physical and chemical characteristics. In: Hidy,
24 G. M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. The character and origins of smog
25 aerosols: a digest of results from the California Aerosol Characterization Experiment (ACHEX). New York,
26 NY: John Wiley & Sons, Inc.; pp. 477-517. (Advances in environmental science and technology: v. 9).
27 Willeke, K.; Baron, P. A., eds. (1993) Aerosol measurement: principles, techniques and applications. New York,
28 NY: Van Nostrand Reinhold Publishers.
29 Willeke, K.; Whitby, K. T. (1975) Atmosperic aerosols: size distribution interpretation. J. AirPollut. Control Assoc.
30 25: 529-534.
31 Williams, E. L., II; Grosjean, D. (1990) Removal of atmospheric oxidants with annular denuders. Environ. Sci.
32 Technol. 24:811-814.
33 Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
34 (2000) The 1998 Baltimore paniculate matter epidemiology-exposure study: part 1 - comparison of ambient,
3 5 residential outdoor, indoor and apartment paniculate matter monitoring. J. Exposure Anal. Environ.
36 Epidemiol.: 10: 518ff.
37 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
38 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
39 Wilson, W. E.; Spiller, L. L.; Ellestad, T. G.; Lamothe, P. J.; Dzubay, T. G.; Stevens, R. K.; Macias, E. S.; Fletcher,
40 R. A.; Husar, J. D.; Husar, R. B.; Whitby, K. T.; Kittelson, D. B.; Cantrell, B. K. (1977) General Motors
41 sulfate dispersion experiment: summary of EPA measurements. J. AirPollut. Control Assoc. 27: 46-51.
42 Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; Wilson, W. E. (2000) Use of continuous measurements of integral aerosol
43 parameters to estimate particle surface area. Aerosol Sci. Technol.: accepted.
44 Wu, C. C.; Suarez, A. E.; Lin, Z.; Kidwell, C. B.; Borgoul, P. V.; Caffrey, P. F.; Ondov, J. M.; Sattler, B. (1998)
45 Application of an Ir tracer to determine soot exposure to students commuting to school on Baltimore public
46 buses. Atmos. Environ. 32: 1911-1919.
47 Yamasaki, H.; Kuwata, K.; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne polycyclic
48 aromatic hydrocarbons. Environ. Sci. Technol. 16: 189-194.
49 Yatin, M.; Tuncel, S. G.; Tuncel, G.; Aras, N. K. (1994) Trace element composition of atmospheric aerosols in
50 Ankara, Turkey, determined by instrumental neutron activation analysis. J. Radioanal. Nucl. Chem.
51 181:401-411.
52 Yu, J.; Cocker, D. R., Ill; Griffin, R. J.; Flagan, R. C.; Seinfeld, J. H. (1998) Gas-phase ozone oxidation of
53 monoterpenes: gaseous and paniculate products. J. Atmos. Chem. 34: 207-258.
54 Yu, J.; Griffin, R. J.; Cocker, D. R., Ill; Flagan, R. C.; Seinfeld, J. H. (1999) Observation of gaseous and paniculate
55 products of monoterpene oxidation in forest atmospheres. Geophys. Res. Lett. 26: 1145-1148.
March 2001 2-117 DRAFT-DO NOT QUOTE OR CITE
-------
1 Zhang, X. Q.; McMurry, P. H. (1987) Theoretical analysis of evaporative losses from impactor and filter deposits.
2 Atmos. Environ. 21: 1779-1789.
3 Zhang, X.; McMurry, P. H. (1992) Evaporative losses of fine paniculate nitrates during sampling. Atmos. Environ.
4 Part A 26: 3305-3312.
5 Zhang, S.-H.; Shaw, M.; Seinfeld, J. H.; Flagan, R. C. (1992) Photochemical aerosol formation from"alpha"-pinene
6 and "beta"-pinene. J. Geophys. Res. [Atmos.] 97: 20,717-20,729.
7 Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
8 submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
9 27: 1593-1607.
March 2001 2-118 DRAFT-DO NOT QUOTE OR CITE
-------
i 3. CONCENTRATIONS, SOURCES, AND EMISSIONS
2 OF ATMOSPHERIC PARTICULATE MATTER
3
4
5 3.1 INTRODUCTION
6 This chapter incorporates material from Chapters 5 (Sources and Emissions of Atmospheric
7 Particles) and Chapter 6 (Environmental Concentrations) of the previous document, Air Quality
8 Criteria for Particulate Matter or " 1996 PM AQCD" (U. S. Environmental Protection Agency,
9 1996) and presents updates to these materials where available.
10 Information about concentrations, the composition, and the spatial and temporal variability
11 of ambient particles across the United States is presented in Section 3.2. Ambient concentration
12 data obtained during the first year of operation of the recently deployed nationwide network of
13 Federal Reference Method PM25 monitors are presented. Results of field studies that have
14 characterized the composition of organic compounds in the ambient aerosol are summarized in
15 Appendix 3 A as a complement to the data for the inorganic composition of ambient particles that
16 was presented in Appendix 6A in the 1996 PM AQCD. Data for characterizing the daily and
17 seasonal variability of PM25 concentrations are discussed in Section 3.2.1, the intraday variability
18 of PM25 concentrations in Section 3.2.2, the relations among different size fractions in
19 Section 3.2.3, the interrelations and correlations among PM components in Section 3.2.4, and the
20 spatial variability of various PM components in Section 3.2.5.
21 Unlike gaseous criteria pollutants (SO2, NO2, CO, O3), which are well-defined chemical
22 entities, atmospheric parti culate matter (PM) is composed of a variety of particles differing in
23 size and chemical composition. Therefore, sources of each component of the atmospheric
24 aerosol must be considered in turn. Differences in the composition of particles emitted by
25 different sources also will lead to spatial and temporal heterogeneity in the composition of the
26 atmospheric aerosol. The nature of the sources and the composition of the emissions from these
27 sources are discussed in Section 3.3. Estimates of contributions of various sources to ambient
28 PM levels given by source apportionment studies also are presented in Section 3.3. More
29 detailed information about the composition of emissions from various sources is given in
30 Appendix 3B. Because PM is composed of both primary and secondary constituents, emissions
March 2001 3-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 of both the primary components and the gaseous precursors of secondary PM must be considered.
2 Nationwide emissions estimates of primary PM and precursors to secondary PM are discussed in
3 Section 3.4.1 and uncertainties in emissions estimates in Section 3.4.2.
4 The organization of topics in this chapter (ambient measurements, source characterization
5 and apportionment, and emissions inventories) reflects, in a broad sense, the order in which these
6 topics are addressed in scientific studies and arguably increasing uncertainties in these areas.
7
8
9 3.2 TRENDS AND PATTERNS IN AMBIENT PM2 5 CONCENTRATIONS
10 A significant amount of data for characterizing PM10 mass concentrations and trends exists
11 and that available up to about 1994 was presented in the 1996 PM AQCD. However, data sets
12 for characterizing PM2 5 and PM(10_2 5) mass or trends were not as extensive. Sources of data on
13 PM25 (fine) and PM(10.25) (coarse), which were discussed in the 1996 PM AQCD, include EPA's
14 Aerometric Information Retrieval System (AIRS) (U. S. Environmental Protection Agency,
15 2000a), IMPROVE (Eldred and Cahill, 1994; Cahill, 1996), the California Air Resources Board
16 (CARB) Data Base (California Air Resources Board, 1995), the Harvard Six-Cities Data Base
17 (Spengler et al., 1986; Neas, 1996), and the Harvard Philadelphia Data Base (Koutrakis, 1995).
18 The Inhalable Particulate Network (IPN) (Inhalable Paniculate Network, 1985; Rodes and Evans,
19 1985) provided TSP, PM15, and PM2 5 data but only a small amount of PM10 data.
20 New sources of PM data include the recently deployed nationwide PM25 compliance
21 monitoring network, which provides mass measurements using a Federal Reference Method
22 (FRM). This section summarizes calendar year 1999 data from this network, and provides an
23 approximate characterization of nationwide PM(10.25) by comparing PM10 to PM25 measurements
24 at sites where both types of compliance monitors are located. In addition, a small number of
25 recent studies in which daily mass and composition measurements are available for extended
26 periods will be discussed in this section.
27 Summary tables giving the results of field studies that obtained data for the composition of
28 particles in the PM2 5, PM(10_2 5), or PM10 size ranges were presented in Appendix A to Chapter 6
29 of the 1996 PM AQCD. The summary tables included data for mass, organic carbon, elemental
30 carbon, nitrate, sulfate, and trace elements. The results of 66 studies were separated and
31 presented for the eastern, western, and central United States. It should be noted that these studies
March 2001 3-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 took place at various times and lasted for various durations over a 20-year period. Summary
2 tables showing data for organic carbon, elemental carbon and organic compounds in ambient
3 particles are given in Appendix 3 A. Data for the gross chemical composition of PM25 particles
4 sampled in rural air by the IMPROVE network and for the aerosol composition studies cited in
5 the 1996 PM AQCD (which were carried out mainly in urban areas) are summarized in
6 Table 3-1. Data are shown separately for the eastern and western United States. The IMPROVE
7 data are annual average concentrations for 1998. Quality assured aerosol composition data for
8 urban areas from the PM2 5 speciation network are not yet available for comparison to the
9 IMPROVE data set. Many features are reflected broadly in both data sets (i.e., increasing
10 organic carbon, nitrate, and minerals and decreasing sulfate in going from east to west). The
11 annual average PM2 5 concentration of 11.0 //g/m3 reported by eastern IMPROVE sites is almost
12 three times higher than 3.9 //g/m3 reported by IMPROVE sites in the western United States. The
13 data shown in Table 3-1 refer only to components that have been identified and quantified.
14 There may be other unidentified components that can represent a significant contribution to the
15 total measured mass, as indicated in the 1996 PM AQCD.
16
17
TABLE 3-1. GROSS CHEMICAL COMPOSITION OF PM2 5 PARTICLES
OBTAINED IN RURAL AREAS OF THE EASTERN AND WESTERN
UNITED STATES BY THE IMPROVE NETWORK AND IN MIXED RURAL,
SUBURBAN, AND URBAN AREAS OBTAINED BY STUDIES SUMMARIZED IN
THE 1996 PARTICULATE MATTER AIR QUALITY CRITERIA DOCUMENT
Eastern U.
IMPROVE1
S. Western U.S.
1996 PM AQCD2
Eastern U
% Contribution
S04=
EC
OC
NO3
Crustal
PM,,
56
5
27
5
7
Reconstructed
11.0
33
6
36
8
17
PM2 5 Concentration (^g/m3)
3.9
44
5
27
1
6
PM25
31.0
.S. Western U.S.
% Contribution"
11
14
38
15
14
Concentration (^g/m3)
37.3
"Note that contributions do not add to 100% because a portion of the measured total mass was not characterized chemically.
Sources: (1) IMPROVE network (1998); (2) U.S. Environmental Protection Agency (1996).
March 2001 3-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 PMig Concentrations and Trends
2 Nationwide PM10 annual mean concentrations from AIRS for calendar year 1999 are shown
3 in Figure 3-1. Concentrations in most areas of the country were below the level of the annual
4 PM10 standard (50 //g/m3) in 1999. Exceptions include central South Carolina, Puerto Rico, and
5 several places in the southwestern United States and central California.
6
7
;i^U> ^^:r—~—
/*'--%> vvfo
•** ° "'^V.Or*,
^Xo /DO^ °!at~
< 4 quarters ^ > 50
C one or more quarters with < 11 samples (j 30-50
U All quarters with at least 11 samples U 20 - 30
U All quarters 75% or more complete U °-20
Figure 3-1. 1999 annual mean PM10 concentrations
Source: Fitz-Simons et al. (2000).
1 Nationwide trends in annual mean PM10 concentrations from 1989 through 1998 based on
2 data obtained at 138 rural sites, 355 suburban sites, and 413 urban sites reporting to AIRS are
March 2001 3-4 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
shown in Figure 3-2 (U.S. Environmental Protection Agency, 2000b). Though average
concentration levels differ among sites, with higher levels at urban and suburban sites, the overall
nationwide trend shows a decline. Figure 3-3 shows the annual mean PM10 trend summarized by
EPA region. Decreases were greater in the western United States than in the eastern United
States, ranging from about 20% in the East to about 38% in the Northwest.
35
30
E 25
c 29
O
05
CD
O
c
°
O
15
m
IU
0
N.
•N.
*'
Rural (138 sites)
Suburban (355 sites)
Urban (413 sites)
89 90 91 92
93 94 95 96 97 98
Year
Figure 3-2. Nationwide trend in ambient PM10 concentration from 1989 through 1998.
Source: U.S. Environmental Protection Agency (2000b).
March 2001
3-5
DRAFT-DO NOT QUOTE OR CITE
-------
34,3
f 38%
41,8
f 36%
27.2
33,3
28.8
24,4
19.8
f 31%
f 21%
32.1
f 21%
f 19%
32.2
f.22%
f 19%
30,0
28.7
23.3
f 19%
§18%
The National Trend
31.7
"*->__ 23.7
f 25%
Alaska is in EPA Region 10; Hawaii, EPA Region 9; and Puerto Rico, EPA Region 2.
Concentrations are ug/m3.
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, 1989 through
1998 (Mg/m3).
Source: U. S. Environmental Protection Agency (2000a).
1
2
3
4
5
9
10
11
PM2 5 Concentrations and Trends
Nationwide PM25 concentrations from the 1999 compliance network are shown in
Figures 3-4a and 3-4b. By the end of 1999 the network consisted of over 1025 monitors. Annual
mean PM2 5 concentrations were above 15 //g/m3 in many areas of the country, especially
throughout the eastern United States, and above 20 //g/m3 in several major urban locations. The
98th percentile 24-h average concentrations were generally below 65 //g/m3. Most of the sites
with levels above 65 //g/m3 are located in California. As shown by the size of the dots on the
maps, the picture for 1999 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.
Further, at the time these maps were created some states such as Massachusetts and New
Hampshire had not reported valid data to AIRS from all monitoring sites. It is premature to
March 2001
3-6
DRAFT-DO NOT QUOTE OR CITE
-------
°o VJUCO^ c
'• )• ,S
_d
A - L.1V*—^ IO^T r.—'o
^ c \ '- - \ ^ ^\ ^' ^^"^- ;^_
J2^* x%!&> ^, \ Jf ^ v>: ' '
<^~ x -•• -, T/- --,'O> -<- ,
"• --- -' '-'" 9 >»' • _- -T1 <•-> I,2)
Alaska Hawaii ^^ \ o . o-
"—0^ Puerto Rico
° < 4 quarters ^ > 20
C one or more quarters with < 11 samples (J 15-20
U All quarters with at least 11 samples l^' 10-15
U All quarters 75% or more complete U °-10
Source: US EPA AIRS Data Base as of 7/12/00 without data flagged as 1, 2, 3, 4, T, W, Y, or X.
Figure 3-4a. 1999 annual mean VM2S concentrations
Source: Fitz-Simons et al. (2000).
1 make judgments on whether an area likely will attain the 1997 PM2 5 standards based on this
2 single year of data, not only because the 1999 data is not complete, but also because the 1997
3 standard is defined in terms of 3-year average concentrations.
4 Annual average PM2 5 obtained as part of health studies conducted in various locations in
5 the United States and Canada from the late 1980s to the early 1990s are shown in Figure 3-5
6 (Bahadori et al., 2000a). These studies include the Harvard six-cities study (Steubenville, OH;
7 Watertown, MA; Portage, WI; Topeka, KS; St. Louis, MO; and Kingston-Harriman, TN),
8 PTEAMS (Riverside, CA), MAACS (Philadelphia, PA; Washington, DC; and Nashville, TN),
9 South Boston Air Quality and Source Apportionment Study (Boston, MA); NPMRMN
March 2001 3-7 DRAFT-DO NOT QUOTE OR CITE
-------
-S^s
•aii>^
&
- - -' -f »• '
- -/
•:u
ort) -TV i ^- '. 0 ~ " - •
' '-'-^- 6' - ^f0'" CW<^<
^n_o'&$%^- °-$ar
, y y cs ^o 'J '43^ i ^
il-— ?c4°£^0'^:?^
^ °^, ^^oc^/
'^,
o,
^^o^^-°^?^"
00 °° o b oO-, i^^^i^g^'"°,-,
"v~1"0, ^^ 'o_Uo_:'0 "^ 5?oov^^-' «^v
'W' (,--, -;, -00j oh r '-' ;
|; c^v-.-*;-1!
•"\ i_ t i I "Y ' ^ ' ii
\ '3 ° i^ ~' V ^"" -
65
O one or more quarters with < 75% of scheduled samples O 40-65
quarters with at least 75% of scheduled samples O 30-40
~) 0-30
~- wii\_» ^i inwiw\-|uaii'^io win i ^ fw/u wi \j\ji i\_>v^u iv^\_i ocii i i|_
O All quarters with at least 75% of scheduled samples
Source: US EPAAIRS Data Base as of 7/12/00 without data flagged as 1, 2, 3, 4, T, W, Y, orX.
Figure 3-4b. 1999 98th percentile 24-h average PM2 5 concentrations
Source: Fitz-Simons et al. (2000).
1 (Phoenix, AZ). Remaining sites were part of the 24-cities study. Sufficient data are not yet
2 available to permit the calculation of nationwide trends of PM25 and PM(10_25); however some
3 general conclusions can be reached. Darlington et al. (1997) proposed that because the consistent
4 reductions in PM10 levels were found in a wide variety of environments ranging from urban to
5 rural over large areas, that common factors or controls might be responsible for these reductions,
6 and that these factors affected fine particles more strongly than coarse particles because fine
7 particles can be transported over longer distances. The longest time series of PM2 5 concentration
8 and composition data have been obtained by the California Air Resources Board. Their data
9 show that annual average PM2 5 concentrations decreased about 50% in the South Coast Air
March 2001
3-8
DRAFT-DO NOT QUOTE OR CITE
-------
g.
to
o
o
H
6
O
2
O
H
O
c
o
H
W
O
^
O
HH
H
W
140H
£
? 120-
.2 100-1
c
0
o
c
o
O
o
'•e
en
Q_
80-
60-
40-
20-
0-
PM25 Annual Standard
(15|jg/m3)
• 95%
1
_
-
T
90%
75%
Mean
50%
125%
10%
• 5%
i
\
n i i n
CDCDCT3C COcoCCD>.COCOCDCCCD
T3= ^ & O = '> O = Q) C CO)^ =
i2 ^-S^^5^ ^ni-c"^ en *~ '~ °-^
CDmCcD.E c —I ? CO >
>J2° ^-^ C*JL~; CD-—
r? 13 C ^ {2 CD^^zE
^co^^l °- co £
_ 0) CD
CD "O c
O ^5 to
to
o
CD o
> -r-
T3
CD
X
= Q.
O .O
CD
O CD CD O
o% £¥
LJJ
CD
CL
CD
O
City
Figure 3-5. Annual distribution of 24-h average PM2 5 concentrations observed in U.S. and Canadian health studies.
Source: Bahadori et al. (2000a).
-------
1 Basin, 35% in the San Joaquin Valley, 30% in the San Francisco Bay Area, and 35% in the
2 Sacramento Valley from 1990 to 1995 (Dolislager and Motallebi, 1999). PM2 5 data have been
3 collected continuously since 1994 as part of the children's health study in 12 communities in
4 southern California (Taylor et al., 1998). Data obtained at all sites show decreases ranging from
5 2% at Santa Maria to 37% at San Dimas/Glendora in PM25 from 1994 through 1998. These
6 decreases were accompanied by decreases in major components such as nitrate, sulfate,
7 ammonium, and acids. However, undefined components showed a mixed pattern of increases
8 and decreases at the same sites.
9 In common usage, the term "background concentrations" refers to concentrations observed
10 in remote areas relatively unaffected by local pollution sources. However, as noted in Chapter 6
11 of the 1996 PM AQCD, several definitions of background concentrations are possible.
12 The two definitions chosen in that document as being most relevant for regulatory purposes
13 are based on estimates of contributions from uncontrollable sources that can affect concentrations
14 in the United States. The first definition refers to the concentration resulting from anthropogenic
15 and natural emissions outside North America and natural sources within North America. The
16 second definition refers to the concentration resulting from natural sources only within and
17 outside of North America. Because of long-range transport from anthropogenic source regions in
18 North America, it is impossible to obtain background concentrations defined above solely on the
19 basis of direct measurement in remote areas in North America. Annual average natural
20 background levels of PM10 (according to definition 1) have been estimated to range from 4 to
21 8 //g/m3 in the western United States and 5 to 11 Mg/m3 in the eastern United States.
22 Corresponding PM2 5 levels have been estimated to range from 1 to 4 //g/m3 in the western
23 United States and from 2 to 5 //g/m3 in the eastern United States (U.S. Environmental Protection
24 Agency, 1996). Although the values shown in Table 3-1 are broadly consistent with those given
25 above, the data shown in Table 3-1 represent only upper limits to background concentrations,
26 because of contributions from long-range transport from anthropogenic sources within North
27 America. Peak 24-h average natural background concentrations may be substantially higher than
28 the annual or seasonal average natural background concentrations. Estimates of levels for
29 background 2 are not yet available. However, recent information about contributions to
30 background concentrations that fall under definitions 1 and 2 because of long-range transport
31 from sources outside the United States is given in Section 3.3.2.
March 2001 3-10 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
10
PM(W.2S) Concentrations
Using AIRS data from the 1999 PM10 and PM2 5 compliance networks it is possible to
construct a preliminary picture of coarse PM across the country. This is accomplished by pairing
data from nearly 400 compliance monitoring sites where PM10 and PM2 5 monitors are col-
located, and subtracting the mass concentrations of PM25 from PM10. The results of this simple
difference method are shown in Figure 3-6. Because of potential problems with this approach,
the results should be viewed with caution. Using this approximate method, annual mean
PM(10_2 5) concentrations are as high as 54 //g/m3, with a nationwide median concentration of
10 //g/m3. The higher values occur mainly in the western United States, particularly in
California.
}
Alaska
Hawaii
Puerto Rico
° < 4 quarters 9 > 30
O one or more quarters with < 11 samples O 15-30
O All quarters with at least 11 samples L' 10-15
U All quarters 75% or more complete U 0-10
Source: US EPA AIRS Data Base as of 7/12/00 without data flagged as 1, 2, 3, 4, T, W, Y, orX.
Figure 3-6. 1999 annual mean PM(10_2 5) concentrations (jUg/m3).
Source: Fitz-Simons et al. (2000).
March 2001
3-11
DRAFT-DO NOT QUOTE OR CITE
-------
1 3.2.1 Daily and Seasonal Variability
2 Data for PM2 5 concentrations obtained as part of the nationwide NAMS/ SLAMS network
3 during 1999 are summarized in Figures 3-7a through c for individual sites in selected urban areas
4 across the United States. As far as possible, the cities were chosen because air pollution-health
5 outcome studies had been performed there, and others were added for the sake of geographic
6 coverage. At least two sites within each of the seven aerosol characteristic regions of the United
7 States identified in Chapter 6 of the 1996 PM AQCD and later adopted by the Health Effects
8 Institute for grouping the results of air pollution-epidemiology studies were chosen. The figures
9 show the range of 24-h average values within each calendar quarter as box and whisker plots and
10 the annual average concentrations for 1999 are shown above each figure. Because FRM
11 measurements of PM25 began only in January, data tend to be limited in many areas, especially
12 for the first quarter. As can be seen from the figures, the pattern of seasonal variability for 1999
13 varied across the United States. At all of the sites shown for the eastern United States, except for
14 the site in Miami, FL, highest quarterly mean values and maximum values occurred during the
15 third quarter (summer) of 1999. This pattern was found, in general, at other sites within the same
16 MS As, although there were exceptions. At sites west of the Mississippi River, highest mean
17 values occurred during the first or fourth quarter (winter or autumn) of 1999, except for the site
18 in Kansas City, where the highest quarterly mean and maximum values occurred in the third
19 quarter. Generally, similar patterns of seasonal variability were found at all other sites within
20 MSAs sampled, although there were exceptions, which may have been related to contributions
21 from local sources as opposed to contributions from regional background sources. At the sites in
22 Miami and Puerto Rico, maximum concentrations occurred during the second quarter, and may
23 have been related to the transport of dust from the Sahara Desert. Because of the limited nature
24 of these data, definitive conclusions regarding long-term patterns of seasonal variability cannot
25 be drawn from these data alone.
26 Longer time series for making more definitive statements about seasonal variations in PM2 5
27 concentrations are available from a few studies which have had as their goal the characterization
28 of PM25 and PM10 concentrations in major urban areas. The Metropolitan Acid Aerosol
29 Characterization Study (MAACS) (Bahadori et al., 2000b) characterized the levels and the
30 spatial and temporal variability of PM25, PM10, and acidic sulfate concentrations in four cities in
31 the eastern United States (Philadelphia, PA; Washington, D.C.; Nashville, TN; and Boston, MA).
March 2001 3-12 DRAFT-DO NOT QUOTE OR CITE
-------
Bangor,
ME Philadelphia, PA
8.9|jg/m3 14.3|jg/m3
50-
CO
E.
— ;
oncentration (
ND
Ol
o
0-
29
*
25
*
:ll
n
Pi ±
* T :r
* *
1 2 3
*
i-Li
5
*
60-
30-
0-
61
*
35
52
*
*
,-rI
pi I I +
rl H ~
*
•k
4 123
RO
by
*
I
+
—
I
*
4
2301900021 4210100041
150-
75-
Atlanta, GA
21.6|jg/m3
80
*
74
*
77
* 81
T *
* * * *
1234
70-
35-
0-
Miami, FL
12.1 pg/m3
77
*
81
*
78
*
41
*
T. T -p
* *
1234
1308920011 1202510161
Detroit, Ml Steubenville, OH
13.2|jg/m3 18.3|jg/m3
70-
co~^
E
c
•° 35-
ro
^
CD
O
O
O
24 25 28
* * * 27
*
TTT
i i i+i i
+ - n 0
H M T 3
* * *
/o-
35-
n _
83
*
71
*
62
*
ST
T
nil
i "•"
T ~
^ * * I
*
1234 1234
2614700051 3908110011
Lexington, KY Baton Rouge, LA
15.4 pg/m3 15.1 ug/m3
70-
35-
0-
30
* 31
* 28
_ *
21
* I JL T
I D H H
Bn M n
1 1 — I— ^L,
*
* *
60-
30-
o-
91
*
88
*
87
* 89
* T
T T i-Lj
0 R H -
Sx U2
sk zfc ^^
* *
1234 1234
2106700121 2203300091
Figure 3-7a. Quarterly distributions of 24-h average PM2 5 concentrations obtained in
eight eastern U.S. cities by the nationwide SLAMS/ NAMS network of PM2 5
FRM monitors during 1999. The data show the lowest, lowest tenth
percentile, lower quartile, median, highest quartile, highest tenth percentile,
and highest PM2 5 values. Values given above the highest extreme value in the
graphs refer to the number of observations. Annual average concentrations
are shown immediately above each graph.
March 2001
3-13
DRAFT-DO NOT QUOTE OR CITE
-------
40-
CD
O
O
O
Fargo, ND
9.4 pg/m3
1234
3801710041
Kansas City, MO-KS
11.6 ug/m3
50-
25-
89
1234
2904700411
60 H
30-
Wichita, KS
12.5 |jg/m3
23
*
*
1234
2017300101
Cheyenne, WY
5.6|jg/m3
30-
ti
CJl
i
a
Conce
27
*
1234
5602100011
Albuquerque, NM
5.2 ug/m3
40-
20-
29 33 32
1234
3504310031
50-
25-
Phoenix, AZ
13.1 ug/m3
43
69
64
*
8?
* *
1234
0401399911
Figure 3-7b. Quarterly distributions of 24-h average PM2 5 concentrations obtained in
six central and mountain U.S. cities by the nationwide SLAMS/ NAMS
network of PM25 FRM monitors during 1999. The data show the lowest,
lowest tenth percentile, lower quartile, median, highest quartile, highest
tenth percentile, and highest PM2 5 values. Values given above the highest
extreme value in the graphs refer to the number of observations. Annual
average concentrations are shown immediately above each graph.
March 2001
3-14
DRAFT-DO NOT QUOTE OR CITE
-------
Bellevue, WA Sacramento, CA
10.3 ug/m3 16.2 ug/rn3
70-
,— ,
1
O)
— '
1 35-
CO
i —
1
o
o
0 -
88
*
86
*
89 83 T
* * JL,
_— J|>
e'
+
f
* * * *
100-
50-
n _
61
*
29
*
_ 41 H
T *
I 77
P * I
h
[
* W *r *
1234 1234
5303300211 0606700101
Bakersfield, CA
27.9|jg/m3
E. Los Angeles, CA
25.9|jg/m3
140 H
70-
70
*
29
*
H
1
H
h
. 88 88
* * I
• IP * >
h
C
i<
100-
50-
n
29
*
29
*
17
*
T
-P 25 i-l-|
I *
I T +
1+ r^ 1
— + 0
T -*- * *
1234 1234
0602900141 0603716011
CD
O
O
O
San Diego, CA Anchorage, AK Honolulu, HI San Juan, PR
16.6 ug/m3 6.5 ug/m3 5.0 ug/m3 10.3 ug/m3
70-
35-
0-
80
*
74
*
82
_ * -,-
T 80
* I
i-L| rn
H T zN
T 9 Q T
-'--'- *
80-
40-
0-
46
*
28
*
54
*
J_ 19 X
^ti r+i
* ™ w *
24-
12-
0-
78
*
t (105)
82
T * 81 *
I
[+i -p
nP *
T U j
"^ \ ^1 1 — 1
40-
20-
0-
66
* 60
59 68
T *
* n
1+1 i -j-
SSgH
* T
* * _L
*
1234 1234 1234 1234
0607300031 0202000181 1500310011 7206100051
Figure 3-7c. Quarterly distributions of 24-h average PM2 5 concentrations obtained in
seven western U.S. cities and San Juan, PR, by the nationwide SLAMS/
NAMS network of PM25 FRM monitors during 1999. The data show the
lowest, lowest tenth percentile, lower quartile, median, highest quartile,
highest tenth percentile, and highest VM25 values. Values given above the
highest extreme value in the graphs refer to the number of observations. For
Honolulu, HI the highest value observed is shown in parentheses. Annual
average concentrations are shown immediately above each graph.
March 2001
3-15
DRAFT-DO NOT QUOTE OR CITE
-------
1 Seasonal variations in PM25 and PM10 concentrations obtained during the course of this study are
2 shown in Figure 3-8.
o
"cc
cu
o
c
o
o
o
03
D_
70
60-
50-
40
30-
20-
10-
0
Philadelphia j Washington | Nashville | Boston
SPSU F W SPSU F W SPSU F W SPSU F W
Season
Figure 3-8. Concentrations of PM2 5 and PM10 measured in the four MAACS cities. The
data show the lowest, lowest tenth percentile, lowest quartile, median highest
quartile, highest tenth percentile, and highest PM2 5 24-h average values. The
dashed line shows the level of the annual PM2 5 standard.
Source: Bahadori et al. (2000b).
1 The data for the four cities included in MAACS are presented as box plots showing the
2 lowest, lowest tenth percentile, lowest quartile, median, highest quartile, highest tenth percentile,
3 and highest PM2 5 and PM10 values in Figure 3-8. Highest PM2 5 and PM10 values are found
4 during the summer in all four cities, and mean values are highest during the summer in all cities,
5 although the seasonal pattern in Boston appears to be more nearly bi-modal with an additional
6 winter peak. This seasonal pattern, based on 2- to 3-year sampling periods for each city during
7 1992 through 1996, is in accord with that obtained from the FRM monitors in the NAMS and
8 SLAMS network.
March 2001
3-16
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
6
7
8
9
Frequency distributions for PM2 5 concentrations obtained in Philadelphia are shown in
Figure 3-9 (Bahadori et al., 2000b). Concentrations predicted from the log-normal distribution,
using geometric mean values and standard deviations derived from the data, also are shown.
In Philadelphia, the highest PM2 5 values were observed when winds were from the southwest
during sunny but hazy high pressure conditions. In contrast, the lowest values were found after
significant rainstorms during all seasons of the year. Day-to-day concentration differences in the
data set are 6.8 ± 6.5 //g/m3 for PM25 and 8.6 ± 7.5 //g/m3 for PM10. Maximum day-to-day
concentration differences are 54.7 //g/m3 for PM2 5 and 50.4 //g/m3 for PM10.
350
300-
250-
"|- 200-
(0
CO
° 150-
d
100-
50 -
PM2.5
geometric mean = 15.2 |jg/m3
Og=1.69
|LJ|LJ|LJ|LJ|LJ|LJ|LJ|
10 20 30 40 50 60
Concentration (|jg/m3)
Figure 3-9. Frequency distribution of 24-h average PM2 5 concentrations measured at the
PBY site in southwestern Philadelphia. Log-normal distribution fit to the data
shown as solid line.
1
2
3
4
Different conclusions could be drawn about data collected elsewhere in the United States.
PM25 concentrations obtained in Phoenix, AZ, from 1995 through 1997 are summarized in
Figure 3-10 and frequency distributions of PM25 concentrations obtained in Phoenix are shown
in Figure 3-11. Day-to-day concentration differences in this data set are 2.9 ±3.0 //g/m3 with a
March 2001
3-17
DRAFT-DO NOT QUOTE OR CITE
-------
40
30-
o 20
-i-i
CO
CD
O
c
o
O
10-
0
Phoenix, AZ
PM2.5
(n = 876)
Y T Y
Mar - May June-Aug Sep-Nov Dec-Feb
Figure 3-10. Concentrations of 24-h average PM25 measured at the EPA site in Phoenix,
AZ. The data show the lowest, lowest tenth percentile, lowest quartile,
median (black circles), highest quartile, highest tenth percentile, and highest
PM2 5 values.
W150-
a>
Q_
E
Si IOC-
'S
o
Z 50-
OJ
I
/
J
\
a. = 1.70
^
n\
iTR^Ln
1 1 I I rrr*f=^^-^_
i i i i i i i i i i i i i i r
5 10 15 20 25 30 35 4
Concentration (pg/m )
Figure 3-11. Frequency distribution of 24-h average PM25 concentrations measured at the
EPA site in Phoenix, AZ.
March 2001
3-18
DRAFT-DO NOT QUOTE OR CITE
-------
1 maximum day-to-day concentration difference of 23 //g/m . PM25 and PM(10_25) data were
2 obtained at a number of sites in California on a sampling schedule of every 6 days with
3 dichotomous samplers from 1989 through 1998. Histograms showing the frequency distribution
4 of the entire set of PM2 5 and PM(10_2 5) concentrations obtained by the CARB network of
5 dichotomous samplers from 1989 to 1998 are shown in Figures 3-12 and 3-13. Also shown are
6 log-normal distributions generated using geometric means and standard deviations derived from
7 the data as input. Although the data for both size fractions appear to be reasonably well
8 simulated by the function, data obtained at individual locations may not be. Data showing the
9 seasonal variability of PM25 obtained at Riverside-Rubidoux are summarized in box plot form in
10 Figure 3-14. The frequency distribution of PM2 5 concentrations obtained at Riverside-Rubidoux
11 from 1989 to 1994 is shown in Figure 3-15. It can be seen that the data are not as well fit by a
12 log-normal distribution as can the data shown in Figure 3-9, for example, mainly as the result of
13 a significant number of days with PM2 5 >100 //g/m3.
14 An examination of the data from the four MAACS cities, Phoenix, AZ, and Riverside, CA,
15 indicates that substantial differences exist in aerosol properties between widely separated
16 geographic regions. Fine-mode particles make up most of the PM10 mass observed in the
17 MAACS cities and appear to drive the daily and seasonal variability in PM10 concentrations
18 there. Coarse-mode particles represent a larger fraction of PM10 mass in Phoenix and Riverside
19 and drive the seasonal variability in PM10 seen there. The ratio of PM25 to PM10 concentrations is
20 much larger in the MAACS cities of Philadelphia (0.72); Washington, DC (0.74); and Nashville
21 (0.63) than in either Phoenix (0.34) or Riverside (0.49). Differences between median and
22 maximum concentrations in any size fraction are much larger at the Riverside site than at either
23 the MAACS or Phoenix sites. Many of these differences could reflect the more sporadic nature
24 of dust suspension at Riverside. In addition, the seasonal variability of PM25 concentrations
25 observed in Phoenix, AZ, and Riverside, CA, appears to be different from that observed in the
26 MAACS cities. These considerations demonstrate the hazards in extrapolating conclusions about
27 the nature of variability in aerosol characteristics inferred at one location to another.
28
29 3.2.2 Diurnal (Orcadian) Variability
30 The variability of PM concentrations on time scales shorter than a day can, in principle, be
31 characterized by measurements made by TEOMs and p-gauge monitors that are currently used
March 2001 3-19 DRAFT-DO NOT QUOTE OR CITE
-------
3000
2500-
2000-
1500-
1000-
500
PM25
geometric mean = 12.8 pg/m3
an = 2.29
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Figure 3-12. Frequency distribution of 24-h average PM25 measurements obtained from all
California Air Resources Board dichotomous sampler sites from 1989 to 1998.
3000
2500
2000-
1500-
1000-
500
PM(iO-2.5)
geometric mean = 15.7 [jg/m3
cin = 2.26
t t t—t t—t t t >—t—t—t-
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Figure 3-13. Frequency distribution of 24-h average PM(10_2 5) concentrations obtained
from all California Air Resource Board Dichotomous sampler sites from 1989
to 1998.
March 2001
3-20 DRAFT-DO NOT QUOTE OR CITE
-------
140-
120-
CO
•§ 100-
1 80~
1 60-
-------
1 to provide Air Quality Index (API) information to the public. A description of these methods
2 was given in Section 2.2.5. However, as shown in Chapter 2, continuous methods are subject to
3 artifacts because, in large part, of heating of their inlets, which results in the loss of components
4 such as water, ammonium nitrate, and semi-volatile organic compounds (cf Sections 2.2.1.1 and
5 2.2.3 for further details concerning the chemistry of volatilizable components), so caution should
6 be used in interpreting results obtained by these techniques.
7 The composite diurnal variation of PM25 concentrations obtained throughout the
8 continental United States by 31 TEOM and P-gauge monitors reporting to AIRS in 1999 is
9 shown in Figure 3-16. As can be seen, there is a distinct pattern with maxima occurring during
10 the morning and evening. Notable exceptions to this pattern occur in California where broad
11 nighttime maxima and daytime minima occur, which may be related to the use of P-gauge
12 monitors with unheated inlets there. It should be noted in examining the diurnal variations
13 shown in Figure 3-16, that there is substantial day-to-day variability in the diurnal profile of
14 PM2 5 measured at the same location which is then smoothed out after a suitably long averaging
15 period is chosen. The large ratio of the interquartile range to the median values supports the
16 view that there is substantial variability in the diurnal profiles.
17 The diurnal variability of PM components is determined by interactions between variations
18 in emissions, the rates of photochemical transformations, and the vertical extent and intensity of
19 turbulent mixing near the surface. Wilson and Stockburger (1990) characterized the diurnal
20 variability of sulfate and lead in Philadelphia. At that time, Pb was emitted mainly by motor
21 vehicles. Pollutants emitted mainly by motor vehicles, such as carbon monoxide, show two
22 distinct peaks occurring during the morning and evening rush hours (see Chapter 3, U.S.
23 Environmental Protection Agency, 2000c). Pollutants, such as sulfate, which are transported
24 long distances in the free troposphere (i.e., above the planetary boundary layer), tend to be mixed
25 downward and have their highest concentrations during the afternoon when the intensity and
26 vertical extent of turbulent mixing are greatest. Secondary aerosol components that are produced
27 by photochemical reactions such as secondary organic compounds may have a daily maximum in
28 the afternoon, by analogy with ozone. PM produced by residential heating (e.g., from wood
29 burning), on the other hand, reach maximum levels during the night.
30 Although the interquartile ranges for hour-to-hour changes in PM25 concentrations shown
31 in Figure 3-16 encompass several //g/m3, extreme values for the hour-to-hour variations can be
March 2001 3-22 DRAFT-DO NOT QUOTE OR CITE
-------
30-
0)
o
c
o
O
o-
6685 6667
6576 6571 6605 6640 6649
\ \ \ \ \ r
\ \ \ r
\ \ \ \ r
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 3-16. 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.
Median, mean (+) and interquartile ranges are shown.
Source: Fitz-Simons et al. (2000).
1
2
3
4
5
6
7
8
9
10
11
much larger (Fitz-Simons et al., 2000). The 98th percentile values for positive and negative
excursions in concentration are all less than 20 //g/m3. Maximum positive excursions were much
larger, ranging from 27 //g/m3 in the Northeast up to 220 //g/m3 in the Southwest, and with
maximum excursions in other regions all less than 125 //g/m3. It should be borne in mind that
the hour-to-hour changes that are reported reflect the effects of a number of processes occurring
during passage through the sampler inlets and on the TEOM measurement elements. These
considerations add uncertainty to the interpretation of the hour-to-hour changes that are observed,
as discussed earlier 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 to the true values that were present.
March 2001
3-23
DRAFT-DO NOT QUOTE OR CITE
-------
1 3.2.3 Relations Among Particulate Matter in Different Size Fractions
2 Relations Among PM2^S, PM(10_2S), andPM10
3 Data obtained in 1999 by collocated PM2 5 and PM10 FRM monitors have been used to
4 calculate the ratio of PM25 to PM10 concentrations and correlations among PM25, PM(10.25) and
5 PM10 concentrations. Results are shown in Table 3-2 for each of the seven aerosol characteristic
6 regions identified in Chapter 6 of CD 96. As can be seen from the table, the ratio of PM2 5 to
7 PM10 concentrations tends to be higher in the eastern United States than in the western United
8 States. This general pattern and the values are consistent with that found for the studies included
9 in Appendix A to Chapter 6 of CD 96. In that compilation, the mean ratio of PM2 5 to PM10 was
10 0.75 in the East, 0.52 in the central United States, and 0.53 in the western United States.
11 Although a large number of paired entries have been included in Table 3-2, seasonal variations
12 and annual averages in a number of regions could not be determined from the data set because of
13 data sparseness, mainly during the early part of 1999. It also can be seen in Table 3-2 that the
14 ratio of PM25 to PM10 was greater than one for a few hundred measurements. There are a number
15 of reasons for these results, many of which arise because the ratios are based on two independent
16 measurements. Measurement imprecisions play a role when the ratios are large and especially
17 when concentrations are small. Differences in the behavior of semi-volatile components in the
18 two samplers also could occur. The results also may be the result of errors in sampler placement,
19 field, laboratory, or data processing procedures.
20
21 Ultrafine Particle Concentrations
22 Data for characterizing the levels of ultrafme particles (<0.15-//m AD) and the relations
23 between ultrafme particles and larger particles are sparse. Perhaps the most extensive data set for
24 ultrafme particle properties is that described by Woo et al. (2000) for a site located 10 km to the
25 northwest of downtown Atlanta, GA. Size distributions from 3 to 2000 nm were being measured
26 every 12 min for 24 mo beginning in August 1998. Approximately 89% of the total number of
27 particles were found to be smaller than 100 nm, whereas 26% were found to be smaller than
28 10 nm. Concentrations tend to be lower during the summer than during the winter.
29 No correlation was found between number concentration and either volume or surface area for
30 particle sizes up to 2 //m. Because the total number of particles is concentrated in the smallest
31 size ranges, these results also indicate that fine particle mass does not correlate with the number
March 2001 3-24 DRAFT-DO NOT QUOTE OR CITE
-------
.
to
o
o
TABLE 3-2. DISTRIBUTION OF RATIOS OF PM25 TO PM10 AND CORRELATIONS BETWEEN PM25 AND PM10,
PM2 5 AND PM(10 2 5), AND PM (10 2 5) AND PM10 FOUND AT COLLOCATED MONITORING SITES IN SEVEN
AEROSOL CHARACTERISTIC (EPA/HEI) REGIONS IN 1999
Percentiles
Correlations
to
Region
Northeast
Southeast
Industrial Midwest
Upper Midwest
Southwest
Northwest
Mean Sites Values 95 90 75 50 25 10 5
0.70 45 1433 0.97 0.95 0.77 0.67 0.60 0.51 0.48
0.70
0.70
0.53
0.38
0.50
Southern California 0.47
76
92
39
23
73
36
2823 1.27 1.06 0.74 0.63 0.54 0.46 0.43
4827 1.09 0.88 0.78 0.68 0.59 0.51 0.47
1446 0.92 0.84 0.62 0.49 0.44 0.34 0.24
701
0.51 0.51 0.47 0.40 0.31 0.23 0.23
3300 0.67 0.65 0.56 0.49 0.44 0.39 0.36
1813 0.70 0.57 0.55 0.48 0.44 0.31 0.24
PM2.5:PM10
0.72a
0.69a
0.71a
0.35a
0.63a
0.69a
0.70a
384 16,343
0.02
-0.04a
0.17a
-0.02
0.49a
0.07a
0.19a
PM(10_2.5):PM1
0.71a
0.69a
0.81a
0.93a
0.99a
0.77a
0.83a
fe
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
"Results considered to be significantly different from zero at the a = 0.01 level.
Source: U.S. EPA Aerometric Information Retrieval System.
-------
1 of ultrafine particles. The high time resolution of the measurements allows some inferences to be
2 made about the possible sources of the ultrafme particles. The number of particles larger than
3 10 nm tends to peak during the morning rush hour (around 8 a.m.) and then to decrease through
4 the day and to increase again after 6 p.m., consistent with a traffic source. Particles smaller than
5 10 nm tend to peak during the mid-afternoon, consistent with nucleation involving products of
6 active photochemistry (McMurry et al., 2000). More direct relations between particle mass
7 observed in different size ranges can be obtained using multi-stage impactors. Keywood et al.
8 (1999) found a correlation between PM2 5 and PM015 of about 0.7, whereas they found
9 correlations of about 0.96 between PMX and PM2 5 and between PM2 5 and PM10 based on samples
10 collected by MOUDIs (Multiple Orifice Uniform Deposit Impactors) in six Australian cities.
11
12 3.2.4 Relations Between Mass and Chemical Component Concentrations
13 Time series of elemental composition data for PM2 5 particles based on X-ray fluorescence
14 (XRF) analyses have been obtained at a few locations across the United States. Time series of
15 components of the organic carbon fraction of the aerosol have not yet been obtained. The filter
16 samples that were collected at the PBY site in southwestern Philadelphia and were used in the
17 construction of Figures 3-8 and 3-9 also were analyzed by XRF. Concentrations of the trace
18 elements and correlations between trace elements and the total mass of particles in the PM2 5 size
19 range are shown in Table 3-3. Also shown in Table 3-3 are similar results obtained for filter
20 samples collected in Phoenix, AZ. Filters from both monitoring studies were analyzed by the
21 same X-ray spectrometer at the EPA facility in Research Triangle Park, NC. As can be seen
22 from inspection of Table 3-3, the analytical uncertainty (given in parentheses next to
23 concentrations) as a fraction of the absolute concentration is highly variable, and it exceeds the
24 concentration for a number of trace metals whose absolute concentrations are low, whereas it is
25 very small for abundant elements such as sulfur.
26 There are a number of distinct differences between the two data sets. For instance, sulfate
27 and associated cations and water appear to constitute a major fraction of the composition of the
28 PM in the Philadelphia data set, whereas they appear to constitute a much smaller fraction of the
29 Phoenix data set. The highest PM2 5 values were observed in Philadelphia during episodes driven
30 by high sulfate abundances and are caused, at least partly, by higher sulfate concentrations.
31 Correlation coefficients between SOJ and PM2 5 were 0.97 during the summer of 1993. Similar
March 2001 3-26 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-3. CONCENTRATIONS OF PM2 5 (^g/m3) AND SELECTED ELEMENTS
(ng/m3) IN THE PM2 5 SIZE RANGE AND CORRELATIONS BETWEEN ELEMENTS
AND PM2 5 MASS
(Values in parentheses refer to analytical uncertainty in X-ray fluorescence
determinations.)
PM25
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Pb
Philadelphia, PAa
Concentration
17.0 ± 0.8 ,wg/m3
4.0 (3 1.1) ng/m3
116(21.1)
8.6 (10.3)
2100 (143)
5.1 (3.4)
60.4 (4.7)
46.6 (4.2)
4.9(4.1)
8.8(1.8)
0.7 (0.7)
3.1(0.8)
109 (10.5)
0.1 (1.4)
7.3 (1.4)
4.8(1.1)
36.9(3.7)
0.6(1.2)
1.5 (0.6)
5.0 (0.9)
17.6 (2.5)
r
1
0.10
0.51
0.31
0.92
-0.01
0.50
0.39
0.44
0.37
0.15
0.39
0.50
0.04
0.22
0.25
0.21
0.18
0.63
0.11
0.19
Phoenix, AZb
Concentration
9.4 ± 0.5 Mg/m3
68.9 (27.2) ng/m3
209 (48.4)
7.6 (4.5)
408 (30.9)
11.4(2.4)
78.6 (8.2)
76.5 (9.7)
7.2(3.3)
0.7 (1.0)
0.4 (0.4)
4.3 (0.6)
112(15.1)
-0.2 (0.8)
0.4 (0.4)
3.3 (0.7)
12.7 (1.7)
1.3 (0.6)
0.3 (0.3)
3.1 (0.6)
4.5(1.0)
r
1
0.23
0.35
0.52
0.16
0.13
0.67
0.51
0.44
-0.28
0.41
0.64
0.80
-0.01
0.38
0.69
0.64
0.50
0.40
0.57
0.69
an= 1105.
bn=643.
1 correlations between SO4 and PM2 5 were found at a site in northeastern Philadelphia (24 km
2 distant from the site under discussion) during the summer of 1993.
March 2001
3-27
DRAFT-DO NOT QUOTE OR CITE
-------
1 Concentrations of "crustal elements" (e.g., Al, Si, K, Ca, Ti, Mn, and Fe) constitute a larger
2 fraction of PM2 5 mass in the Phoenix data set than they do in the Philadelphia data set. Sulfur is
3 very highly correlated with PM2 5 r = 0.92) in the Philadelphia data set, whereas it is only weakly
4 correlated r = 0.16) with PM2 5 in the Phoenix data set. Trace metals (e.g., Cr, Co, Ni, Cu, Zn,
5 As, and Pb) are not well correlated ( 0.04 < r < 0.25) with PM2 5 in the Philadelphia data set,
6 whereas they are more variably correlated (0.01 < r < 0.69) with PM25 in the Phoenix data set.
7 The uncertainty in the concentration measurement most probably plays a role in determining a
8 species' correlation with PM2 5, especially when the analytical uncertainty is high relative to
9 concentration (e.g., for trace metals such as Co).
10
ll 3.2.5 Spatial Variability
12 Intersite correlation coefficients for PM25 can be calculated based on the results of FRM
13 monitors placed at multiple sites within Metropolitan Statistical Areas (MSAs) across the United
14 States. Pearson correlation coefficients for PM2 5 monitors located in the Atlanta, GA; Detroit,
15 MI; Phoenix, AZ; and Seattle, WA, MSAs are shown in Tables 3-4a through d. Only sites with
16 at least 100 measurement days were chosen, and, furthermore, only days with concurrent
17 measurements were selected from this subset of monitoring sites. As can be seen from
18 Table 3-4a through d, PM25 concentrations tend to be highly correlated among sites within all of
19 the MSAs shown, although there can be exceptions, as shown in the results for Atlanta, GA.
20 There are a number of factors that could lower intersite correlations. These include field
21 measurement and laboratory analysis errors, placement of monitors close to active sources, and
22 transient local events.
23 In the Philadelphia area, PM2 5 was found to be strongly correlated (r > 0.9) between seven
24 urban sites and one background site (Valley Forge, PA) during the summer of 1993 (Suh et al.,
25 1995). Similar relationships also were found during the summer of 1994 at four monitoring sites
26 as part of a separate study (Pinto et al., 1995). The data collected in these studies also indicate
27 that PM2 5 and SOJ concentrations are spatially uniform throughout the Philadelphia
28 metropolitan area, and that variability in PM10 levels is caused largely by variability in PM2 5
29 (Wilson and Suh, 1997).
30 Three methods for comparing the chemical composition of aerosol databases obtained at
31 different locations and times were discussed by Wongphatarakul et al. (1998). Log-log plots of
March 2001 3-28 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-4. PEARSON CORRELATION COEFFICIENTS FOR THE SPATIAL
VARIATION OF PM2 5 CONCENTRATIONS IN SELECTED METROPOLITAN
STATISTICAL AREAS
(AIRS site ID numbers without the state code are used to identify sites.)
(a) Atlanta, GA (n = 46)
08920011
12100321
12110011
06300911
12100391
08900021
08920011 12100321 12110011 06300911
1 0.87 0.89 0.86
1 0.88 0.78
1 0.93
1
12100391
0.90
0.91
0.90
0.86
1
08900021
0.48
0.33
0.42
0.39
0.36
1
(b) Detroit, MI (n = 52)
09900091
14700051
16300011
16300161
(c) Phoenix-Mesa, AZ (n = 45)
01399901
01399971
01399911
01399921
(d) Seattle-Bellevue-Everett,
WA (n = 78)
03300211
03300571
03300801
03320041
06110071
09900091 14700051 16300011 16300161
1 0.98 0.96 0.97
1 0.94 0.93
1 0.95
1
01399901 01399971 01399911 01399921
1 0.92 0.89 0.84
1 0.97 0.90
1 0.89
1
03300211 03300571 03300801 03320041 06110071
1 0.96 0.95 0.95 0.81
1 0.93 0.92 0.80
1 0.94 0.83
1 0.90
1
Source: Fitz-Simons et al. (2000); data from EPA Aerometric Information Retrieval System (AIRS).
March 2001
3-29
DRAFT-DO NOT QUOTE OR CITE
-------
1 chemical concentrations obtained at pairs of sampling sites accompanied by the coefficient of
2 divergence (COD) were examined as a way to provide an easily visualized means of comparing
3 two data sets1. Examples comparing downtown Los Angeles with Burbank and with Riverside-
4 Rubidoux are shown in Figures 3-17 and 3-18, respectively. As the composition of two sampling
5 sites become more similar, the COD approaches zero; as their compositions diverge, the COD
6 approaches one. Cluster analyses based on the COD between individual data sets can be used to
7 determine the degree of similarity among a number of data sets. Correlation coefficients
8 calculated between components can be used to show the degree of similarity between pairs of
9 sampling sites. In addition to calculating correlation coefficients for total mass or for individual
10 components, correlation coefficients for characterizing the spatial variation of the contributions
11 from given source types can also be calculated by averaging the correlation coefficients of the set
12 of chemical components that represent the source type. The first two methods could be applied
13 either to aerosol data sets collected at multiple sites within a given geographic region or to
14 aerosol data sets collected at widely different locations or times while the third method is best
15 used to characterize sites within a particular geographic region.
16 Correlation coefficients showing the spatial relations among PM2 5 (total) and contributions
17 from different source categories obtained at various sites in the South Coast Air Basin (SoCAB)
18 Study are shown in Table 3-5. In Wongphatarakul et al. (1998), crustal material (crustal), motor
19 vehicle exhaust (mv), residual oil emissions (residual oil), and secondary PM (sec) were
20 considered as source categories. Al, Si, Fe, and Ca were used as markers for crustal material
21 (crustal). V and Ni were used as markers for fuel oil combustion (residual oil). Pb, Br, and Mn
22 were used as markers for motor vehicle exhaust (mv). NO3", NH4+, and SO4= represent secondary
23 PM components (sec). The average of the correlation coefficients of marker elements within
24 each source category are shown in Table 3-3. Values of rsec and rmv are much higher than those
25 for rcrustal and rresidual oil throughout the SoCAB suggesting a more uniform distribution of the
lrThe COD for two sampling sites is defined as follows:
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.
March 2001 3-30 DRAFT-DO NOT QUOTE OR CITE
-------
102
«£ 10°
c
CO
I 1Q-1
QQ
10-2-
10
-3
COD=0.099
so,
2-
Source: Wongphatarakul (1997)
T
T
10"
1Q-2 1Q-1 1Q0
Downtown Los Angeles (|jg/m3)
102
*Jan. 2 - Dec. 28,1986 (63 data points), 24 hours sampling, sampling every 6 days, dp < 2.5 pm
**Jan. 2 - Dec. 28,1986 (61 data points), 24 hours sampling, sampling every 6 days, dp < 2.5 |jm
Figure 3-17. PM2 5 chemical components in downtown Los Angeles and Burbank (1986)
have similar characteristics.
Source: Wongphatarakul, etal. (1998).
March 2001
3-31
DRAFT-DO NOT QUOTE OR CITE
-------
X
D
O
102
E 10°-
10"-
10
-3
10
,-3
COD=0.230
Unknown
Gas Al
K
Zn
Source: Wongphatarakul(1997)
T
10
T
T
10" 10U
Downtown Los Angeles (pg/m3)
102
*Jan. 2 - Dec. 28,1986 (63 data points), 24 hours sampling, sampling every 6 days, dp < 2.5 urn
**Jan. 2 - Dec. 28,1986 (60 data points), 24 hours sampling, sampling every 6 days, dp <2.5 urn
Figure 3-18. 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 3-13).
Source: Wongphatarakul (1998).
March 2001
3-32
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-5. CORRELATION COEFFICIENTS FOR SPATIAL VARIATION OF
PM2 5 MASS AND DIFFERENT SOURCES FOR PAIRS OF SAMPLING
SITES IN THE SOUTH COAST AIR BASIN (1986)
Hawthorne and Rubidoux
Long Beach and Rubidoux
Anaheim and Rubidoux
Downtown Los Angeles and Rubidoux
Burbank and Rubidoux
Hawthorne and Anaheim
Long Beach and Anaheim
Burbank and Anaheim
Downtown Los Angeles and Anaheim
Downtown Los Angeles and Hawthorne
Burbank and Hawthorne
Long Beach and Burbank
Long Beach and Hawthorne
Downtown Long Angeles and Long Beach
Downtown Los Angeles and Burbank
^total
-0.027
0.051
0.066
0.095
0.120
0.760
0.852
0.770
0.827
0.808
0.704
0.731
0.880
0.842
0.928
crustal
0.034
0.075
0.105
0.143
0.568
0.599
0.633
0.649
0.653
0.825
rsec
0.768
0.888
0.749
0.804
0.854
0.790
0.737
0.909
0.817
0.960
rmv
0.492
0.504
0.579
0.556
0.669
0.688
0.714
0.861
0.719
0.871
residual oil
0.170
0.150
0.161
0.233
0.533
0.491
0.295
0.482
0.378
0.606
Source: Wongphatarakuletal. (1998).
1 contributions from secondary PM formation and automobiles than from crustal material and
2 localized stationary sources.
3 Correlation coefficients in Philadelphia air for PM2 5 (total), crustal components (Al, Si, Ca,
4 and Fe), the major secondary component (sulfate), organic carbon (OC) and elemental carbon
5 (EC) are shown in Table 3-6. Because these data were obtained after Pb had been phased out of
6 gasoline, a motor vehicle contribution could not be estimated from the data. Pb also is emitted
7 by discrete point sources, such as the Franklin smelter. Sulfate in aerosol samples collected in
8 Philadelphia arises mainly from long-range transport from regionally dispersed sources (Dzubay
9 et al., 1988). This conclusion is strengthened by the high correlations in sulfate between different
March 2001
3-33
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-6. CORRELATION COEFFICIENTS FOR SPATIAL VARIATION OF
PM2 5 MASS AND DIFFERENT COMPONENTS FOR PAIRS OF SAMPLING
SITES IN PHILADELPHIA (1994)
Castor Ave. and Roxboro
Castor Ave. and NE Airport
Castor Ave. and Broad St.
Roxboro and NE Airport
Roxboro and Broad St.
NE Airport and Broad St.
rtot
0.92
0.93
0.93
0.98
0.95
0.95
crustal
0.52
0.47
0.57
0.67
0.90
0.69
rsec
0.98
0.99
0.99
0.98
0.98
0.99
roc
0.88
0.88
0.85
0.83
0.86
0.84
TEC
0.84
0.77
0.89
0.82
0.79
0.63
rPb
0.43
-0.07
0.11
0.20
0.47
0.11
Source: Pinto et al. (1995).
1 monitoring sites and the uniformity in sulfate concentrations observed among the sites.
2 Widespread area sources (e.g., motor vehicle traffic) also may emit pollutants that are relatively
3 spatially uniform and are highly correlated between sites with uniform traffic density and
4 emissions patterns. Very few studies have compared aerosol composition in urban areas to that
5 in nearby rural areas. Tanner and Parkhurst (2000), for example, found that sulfate constituted a
6 larger fraction of fine particle mass at rural sites in the Tennessee Valley PM25 monitoring
7 network than did organic carbon. For urban sites, they found the situation was largely reversed,
8 with organic carbon constituting a larger fraction of aerosol mass than sulfate.
9
10
11 3.3 SOURCES OF PRIMARY AND SECONDARY PARTICIPATE
12 MATTER
13 Information about the nature and relative importance of sources of ambient PM is presented
14 in this section. Table 3-7 summarize anthropogenic and natural sources for the major primary
15 and secondary aerosol constituents of fine and coarse particles. Major sources of each
16 constituent are shown in boldface type. Anthropogenic sources can be further divided into
17 stationary and mobile sources. Stationary sources include fuel combustion for electrical utilities,
18 residential space heating and industrial processes; construction and demolition; metals, minerals,
19 and petrochemicals; wood products processing; mills and elevators used in agriculture; erosion
March 2001 3-34 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-7. CONSTITUENTS OF ATMOSPHERIC PARTICLES AND THEIR MAJOR SOURCES
1,2
g. Sources
^ Primary (PM <2 . 5 ^m) Primary (PM >2 . 5 ^m) Secondary PM Precursors (PM <2 . 5 /j,m)
^ Aerosol
species Natural Anthropogenic Natural
SO4= Sea spray Fossil fuel combustion Sea spray
Sulfate
Anthropogenic Natural
— Oxidation of reduced sulfur
gases emitted by the oceans and
wetlands and SO2 and H2S
emitted by volcanism and forest
fires
Anthropogenic
Oxidation of SO2 emitted
from fossil fuel
combustion
N03-
Nitrate
Minerals
Oxidation of NO,, produced by
soils, forest fires, and lighting
Oxidation of NOX emitted
from fossil fuel
combustion and in motor
vehicle exhaust
Erosion and
reentrainment
Fugitive dust paved
and unpaved roads,
agriculture, and
forestry
Erosion and reentrainment
Fugitive dust, paved
and unpaved road
dust, agriculture, and
forestry
wv
O
;>
5
H
6
o
0
H
/O
r*^S
O
H
W
O
o
I— H
H
W
NH4+
Ammonium
Organic
carbon (OC)
Elemental
carbon
(EC)
Metals
Bioaerosols
—
Wild fires
Wild fires
Volcanic
activity
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
—
— —
— Tire and asphalt wear
and paved road dust
— —
Erosion, reentrainment, —
and organic debris
Plant and insect fragments, —
pollen, fungal spores, and
bacterial agglomerates
Emissions of NH3 from wild Emissions of NH3 from
animals, and undisturbed soil animal husbandry,
sewage, and fertilized land
Oxidation of hydrocarbons Oxidation of hydrocarbons
emitted by vegetation (terpenes, emitted by motor vehicles,
waxes) and wild fires prescribed burning, and
wood burning
— —
— —
— —
'Dash (-) indicates either very minor source or no known source of component.
2Major source of each component
shown in boldface type.
-------
1 from tilled lands; waste disposal and recycling; and fugitive dust from paved and unpaved roads.
2 Mobile, or transportation-related, sources include direct emissions of primary PM and secondary
3 PM precursors from highway and off-highway vehicles and nonroad sources. In addition to
4 fossil fuel combustion, biomass in the form of wood can be burned for fuel. Vegetation can be
5 burned to clear new land for agriculture and for building construction, to dispose of agricultural
6 and domestic waste, to control the growth of animal or plant pets, and to mange forest resources
7 (prescribed burning). Also shown are sources for precursor gases whose oxidation forms
8 secondary particulate matter. A description of the atmospheric chemical processes producing
9 secondary PM is given in Section 3.4.
10 In general, the nature of sources of fine particulate matter is very different from that for
11 coarse particulate matter. A large fraction of the mass in the fine size fraction is derived from
12 material that has been formed during combustion (primary), has been volatilized in combustion
13 chambers and then recondensed to form primary PM, or has been formed in the atmosphere from
14 precursor gases as secondary PM. Because precursor gases and fine particulate matter are
15 capable of traveling great distances, it is difficult to identify individual sources of constituents.
16 The coarse PM constituents have shorter lifetimes in the atmosphere, so their impacts tend to be
17 more localized. Only major sources for each constituent within each broad category shown at the
18 top of Table 3-7 are listed. Chemical characterizations of primary particulate emissions from a
19 wide variety of natural and anthropogenic sources as shown in Table 3-7 were given in Chapter 5
20 of 1996 PM AQCD. Summary tables of the composition of source emissions presented in 1996
21 PM AQCD and updates are given in Appendix B. These profiles were based in large measure on
22 the results of various studies collecting source signatures for use in source apportionment studies.
23 Natural sources of primary PM include windblown dust from undisturbed land, sea spray,
24 and plant and insect debris. The oxidation of a fraction of terpenes emitted by vegetation and
25 reduced sulfur species from anaerobic environments leads to secondary PM formation.
26 Ammonium (NH4+) ions, that regulate the pH of particles, are derived from emissions of
27 ammonia (NH3) gas. Source categories for NH3 have been divided into emissions from
28 undisturbed soils (natural) and emissions that are related to human activities (e.g., fertilized
29 lands, domestic and farm animal waste). There is considerable debate about characterizing
30 emissions from wild fires (i.e., unwanted fires) as either natural or anthropogenic.
31 Approximately 70 to 90% of wildfires may be ignited directly as the result of human activities,
March 2001 3-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 either deliberately through prescribed burning and arson, or inadvertently through the improper
2 disposal of flammable material or fugitive sparks (Andreae, 1991), with the remainder ignited by
3 lightning strikes. On the other hand, human intervention also suppresses wildland fires that are
4 ignited by natural causes (e.g., lightning strikes). Fire suppression allows the buildup of fire
5 fuels and increase the susceptibility of forests to more severe and infrequent fires from whatever
6 cause. Prescribed burning may limit the growth of these fuels and the chances for more
7 catastrophic fires.
8 Receptor models are perhaps the primary means used to estimate source category
9 contributions to particulate matter at individual monitoring sites. Dispersion models (i.e., three-
10 dimensional chemistry and transport models) are formulated in a prognostic manner (i.e., they
11 attempt to predict species concentrations using a tendency equation that includes terms based on
12 emissions inventories, atmospheric transport, chemical transformations, and deposition).
13 Receptor models are diagnostic in their approach (i.e., they attempt to derive source contributions
14 based either on ambient data alone or in combination with data from the chemical composition of
15 sources). These methods have the advantage that they do not invoke all of the uncertainties
16 inherent in emissions inventories or in parameterizing atmospheric transport processes in grid
17 point models. There are two main approaches to receptor modeling. Receptor models such as
18 the chemical mass balance (CMB) model (Watson et al., 1990a) relate source category
19 contributions to ambient concentrations based on analyses of the composition of ambient
20 particulate matter and source emissions samples. This technique has been developed for
21 apportioning source categories of primary particulate matter and was not formulated to include
22 the processes of secondary particulate matter formation. In the second approach, various forms
23 of factor analysis are used. They rely on the analysis of time series of compositional data from
24 ambient samples to derive both the composition of sources and the source contributions.
25 Standard approaches such as factor analysis or Principal Component Analysis (PCA) can
26 apportion only the variance and not the mass in an aerosol composition data set. Positive matrix
27 factorization (PMF) is a recently developed multivariate technique (Paatero and Tapper, 1993,
28 1994) that overcomes many of the limitations of standard techniques, such as principal
29 components analysis (PCA), by allowing for the treatment of missing data and data near or below
30 detection limits. This is accomplished by weighting elements inversely according to their
31 uncertainties. Standard methods such as PCA weight elements equally regardless of their
March 2001 3-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 uncertainty. Solutions also are constrained to yield nonnegative factors. Both the CMB and the
2 PMF approaches find a solution based on least squares fitting and minimize an object function
3 and both methods provide error estimates. For a complete apportionment of mass, all of the
4 major sources affecting a monitoring site must be sampled for analysis by CMB, whereas there is
5 no such restriction in the use of PMF.
6 A number of specialty conference proceedings, review articles, and books have been
7 published to provide greater detail about source category apportionment receptor models as
8 described in CD 96. A review of the various methods used to apportion PM in ambient samples
9 among its source categories was given in Section 5.5.2 of 1996 PM AQCD. The collection of the
10 source category characterization profiles shown in Appendix 3B has been motivated in many
11 cases by the need to use them in receptor modeling applications.
12
13 3.3.1 Source Contributions to Ambient Particulate Matter
14 The results of several source apportionment studies will be discussed in this section to
15 provide an indication of different sources of particulate matter across the United States. First,
16 results obtained mainly by using the chemical mass balance (CMB) approach for estimating
17 contributions to PM25 from different source categories at monitoring sites in the United States
18 will be discussed and presented in Table 3-8. More recent results using the PMF approach are
19 included for Phoenix, AZ. Results obtained at a number of monitoring sites in the central and
20 western United States by using the CMB model for PM10 are shown in Table 3-9. The sampling
21 sites represent a variety of different source characteristics within different regions of Arizona,
22 California, Colorado, Idaho, Illinois, Nevada, and Ohio. Several of these are background sites,
23 specifically Estrella Park, Gunnery Range, Pinnacle Peak, and Corona de Tucson, AZ, and
24 San Nicolas Island, CA. Definitions of source categories also vary from study to study. The
25 results of the PM10 source apportionment studies were given in 1996 PM AQCD and are
26 presented here to allow easy comparison with results of PM25 source apportionment studies.
27 There are several differences between the source categories shown at the tops of Tables 3-8
28 and 3-9. These differences reflect the nature of sources that are important for producing fine and
29 coarse particulate matter shown in Table 3-7. They also are related to improvements in the
30 ability to distinguish between sources of similar nature (e.g., diesel and gasoline vehicles, meat
31 cooking and vegetation burning). It has been only recently that motor vehicle emissions can be
March 2001 3-38 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-8. RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
-2.5
g.
to
o
o
U)
UJ
VO
o
£
H
6
o
0
H
O
C^j
o
H
W
O
o
1 — I
— ]
w
% Contribution3
Misc. Misc.
Sampling Site
Pasadena, CA (Schauer et al., 1996)
Downtown LA, CA (Schauer et al., 1996)
West LA, CA (Schauer et al., 1996)
Rubidoux, CA (Schauer et al., 1996)
Sacramento, CA (Motallebi, 1999)
Bakersfield, CA (Magliano et al., 1998)
Fresno, CA (Magliano et al., 1998)
Philadelphia, PA (Dzubay et al., 1988)
Camden, NJ (Dzubay et al., 1988)
Clarksboro, NJ (Dzubay et al., 1988)
Grover City IL ENEj; (Glover et al., 1991)
Grover City, IL SSW; (Glover et al., 1991)
Grover City, IL WNW; (Glover et al., 1991)
Grover City, IL NNW; (Glover et al., 1991)
Welby, CO (Lawson and Smith, 1998)
Brighton, CO (Lawson and Smith, 1998)
Reno, NV (Gillies et al., 2000)
Phoenix, AZ (Ramadan et al., 2000)
Phoenix, AZ (Ramadan et al., 2000)
"Secondary and other organic compounds.
""Secondary ammonium.
°Meat cooking.
Time Period
1982
1982
1982
1982
Winter
1991-1996
Winter 1996
Winter 1996
Summer 1982
Summer 1982
Summer 1982
1986 to 1987
1986 to 1987
1986 to 1987
1986 to 1987
Winter 1997
Winter 1997
Summer 1998
Summer
1995-1998
Winter
1995-1998
Gasoline Road Dust,
Diesel Vehicles Soil
18.8 5.7 12.4
35.7 6.5 11.1
18.0 5.7 12.2
12.8 0.7 13.1
24.5e — 1.2
16e — <3
13e — <3
8.5e — 4.4
9.2e — 3.2
5.8e — 2.7
— — 2.3
2.4e — 5.1
— — 3.1
10 28 16
10 26 11
68e — 14.5
10.9 36.2 1.8
14.5 38.9 1.1
Vegetation
Burning
9.6
5.8
11.0
1.2
18.1
20
19
—
—
—
—
—
—
5
2
4
15.0
8.9
Secondary
Sulfate
20.9
20.3
24.1
13.8
4.5
7
5
81. 9f
81.3f
84.6f
83. 2f
59.0f
88. 5f
86.6f
10°
15°
11
—
—
Secondary
Nitrate
7.4
9.2
7.8
24.7
36.6
34
32
—
0.4
—
—
—
—
25"
32"
2
—
—
Including associated cations and water.
8Incinerators.
hOil fly ash.
Source Source
1
5.3"
3.7"
4.1"
4.5"
—
—
—
2.28
2.58
0.88
9.7k
11.6"
2.8k
3.4"
4C
T
0.6"
20.8"
9.5"
kLead smelter.
"Iron works.
"Copper smelter
2
9.2b
9.2b
9.4b
12.1b
—
—
—
1.9h
2 5h
1.5"
3.0"
11.9"
—
3.0"
2"
2"
—
4.9r
4.5r
Misc. Misc. Measured
Source Source PM25
3 4 Concentration
8.5C l.ld 28.2
5.2C 0.6d 32.5
8.2C 1.6d 24.5
4.5C 0.5d 42.1
— — 39.5
— — 52
— — 63
0.41 — 27.0
0.71 — 28.3
0.41 — 26.0
1.28 —
4. 18 4.6m
— —
— —
— —
— —
— — 7.8
6.7s 3.6" 8.3
18.7s 4.1" 13.8
pAs ammonium nitrate.
"Sea salt.
"Wood burning.
•"Vegetative detritus.
'Value represents sum of diesel and gasoline vehicle exhaust.
'Fluidized catalyst cracker.
jWind direction.
"Coal power plant.
"As ammonium sulfate.
sNonferrous smelting.
-------
TABLE 3-9. RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
10
.
to
o
o
fe
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
% Contribution
Sampling Site
Central Phoenix, AZ (Chow et al., 1991)
Craycroft, AZ (Chow et al., 1992a)
Hayden 1, AZ (Garfield) (Ryan et al., 1988)
Hayden 2, AZ (Jail) (Ryan et al., 1988)
Phoenix, AZ (Estrella Park) (Chow et al., 1991)
Phoenix, AZ (Gunnery Rg.) (Chow et al., 1991)
Phoenix, AX (Pinnacle Pk.) (Chow et al., 1991)
Rillito, AZ (Thanukos et al., 1992)
Scottsdale, AZ (Chowetal., 1991)
West Phoenix, AZ (Chow et al., 1991)
Bakersfield, CA(Magliano, 1988)
Bakerfield, CA (Chow et al., 1992b)
Crows Landing, CA (Chow et al., 1992b)
Fellows, CA (Chow et al., 1992b)
Fresno, CA(Magliano, 1988)
Fresno, CA (Chow et al., 1992b)
Indio, CA (Kim et al., 1992)
Kern Wildlife Refuge, CA (Chow et al., 1992b)
Long Beach, CA (Gray et al., 1988)
Long Beach, CA (Summer) (Watson et al., 1994)
Long Beach, CA (Fall) (Watson et al., 1994)
Riverside, CA (Chow et al., 1992c)
Rubidoux, CA (Gray et al., 1988)
Rubidoux, CA (Summer) (Watson et al., 1994)
Rubidoux, CA (Fall) (Watson et al., 1994)
Rubidoux, CA (Chow et al., 1992c)
San Nicolas Island, CA (Summer) (Watson et al.,
Time Period
Winter 1989-1990
Winter 1989-1990
1986
1986
Winter 1989-1990
Winter 1989-1990
Winter 1989-1990
1988
Winter 1989-1990
Winter 1989-1990
1986
1988-1989
1988-1989
1988-1989
1986
1988-1989
1988-1989
1986
Summer 1987
Fall 1987
1988
1986
Summer 1987
Fall 1987
1988
Summer 1987
Primary
Geological
51.6
55.6
4.8
35.6
67.3
74.1
58.3
53.7
45.5
43.5
40.5
53.9
61.3
53.1
35.6
44.5
56.9
31.6
39.9
24.1
11.8
50.9
49.3
30.4
17.1
55.2
9.2
Primary
Motor
Primary Vehicle
Construction Exhaust
0.0
0.0
1.9b
6.8b
0.0
0.0
0.0
17.4b
0.0
0.0
4.4
2.0
0.0
2.6
1.5
0.0
5.2
4.2
0.0
0.0
0.0
0.0
4.6
3.9
14.4
0.0
0.0
39.0
35.5
0.0
0.0
18.2
20.4
24 2
1.5f
34.5
36.2
8.1
9.7
4.2
3.8
8.3
9.5
7.6
4.6
9.81
13.7
44.5
10.9
6.41
15.1
27.1
11.7
5.2
Primary
Vegetative
Burning
3.6
0.0
0.0
0.0
1.6
0.0
8.3
0.0
13.5
14.5
14.21
8.2
6.5
6.2
19.11
7.1
12.2
8.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Secondary
Ammonium
Sulfate
0.3
3.0
3.8
6.8
2.9
3.7
7.5
0.0
1.1
0.6
8.3
6.9
5.3
9.3
3.7
5.0
6.2
6.9
15.4
23.6
4.0
7.5
7.3
8.3
1.9
6.1
21.3
Secondary Misc.
Ammonium Source
Nitrate 1
4.4
2.6
0.0
0.0
0.0
0.0
0.0
0.0
6.5
4.5
0.0
16.0
12.4
13.7
0.0
14.5
7.1
3.1
17.7
1.7
24.1
33.4
24.4
23.9
28.2
24.9
2.9
0.0
5.1'
70. 5C
47.5C
0.0
0.0
0.0
14.68
0.0
0.0
0.7j
1.3™
1.0m
12.8m
0.2j
0.4m
0.3j
1.0m
0.2j
0.2j
0.0"
0.5j
0.3j
0.0"
0.0"
0.&
0.0"
Misc.
Source
2
0.0
0.0
4.8d
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.9"
1.9"
2.6"
0.0
1.9"
1.7"
3.1"
3.9h
4.8h
2.8h
2.0h
1.1"
4.4h
1.0h
1.7"
24.7h
Misc.
Source
3
0.0
0.0
1.0e
\.T
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.8k
2.3k
2.6k
0.0
O.lk
0.0
1.5k
12.3k
0.0
0.0
1.7°
6.8k
0.0
0.0
6.6°
0.0
Misc. Measured
Source PM10
4 Concentration
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
64.0
23.4
105.0
59.0
55.0
27.0
12.0
79.5
55.0
69.0
67.6
79.6
52.5
54.6
48.1
71.5
58.0
47.8
51.9
46.1
96.1
64.0
87.4
114.8
112.0
87.0
17.4
1994)
-------
TABLE 3-9 (cont'd). RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
10
.
to
o
o
fe
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
% Contribution
Sampling Site
Stockton, CA (Chow et al., 1992b)
Pocatello, ID (Houck et al., 1992)
S. Chicago, IL (Hopke et al., 1988)
S.E. Chicago, IL (Vermette et al., 1992)
Reno, NV (Chow et al., 1988)
Sparks, NV (Chow et al., 1988)
Follansbee, WV (Skidmore et al., 1992)
Mingo, OH (Skidmore et al., 1992)
Steubenville, OH (Skidmore et al., 1992)
Time Period
1989
1990
1986
1988
1986-1987
1986-1987
1991
1991
1991
Primary
Geological
55.1
8.3
34.0
35.9Y
497
36.8
15.2
20.0
18.0
Primary
Construction
0.8
7.5"
3.0
0.0
0.0
0.0
0.0
0.0
0.0
Primary
Motor
Vehicle
Exhaust
8.3
0.1
3.5
2.2f
33.3
28.3
53.0
23.3
30.4
Primary
Vegetative
Burning
7.7
0.0
0.0
0.0
6.3
32.7
0.0
6.8
1.7
Secondary
Ammonium
Sulfate
5.0
0.0
19.2"
18.8
4.3
6.6
24.2
25.0
30.4
Secondary Misc.
Ammonium Source
Nitrate 1
11.2 l.lm
0.0 0.0
— 18.9'
— 2.0'
2.0 0.0
2.2 0.0
— 14.1'
— 5.7'
— 8.3'
Misc.
Source
9
2 9°
0.0
2.7"
0.7"
0.0
0.0
0.0
18.3"
10.9"
Misc.
Source
3
0.0k
84.1r
0.0
2.7™
0.0
0.5k
0.0
0.0
0.0
Misc.
Source
4
0.0
0.0
0.0
18. 88
0.0
0.0
0.0
0.0
0.0
Measured
PM10
Concentration
62.4
100.0
80.1
41.0
30.0
41.0
66.0
60.0
46.0
"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.
8Background aerosol.
'Marine 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 + NaNO3.
"Lime.
pRoad sanding material.
'Asphalt industry.
•Phosphorus/phosphate industry.
'Regional sulfate.
'Steel mills.
"Refuse incinerator.
vLocal road dust, coal yard road dust, and steel haul road dust.
"Incineration.
"Unexplained mass.
-------
1 broken down into contributions from diesel and gasoline vehicles through the use of organic
2 tracers. Meat cooking is also distinguished from vegetation burning in more recent studies,
3 although both are considered to be part of biomass burning. Vegetation burning consists of
4 contributions from residential fuel wood burning, wild fires, prescribed burning and burning of
5 agricultural waste. Miscellaneous sources of fine particles include contributions from
6 combustion sources, whereas miscellaneous sources of coarse particles consist of contributions
7 from soil and sea spray and industrial processing of geological material (e.g., cement
8 manufacturing). Although a large number of elements and chemical compounds are used to
9 differentiate among source categories, it can be seen from Tables 3-8 and 3-9 that only a
10 relatively small number of sources are needed to account for the mass of PM2 5 and PM10.
11 Secondary sulfate is the dominant component of PM2 5 samples collected in the studies of
12 Dzubay et al. (1988) and Glover et al. (1991). Both studies found that sulfate at their monitoring
13 site arose from regionally dispersed sources. Sulfate, associated cations and water also represent
14 the major components of PM2 5 found in monitoring studies in the eastern United States
15 (Table 3-1). Motor vehicle emissions, arising mainly from diesels, are other major sources of
16 PM25. Contributions from road dust and soils are relatively minor, typically constituting less
17 than 10% of PM2 5 in the studies shown in Table 3-8. Studies in the western United States shown
18 in Table 3-8 have found larger contributions from motor vehicles, fugitive dust and ammonium
19 nitrate. The most notable difference in the relative importance of major source categories of
20 PM25 shown in Table 3-8 and PM10 shown in Table 3-9 involves crustal material, (e.g., soil, road
21 dust), which represents about 40% on average of the total mass of PM10 in the studies shown in
22 Table 3-9. The fraction is higher in locations located away from specific sources such as sea
23 spray or smelters. Emissions of fugitive dust are concentrated mainly in the PM(10.25) size range.
24 The average fugitive dust source contribution is highly variable among sampling sites within the
25 same urban areas, as seen by differences between the Central Phoenix (33 //g/m3) and Scottsdale
26 (25 //g/m3) sites in Arizona, and it is also seasonally variable, as evidenced by the summer and
27 fall contributions at Rubidoux, CA. The variability in fugitive dust loadings reflects the sporadic
28 nature of its emissions and its short lifetime in the atmosphere.
29 In Table 3-9, primary motor vehicle exhaust contributions account for up to 40% of average
30 PM10 at many of the sampling sites. Vehicle exhaust contributions are also variable at different
31 sites within the same study area. The mean value and the variability of motor vehicle exhaust
March 2001 3-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 contributions reflects the proximity of sampling sites to roadways and traffic conditions during
2 the time of sampling. Many studies were conducted during the late 1980s, when a portion of the
3 vehicle fleet still used leaded gasoline. Pb and Br in motor vehicle emissions facilitated the
4 distinction of motor vehicle contributions from other sources. Vehicles using leaded fuels have
5 higher emission rates than vehicles using unleaded fuels. Pb also poisons automobile exhaust
6 catalysts and produces adverse human health effects. As a result, Pb virtually has been
7 eliminated from vehicle fuels. However, organic species have replaced Pb as a source marker for
8 motor vehicle emissions (e.g., Rogge et al., 1993).
9 Marine aerosol is found, as expected, at coastal sites such as Long Beach (average 3.8% of
10 total mass) and San Nicolas Island (25%). These contributions are relatively variable and are
11 larger at the more remote sites. Individual values reflect proximity to local sources. Of great
12 importance are the contributions from secondary ammonium sulfate in the eastern United States
13 and ammonium nitrate in the western United States. These are especially noticeable at sites in
14 California's San Joaquin Valley (Bakersfield, Crows Landing, Fellows, Fresno, Kern Wildlife,
15 and Stockton) and in the Los Angeles area.
16 Samples selected for chemical analysis are often biased toward the highest PM10 mass
17 concentrations in the studies shown in Table 3-9, so average source contribution estimates are
18 probably not representative of annual averages. For example, the study by Motallebi (1999)
19 considered only days when the PM10 concentration was greater than 40 //g/m3. Quoted
20 uncertainties in the estimated contributions of the individual sources shown in Table 3-6 range
21 from 10 to 50%. Uncertainties of source contribution estimates are not usually reported with the
22 average values summarized in Table 3-9. Estimates of standard errors are calculated in source
23 apportionment studies, and typically range from 15 to 30% of the source contribution estimate.
24 They are much higher when the chemical source profiles for different sources are highly
25 uncertain or too similar to distinguish one source from another.
26 Very few source apportionment studies using the CMB modeling technique have examined
27 the spatial variability of source contributions at different sites within an urban area. As can be
28 seen from Table 3-8, Dzubay et al. (1988) found a uniform distribution of sulfate among the NE
29 Airport in Philadelphia, PA; downtown Camden, NJ; and Clarksboro, NJ, during the summer of
30 1982. The longest distance between two monitoring sites (NE Airport and Clarksboro) was
31 approximately 40 km. Magliano et al. (1998) examined the spatial variability of PM10 source
March 2001 3-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 contributions at a number of sites in Fresno and Bakersfield, CA, during the winter of 1995-1996
2 and reported values for 1 day, December 27, 1995. During that day, mobile sources contributed
3 from 13.0 to!5.8 //g/m3, vegetation burning contributed from 5.1 to 11.1 //g/m3, ammonium
4 sulfate contributed 2.4 to 3.4 //g/m3, and ammonium nitrate contributed 19.3 to 24.6 //g/m3 to
5 PM10 at the sites in Bakersfield. Mobile sources contributed 13.9 to 22.5 //g/m3, vegetation
6 burning contributed 8.2 to 15.7 //g/m3, ammonium sulfate contributed 1.8 to 2.3 //g/m3, and
7 ammonium nitrate contributed 14.5 to 18.9 //g/m3 at the sites in Fresno. All of these components
8 are expected to be found mainly in the PM2 5 size fraction. As can be seen, source contributions
9 at different sites varied by factors of 1.2 to 2.2 in Bakersfield and by factors of 1.3 to 1.9 in
10 Fresno on that day.
11
12 3.3.2 Long-Range Transport of Particulate Matter from Sources Outside the
13 United States
14 Apart from sources within the continental United States, particulate matter can be brought
15 in by long-range transport from sources outside the United States. For example, the transport of
16 PM from uncontrolled biomass burning in Central America and southern Mexico resulted in
17 anomalously high PM levels observed in southern Texas and generally elevated PM
18 concentrations throughout the entire central and southeastern United States during the spring and
19 early summer of 1998. Windblown dust from individual dust storms in the Sahara desert has
20 been observed in satellite images as plumes crossing the Atlantic Ocean and reaching the
21 southeast coast of the United States (e.g., Ott et al., 1991). Dust transport from the deserts of
22 Asia across the Pacific Ocean also occurs (Prospero, 1996). Most dust storms in the deserts of
23 China occur in the spring following the passage of strong cold fronts after the snow has melted
24 and before a surface vegetation cover has been established. Strong winds and unstable
25 conditions result in the rapid transport of dust into the middle and upper troposphere (4 to 5 km
26 altitude), where it is transported by strong westerly winds out over the Pacific Ocean (Duce,
27 1995). Satellite images have been used to track the progress of a dust cloud from the Gobi desert
28 to the northwestern United States during the spring of 1998 (Husar et al., 2000).
29 Satellite images obtained at visible wavelengths cannot track mineral dust across the
30 continents because of a lack of contrast between the plume and the underlying surface. Other
31 means must be used to track the spread of North African dust through the eastern United States.
March 2001 3-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 Perry et al. (1997) used two criteria (PM25 soil concentration > 3 //g m"3 and Al/Ca > 3.8) to
2 distinguish between soil of local origin from soil originating in North Africa in characterizing the
3 sources of PM in aerosol samples collected in the IMPROVE (Interagency Monitoring of
4 Protected Visual Environments) network. Their analysis indicates that incursions of Saharan
5 dust into the continental United States have occurred, on average, about three times per year from
6 1992 to 1995. These events have persisted for about 10 days, principally during the summer.
7 As can be expected, the frequency of dust events is highest in the southeastern United States;
8 about half are observed only within the state of Florida, with these being associated with dense
9 hazes in Miami (Figure 3-19) during the summer (Prospero et al., 1987), such that African dust is
10 the dominant aerosol constituent in southern Florida during the summer (Prospero, 1999). The
11 mass median diameter of mineral dust over the oceans is typically between 2 and 3 //m (Duce,
12 1995). North African dust has been tracked as far as Illinois (Gatz and Prospero, 1996) and to
13 Maine (Perry et al., 1997). Larger scale events typically covered from 15 to 30% of the area of
14 the continental United States and resulted in increases of PM2 5 levels of 8.7 ± 2.3 //g m"3
15 throughout the affected areas, with mean maximum dust contributions of 19.7 ± 8.4 //g m"3
16 during these events and a peak contribution of 32 //g m"3 to 24-h average PM2 5 levels.
17 Husar et al. (2000) documented transport of dust from the Gobi and Taklimakan deserts to
18 North America during April 1998. The PM10 concentration averaged over 150 stations in
19 Washington, Oregon, California, Nevada, and Idaho reporting data to AIRS was 65 //g/m3
20 between April 26 and May 1, compared to about 20 //g/m3 during the rest of April and May.
21 Data from the IMPROVE network indicated that PM10 concentrations were over 100 //g/m3 in
22 central British Columbia, Washington State, and Oregon. The highest PM concentrations
23 observed were 120 //g/m3 for PM10 and 50 //g/m3 for PM2 5 at Chilliwack Airport in northwestern
24 Washington State (Figure 3-20). Aircraft measurements made over the northwestern United
25 States were consistent with a mass median diameter of the dust being between 2 and 3 //m.
26 Three-dimensional model simulations of the transport of inert tracers from Asia indicate that
27 substantial additions to PM concentrations also occurred throughout the north and mid-western
28 United States and southwestern Canada (Hanna et al., 1999).
29 Biomass burning for agricultural purposes occurs normally during the spring of each year in
30 Central America and southern Mexico. During the spring of 1988, fires burned uncontrollably
31 because of abnormally hot and dry conditions associated with the intense El Nino of 1997 to
March 2001 3-45 DRAFT-DO NOT QUOTE OR CITE
-------
Year
Figure 3-19. Monthly average Saharan dust components of the aerosol sampled in Miami,
FL, during 1974 to 1996.
Source: Prospero (1999).
1 1998. PM10 concentrations observed in the southern Rio Grande Valley were elevated
2 substantially during the passage northward of the biomass burning plume produced by these fires
3 as shown in Figure 3-21. Elevated PM2 5 and PM10 concentrations also were found as far north as
4 St. Louis, MO (Figure 3-22). As can be seen from Figure 3-21 and Figure 3-22, the elevations in
5 PM concentrations were limited in duration. However, uncontrolled wildfires occur in the
6 United States every year, but their effects on air quality throughout the United States still need to
7 be evaluated systematically. These fires can be widespread. For example, approximately
8 26,000 km2 were consumed during 2000 in the western United States.
9 Wildfires also occur in the boreal forests of northwestern Canada. Wotowa and Trainer
10 (2000) suggested that the plume from fires occurring in the Northwest Territories of Canada in
March 2001
3-46
DRAFT-DO NOT QUOTE OR CITE
-------
140
4/25/98 4/27/98 4/29/98 5/1/98 5/3/98 5/5/98
Figure 3-20. PM2 5 and PM10 concentrations measured at Chilliwack Airport, located in
northwestern Washington State, just before and during the Asian desert dust
episode of April and May 1998.
Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
1
2
3
4
5
6
early July 1995 may have extended throughout most of the eastern United States, resulting in
elevated levels of carbon monoxide (CO) and ozone. 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 about 5 //g/m3 in the
Southeast and increasing to close to 100 //g/m3 in the northern Plains States.
March 2001
3-47
DRAFT-DO NOT QUOTE OR CITE
-------
600
5/5 5/6 5/7 5/8 5/9 5/10 5/11 5/12 5/13 5/14 5/15 5/16 5/17 5/18 5/19 5/20 5/21 5/22 5/23 5/24 5/25 5/26 5/27 5/28
Time, days
Figure 3-21. Time series of 24-h average PM10 concentrations observed in the Rio Grande
Valley during May 1998.
Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
200
180-
160-
140-
120-
100-
80-
60-
40
20-|
0
Smoke
Event
Pm10 24 hr Standard
5 6 7 8 9 10 11 121314151617
May 1998
Figure 3-22. PM10 concentrations observed in St. Louis, MO, during May 1998.
Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
March 2001 3-48 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.4 EMISSIONS ESTIMATES AND THEIR UNCERTAINTIES
2 In principle, source contributions to ambient PM also could be estimated on the basis of
3 predictions made by chemistry-transport models (CTM) or even on the basis of emissions
4 inventories alone. Uncertainties in emissions inventories have arguably been regarded as
5 representing the largest source of uncertainty in CTMs (Calvert et al., 1993). Apart from
6 uncertainties in emission inventories, a number of other factors limit the ability of an emissions
7 inventory driven CTM to determine the effects of various sources on particle samples obtained at
8 a particular location. Air pollution model predictions represent averages over the area of a grid
9 cell, which in the case of the Urban Airshed Model typically has been 25 km2 (5 km x 5 km).
10 The contributions of sources to pollutant concentrations at a monitoring site are controlled
11 strongly by local conditions that cannot be resolved by an Eulerian grid-cell model. Examples
12 would be the downward mixing of tall stack emissions and deviations from the mean flow caused
13 by buildings. The impact of local sources at a particular point in the model domain may not be
14 predicted accurately, because their emissions would be smeared over the area of a grid cell or if
15 the local wind fields at the sampling point deviated significantly from the mean wind fields
16 calculated by the model. CTMs also have problems in predicting pollutant concentrations
17 because of uncertainties in vertical mixing and in predicting concentrations of pollutants from
18 stationary combustion sources resulting from uncertainties in estimates of plume rise.
19 Estimates of nationwide emissions of primary PM25 and gaseous precursors to secondary
20 PM formation are given in Section 3.4.1. Uncertainties in emissions estimates are discussed in
21 Section 3.4.2.
22
23 3.4.1 Emissions Estimates for Primary Particulate Matter and Sulfur
24 Dioxide, Nitrogen Oxides, and Volatile Organic Compounds in the
25 United States
26 Estimated emissions of primary PM25 from different sources in the United States are
27 summarized in Figure 3-23. The estimates are based on information presented in the EPA
28 National Air Pollutant Emission Trends Report, 1900-1998 (U.S. Environmental Protection
29 Agency, 2000b), to which the reader is referred for detailed tables showing trends in PM2 5
30 emissions from a number of source categories froml990 to 1998, descriptions of the
31 methodology used in the construction of these tables, and descriptions of the uncertainties
March 2001 3-49 DRAFT-DO NOT QUOTE OR CITE
-------
PM26Total Emissions (1998) = 8.4 Tg yr
-1
On-Road
2.4%
Incineration
0.6%
Industry
4.7%
Non-Road
4.9%
Fossil Fuel
Combustion
5.2%
Biomass
Burning
17.1%
Wind Erosion
9.5%
Agriculture
11.5%
Fugitive Dust
44.2%
Figure 3-23. 1998 directly emitted national particulate matter (PM25) emissions by
principal source categories for nonfugitive dust sources (see Section 3.4.2 for
discussion of uncertainties associated with emissions estimates).
Source: U.S. Environmental Protection Agency (2000b).
1
2
3
4
5
involved in the estimation process. This document also provides information about emissions of
PM10, sulfur dioxide (SO2), nitrogen oxides (NOX), volatile organic compounds (VOC), and
ammonia (NH3). Although uncertainties associated with these estimates are not given, a
discussion of uncertainties in emissions estimates is given in Section 3.4.2.
Estimated total nationwide emissions of primary PM25 were 8.4 Tg year"1 in 1998. The
category of fossil fuel combustion referred to in Figure 3-23 includes fossil fuel burning by
electric utilities, industry, and residences. The industry category includes contributions from
March 2001
3-50
DRAFT-DO NOT QUOTE OR CITE
-------
1 metals processing, petroleum refining, agricultural products processing, mining, and the storage
2 and transport of industrial goods. Incineration refers to the burning of nonbiomass waste by
3 residences and municipalities. The on-road vehicles category includes contributions from
4 gasoline- and diesel-powered vehicles. The nonroad engines and vehicles category includes
5 contributions from transportation (marine vessels, aircraft, trains, etc.); construction; and other
6 commercial, industrial, and recreational activities. Wind erosion refers to the raising of crustal
7 material by the wind. The biomass burning category includes contributions from residential
8 wood burning, open burning of vegetation, and forest fires. The agriculture category includes
9 contributions from emissions of crustal material related to the production of agricultural crops
10 and livestock. Fugitive dust refers mainly to crustal material raised by on-road and nonroad
11 vehicles during their operation. As can be seen from inspection of Figure 3-23, the raising of
12 crustal material by wind erosion, agriculture, and as fugitive dust emissions constitutes the
13 largest source (65.2%) of primary PM25 on a nationwide basis. Note that wind erosion emissions
14 are difficult to interpret, owing to the relatively short duration of wind gusts. Biomass burning
15 constitutes the second largest source (17.1%) of primary PM25. The gross composition of
16 emissions from most of these categories is summarized in Table 3-5 in the EPA report, National
17 Air Pollution Emission Trends, 1900-1998 (U. S. Environmental Protection Agency, 2000b).
18 Total emissions of primary PM25, as well as contributions from individual source categories,
19 were relatively constant over the period from 1990 to 1998 (U.S. Environmental Protection
20 Agency, 2000b).
21 Estimated contributions from individual sources to emissions of gaseous precursors to
22 secondary PM formation are summarized in Figure 3-24 for SO2, NOX , VOC, and NH3.
23 Information about the yield of PM formed during the oxidation of VOC is given in Section 3.4.
24 Although total emissions of gaseous precursors (SO2, NOX, VOC, and NH3) are shown in
25 Figure 3-24, it should be remembered that these values cannot be translated directly into
26 production rates of PM. Dry deposition and precipitation scavenging of some of these gases can
27 occur before they are converted to PM in the atmosphere. In addition, some fraction of these
28 gases are transported outside of the domain of the continental United States before being
29 oxidized. Likewise, emissions of these gases from areas outside the United States can result in
30 the transport of their oxidation products into the United States. Although the chemical oxidation
31 of SO2 will lead quantitatively to the formation of SO4=, the yield of paniculate matter from the
March 2001 3-51 DRAFT-DO NOT QUOTE OR CITE
-------
SO2 Total Emissions (1998)
= 18 TgyH
NO Total Emissions (1998)
Fuel Combustion-
Electrical Utility -
Fuel Combustion-
Industrial
Fuel Combustion-
Other
Metals Processing
On-Road and
Non-Road Engines
and Vehicles
On-Road and
Non-Road Engines
and Vehicles
= 21 Tgyr
.Fuel Combustion-
Electrical Utility
Fuel Combustion-
Other
Fuel Combustion-
Industrial
VOCs Total Emissions (1998)
= 17 Tgyr-1
NH3 Total Emissions (1998)
= 2.9 Tgyr-1
On-Road and
Non-Road Engines
and Vehicles
Solvent Utilization
On-Road and
Non-Road Engines
and Vehicles
Storage and
Transport
Chemicals Allied
Product Mfg.
Miscellaneous
(includes livestock
and fertilizer)
Figure 3-24. Nationwide emissions of SO2, NOX, VOC, and NH3 from various source
categories (see Section 3.4.2 for discussion of uncertainties associated with
emissions estimates).
Source: U.S. Environmental Protection Agency (2000b).
1 oxidation of VOC will be much less because only a small fraction of VOC react to form
2 particles, and those that do have efficiencies less than 10% (see Section 3.4).
3 The values shown in this section are based on annual totals. However, annual averages do
4 not reflect the seasonality of a number of emissions categories. Residential wood burning in
5 fireplaces and stoves, for example, is a seasonal practice that reaches its peak during cold
6 weather. Cold weather also affects motor vehicle exhaust particulate matter emissions, both in
7 terms of chemical composition and emission rates (e.g., Watson et al., 1990b; Huang et al.,
8 1994). Planting, fertilizing, and harvesting are also seasonal activities. Forest fires occur mainly
9 during the local dry season and during periods of drought. Maximum dust production by wind
10 erosion in the United States occurs during the spring, whereas the minimum occurs during the
March 2001
3-52
DRAFT-DO NOT QUOTE OR CITE
-------
1 summer (Gillette and Hanson, 1989). Efforts are being made to account for the seasonal
2 variations of emissions in the nationwide emissions inventories.
3 Trends in nationwide, annual average concentrations of PM10, and precursor gases (SO2,
4 NO2, and VOC) over the 10 years from 1989 to 1998 are shown in Table 3-10. As can be seen
5 from Table 3-10, there have been substantial decreases in the ambient concentrations of
6 PM10,SO2 and NO2. Not enough data are available to define trends in concentrations of VOC;
7 there also have been substantial decreases in the emissions of all the species shown in
8 Table 3-10, except for NO2, although its average concentration has decreased by 14%. These
9 entries suggest that decreases in the average ambient concentration of PM10 could have been
10 produced by both decreases in emissions of primary PM10 and the formation of secondary PM10.
11 The large reductions in ambient SO2 concentrations have resulted in reductions in sulfate
12 formation that would have been manifest in PM2 5 concentrations on the regional scale in the
13 eastern and central United States, where sulfate has constituted a larger fraction of PM2 5 than in
14 the West. Likewise, reductions in NO2 concentrations would have had a more noticeable impact
15 on PM25 concentrations in the western United States than in the eastern United States because
16 nitrate is a larger component of the aerosol in the western United States. Trends in aerosol
17 components (i.e., nitrate, sulfate, carbon, etc.) are needed for a more quantitative assessment of
18 the effects of changes in emissions of precursors. Measurements of aerosol nitrate and sulfate
19 concentrations have been obtained at North Long Beach and Riverside, CA, since 1978
20 (Dolislager and Motallebi, 1999). Downward trends in aerosol nitrate have tracked downward
21 trends in NOX concentrations, and SO2 and sulfate concentrations have both decreased. However,
22 the rate of decline of sulfate has been smaller than that of SO2 indicating the long range transport
23 of sulfate from outside the air shed may be an important source in addition to the oxidation of
24 locally generated SO2. There are a number of reasons why pollutant concentrations do not track
25 estimated reductions in emissions. Some of these reasons are related to atmospheric effects such
26 as meteorological variability and secular changes in the rates of photochemical transformations
27 and deposition (U.S. Environmental Protection Agency, 2000b). Other reasons are related to
28 uncertainties in ambient measurements and in emissions inventories.
29
30
March 2001 3-53 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-10. NATIONWIDE CHANGES IN AMBIENT CONCENTRATIONS AND
EMISSIONS OF PM10 AND GASEOUS PRECURSORS TO SECONDARY
PARTICULATE MATTER FROM 1989 TO 1998
% Change 1989-1999
Ambient Concentration(1) Emissions(1)
PM10 -25% -35%
PM2 5 Urban east - 5%(2) +5% (1990 to 1998)
Rural east +3%(2)
Rural west-11%(2)
SO47SO2 -39%(sulfate) -16%(SO2)
NO3-/NOX -14% (nitrate) +2% (NOX)
VOC — -20%
Source: (1) U. S. Environmental Protection Agency (2000a); (2) U. S. Environmental Protection Agency (2000d).
1 3.4.2 Uncertainties of Emissions Inventories
2 As described in the 1996 PM AQCD, it is difficult to assign uncertainties quantitatively to
3 entries in emissions inventories. Methods that can be used to verify or place constraints on
4 emissions inventories are sparse. In general, the overall uncertainty in the emissions of a given
5 pollutant includes contributions from all of the terms used to calculate emissions (i.e., activity
6 rates, emissions factors, control device efficiencies). Additional uncertainties arise during the
7 compilation of an emissions inventory because of missing sources and computational errors.
8 The variability of emissions can cause errors when annual average emissions are applied to
9 applications involving shorter time scales.
10 Activity rates for well-defined point sources (e.g., power plants) should have the smallest
11 uncertainty associated with their use, because accurate production records need to be kept.
12 On the other hand, activity rates for a number of really dispersed fugitive sources are difficult to
13 quantify. Emissions factors for easily measured fuel components that are released quantitatively
14 during combustion (e.g., CO2, SO2) should be the most reliable. Emissions of components
15 formed during combustion are more difficult to characterize as the emissions rates are dependent
March 2001 3-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 on factors specific to individual combustion units and on combustion stage (i.e., smoldering or
2 active). Although the AP-42 emissions factors (U.S. Environmental Protection Agency, 1995)
3 contain extensive information for a large number of source types, these data are very limited in
4 the number of sources sampled. The efficiency of control devices is determined by their design,
5 their age, their maintenance history, and operating conditions. It is virtually impossible to assign
6 uncertainties in control device performance because of these factors. It should be noted that the
7 largest uncertainties occur for those devices that have the highest efficiencies (>90%). This
8 occurs because the efficiencies are subtracted from one, and small errors in assigning efficiencies
9 can lead to large errors in emissions.
10 Ideally, an emissions inventory should include all major sources of a given pollutant. This
11 may be an easy task for major point sources. However, area sources of both primary PM and
12 precursors to secondary PM formation are more difficult to characterize than point sources and
13 thus, they require special emphasis when preparing emission inventories. Further research is
14 needed to better characterize the sources of pollutants to reduce this source of uncertainty. Errors
15 can arise from the misreporting of data, and arithmetic errors can occur in the course of
16 compiling entries from thousands of individual sources. A quality assurance program is required
17 to check for outliers and arithmetic errors.
18 Because of the variability in emissions rates, there can be errors in the application of
19 inventories developed on an annually averaged basis (as are the inventories shown in
20 Figures 3-23 to 3-24) to episodes occurring on much shorter time scales. As an example, most
21 modeling studies of air pollution episodes are carried out for periods of a few days.
22 Uncertainties in annual emissions were estimated to range from 4 to 9% for SO2 and from
23 6 to 11% for NOX in the 1985 NAPAP inventories for the United States (Placet et al., 1991).
24 Uncertainties in these estimates increase as the emissions are disaggregated both spatially and
25 temporally. The uncertainties quoted above are minimum estimates and refer only to random
26 variability about the mean, assuming that the variability in emissions factors was adequately
27 characterized, and that extrapolation of emissions factors to sources other than those for which
28 they were measured is valid. The estimates do not consider the effects of weather or variations in
29 operating and maintenance procedures.
30 Fugitive dust sources, as mentioned above, are extremely difficult to quantify, and stated
31 emission rates may represent only order-of-magnitude estimates. Although crustal dust
March 2001 3-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 emissions constitute over 60% of the total PM25 inventory, they constitute less than about 15% of
2 the source strengths inferred from the PM composition data shown in Table 3-1 and the receptor
3 modeling studies shown in Table 3-6. However, it should be remembered that secondary
4 components (sulfate, nitrate, and some fractions of organic carbon) often represent a significant
5 fraction of ambient samples. Therefore, this discrepancy is smaller than the factor of four that is
6 obtained by comparing primary dust emissions to the sum of primary and secondary components
7 in the ambient aerosol. The reasons for this apparent discrepancy are not clear. In addition to
8 errors in inventories or source apportionments, weather-related factors (wind speed and ground
9 wetness) and the dominance of local sources on spatial scales too small to be captured in
10 inventories may be involved. It should be remembered that dust emissions are dispersed widely
11 and are highly sporadic. Dust particles also have short atmospheric residence times and as a
12 result their dominance in emissions inventories may not be reflected in samples collected near
13 specific sources.
14 Although mineral dust sources account for most of the emissions, their contributions are
15 distributed much more widely than are those from combustion sources. Watson and Chow
16 (1999) examined a number of possible causes for this discrepancy. In large part, it is related to
17 the method used to obtain emissions factors for fugitive dust. The standard methods use data
18 obtained by particle monitors stacked at several elevations from 1 to 2 m up to 7 to 10 m above
19 the surface. However, small-scale turbulent motions, not stable winds, characterize atmospheric
20 flow patterns immediately adjacent to the surface (Garratt, 1994). The depth of this turbulent
21 layer is determined by surface roughness elements, and, if particle monitors are sampling within
22 this layer, there is a high probability of particles being entrained in turbulent eddies and
23 redepositing on the ground. In addition to the source sampling problem referred to above, it
24 should be remembered that dust often is raised in remote areas far removed from population
25 centers. Gravitational settling can be an important loss mechanism for particles larger than a few
26 microns in aerodynamic diameter and precipitation or scavenging by cloud droplets also removes
27 smaller particles during transport from the source area.
28 As rough estimates, uncertainties in emissions estimates could be as low as 10% for the
29 best characterized source categories, whereas emissions figures for windblown dust should be
30 regarded as order-of-magnitude estimates. Given (a) uncertainties in the deposition of SO2 and
31 its oxidation rate; (b) the variability seen in OC and EC emissions from motor vehicles, along
March 2001 3-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 with the findings from past verification studies for NMHC and CO to NOX ratios; (c) ranges of
2 values found among independent estimates for emissions of individual species (NH3, OC); and
3 (d) the predominance of fugitive emissions, PM emissions rates should be regarded as
4 order-of-magnitude estimates.
5 There have been few field studies designed to test emissions inventories observationally.
6 The most direct approach would be to use aircraft to obtain cross-sections of pollutants upwind
7 and downwind of major urban areas. The computed mass flux through a cross section of the
8 urban plume can then be equated to emissions from the city chosen. This approach has been
9 attempted on a few occasions. Results have been ambiguous because of contributions from
10 fugitive sources, nonsteady wind flows, and general logistic difficulties.
11
12
13 3.5 SUMMARY AND CONCLUSIONS
14 The recently deployed PM2 5 FRM network has returned data for a large number of sites
15 across the United States. Annual mean PM25 concentrations range from about 5 //g/m3 to over
16 20 //g/m3. In the eastern United States, the 1999 data indicate that highest quarterly mean
17 concentrations and maximum concentrations were reached during the summer. In the western
18 United States, highest quarterly mean values and maximum values occurred during the winter at
19 a number of sites, although there were exceptions to these general patterns. These findings are
20 generally consistent with those based on longer term data sets such as MAAQS in the eastern
21 United States and the CARB network of dichotomous samplers in California. The 1999 FRM
22 PM2 5 data indicate that, in general, PM2 5 concentrations are highly correlated among sites in
23 several MSAs (Atlanta, GA; Detroit, MI; Phoenix-Mesa, AZ; and Seattle-Bellevue-Everett,
24 WA), although there are exceptions to this rule. These findings are consistent with those of
25 earlier studies in Philadelphia, PA, and Los Angeles, CA, examining the spatial variability of
26 PM2 5 and its components. PM2 5 to PM10 ratios were generally higher in the East than in the
27 West, and values are consistent with those found in numerous earlier studies presented in the
28 1996PMAQCD.
29 Ambient particulate matter contains both primary and secondary components. The results
30 of ambient monitoring studies and receptor modeling studies in the eastern United States indicate
31 that PM2 5 is dominated by secondary components. Depending on the origin of OC in ambient
March 2001 3-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 samples, PM2 5, on average, also may be dominated by secondary components throughout the rest
2 of the United States. Primary constituents represent smaller but still important components of
3 PM25, on average. Crustal materials constitute the largest fraction of PM(10_2 5) throughout the
4 United States. Data collected in several airsheds including the Los Angeles Basin, Bakersfield
5 and Fresno, CA; and Philadelphia, PA, suggest that secondary PM components are more
6 uniformly distributed than are primary components. Compositional data obtained at multiple
7 sites in other urban areas are sparse.
8 Because of the complexity of the composition of ambient PM2 5 and PM(10_2 5), sources are
9 best discussed in terms of individual constituents of both primary and secondary PM25 and
10 PM(10_25). Each of these constituents can have anthropogenic and natural sources, as shown in
11 Table 3-7. The distinction between natural and anthropogenic sources is not always obvious.
12 Although windblown dust might seem to be the result of natural processes, highest emission rates
13 are associated with agricultural activities in areas that are susceptible to periodic drought.
14 Examples include the dust bowl region of the midwestern United States and the Sahel of Africa.
15 Also, most forest fires in the United States could be classified as human in origin, either through
16 prescribed burning; by accident; or through forest management practices that allow the buildup
17 of combustible material, thereby increasing the likelihood of fire from whatever cause.
18 Emissions inventories are generally not the most appropriate way to apportion material in
19 ambient samples. Receptor modeling has proven to be an especially valuable tool in this regard.
20 Receptor modeling can help bound emission inventories and establish uncertainty estimates.
21 Compositional profiles developed for receptor modeling applications are perhaps the most
22 accessible and reliable means of characterizing the composition of emissions. Techniques are
23 under development to use emission inventories and receptor modeling to reduce the uncertainty
24 in the overall source apportionment (U. S. Environmental Protection Agency, 2000e).
25 The results of receptor modeling studies throughout the United States indicate that the
26 combustion of fossil and biomass fuels is a major source of PM25. Fugitive dust, found mainly in
27 the PM(10_2 5) range size, represents the largest source of PM10 in many locations in the western
28 United States. Quoted uncertainties in source apportionments of constituents in ambient aerosol
29 samples typically range from 10 to 50%. It is apparent that a relatively small number of source
30 categories, compared to the total number of chemical species that typically are measured in
March 2001 3-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 ambient monitoring-source receptor model studies, are needed to account for most of the
2 observed mass of PM in these studies.
3 Improvements in the ability of receptor models to allocate sources of ambient PM continue
4 to be made. Recently developed techniques such as positive matrix factorization allow
5 quantitative determinations of contributions from different categories of PM sources to be made
6 on the basis of ambient data alone. Improvements in the accuracy of PM emissions inventories
7 also continue to be made. Recent studies have identified causes of the overrepresentation of
8 crustal material in emissions inventories. The causes are related to the neglect of near-source PM
9 deposition in the development of emissions factors.
10 As seen in Table 3-7, emissions of mineral dust, organic debris, and sea spray are
11 concentrated mainly in the coarse fraction of PM10 (>2.5 //m aerodynamic diameter). A small
12 fraction of this material is in the PM2 5 size range (< 2.5 //m aerodynamic diameter).
13 Nevertheless, concentrations of crustal material can be appreciable, especially during dust events.
14 It also should be remembered that much of the Saharan dust reaching the United States is in the
15 PM25 size range. Emissions from combustion sources (mobile and stationary sources and
16 biomass burning) are also predominantly in the PM25 size range.
17 Uncertainties in emissions inventories are difficult to quantify. They may be as low as 10%
18 for well-defined sources (e.g., for SO2) and may range up to a factor of 10 or so for windblown
19 dust. As a rule, total PM emissions rates should be regarded as order-of-magnitude estimates.
20 Because of the large uncertainty associated with emissions of suspended dust, trends of total
21 PM2 5 PM10 emissions should be viewed with caution, and emissions of specific components are
22 best discussed on an individual basis. Receptor modeling, especially when coupled to accurate
23 measurements of the composition of emissions, can be useful in providing bounds for emission
24 inventories.
25 Although most emphasis in this chapter has been on sources within the United States,
26 it also should be remembered that sources outside the United States contribute to ambient PM
27 levels that can, at times, exceed the ambient NAAQS level for PM. Perry et al. (1997) have
28 found that the highest concentrations of mineral dust in the PM2 5 fraction are found in the eastern
29 United States during the summer and not in arid areas of the western United States. This dust
30 has been emitted in the Sahara Desert and then transported across the Atlantic Ocean.
31 Large-scale dust storms in the deserts of central Asia recently have been found to contribute to
March 2001 3-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM levels in the Northwest on an episodic basis. Uncontrolled biomass burning in central
2 America and Mexico may have contributed to elevated PM levels that exceeded the daily
3 NAAQS level for PM in Texas. Wildfires throughout the United States, Canada, Mexico, and
4 Central America all contribute to background concentrations of PM in the United States.
5
March 2001 3-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Andreae, M. O. (1991) Biomass burning: its history, use, and distribution and its impact on environmental quality
3 and global climate. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
4 implications. Cambridge, MA: MIT Press; pp. 1-21.
5 Bahadori, T.; Suh, H.; Koutrakis, P. (2000a) A cross-study analysis of ambient concentrations of fine mass. J. Air
6 Waste Manage. Assoc.: in press.
7 Bahadori, T.; Suh, H.; Koutrakis, P. (2000b) PM2 5 and PM10 concentrations in four metropolitan cities in eastern
8 United States. J. Air Waste Manage. Assoc.: in press.
9 Cahill, T. A. (1996) [Data from IMPROVE data base.] Fort Collins, CO: Air Resource Specialists, Inc.
10 California Air Resources Board. (1995) [Data base provided by John R. Holmes, Chief, Research Division.]
11 Sacramento, CA: California Air Resources Board.
12 Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some
13 reflections on controlling vehicle emissions. Science (Washington, DC) 261: 37-45.
14 Chow, J. C.; Watson, J. G.; Frazier, C. A.; Egami, R. T.; Goodrich, A.; Ralph, C. (1988) Spatial and temporal
15 source contributions to PM10 and PM2 5 in Reno, NV. In: Mathai, C. V.; Stonefield, D. H., eds. PM.10:
16 implementation of standards, an APCA/EPA international specialty conference; February; San Francisco,
17 CA. Pittsburgh, PA: Air Pollution Control Association; pp. 438-457. (APCA transactions series TR-13).
18 Chow, J. C.; Watson, J. G.; Richards, L. W.; Haase, D. L.; McDade, C.; Dietrich, D. L.; Moon, D.; Sloane, C.
19 (1991) The 1989-90 Phoenix PM10 study. Volume II: source apportionment. Final report. Phoenix, AZ:
20 Arizona Department of Environmental Quality; Desert Research Institute document no. 8931.6F1.
21 Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Frazier, C. A.; Hinsvark, B. A.; Prichett, L. C.; Neuroth, G. R.
22 (1992a) Wintertime PM10 and PM2 5 chemical compositions and source contributions in Tucson, Arizona.
23 In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source controls, an
24 A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air &
25 Waste Management Association; pp. 231-243. (A&WMA transactions series no. 22).
26 Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W.
27 (1992b) PM10 source apportionment in California's San Joaquin Valley. Atmos. Environ. Part A
28 26: 3335-3354.
29 Chow, J. C.; Liu, C. S.; Cassmassi, J.; Watson, J. G.; Lu, Z.; Pritchett, L. C. (1992c) A neighborhood-scale study of
30 PM10 source contributions inRubidoux, California. Atmos. Environ. Part A 26: 693-706.
31 Darlington, T. L.; Kahlbaum, D. F.; Heuss, J. M.; Wolff, G. T. (1997) Analysis of PM10 trends in the United States
32 from 1988 through 1995. J. Air Waste Manage. Assoc. 47: 1070-1078.
33 Dolislager, L. J.; Motallebi, N. (1999) Characterization of paniculate matter in California. J. Air Waste Manage.
34 Assoc. 49: PM-45-56.
3 5 Duce, R. A. (1995) Sources, distributions, and fluxes of mineral aerosols and their relationship to climate.
3 6 In: Charlson, R. J.; Heintzenberg, J., eds. Aerosol forcing of climate: report of the Dahlem workshop on
37 aerosol forcing of climate; April 1994; Berlin, Federal Republic of Germany. Chichester, United Kingdom:
38 John Wiley & Sons, Ltd.; pp. 43-72.
39 Dzubay, T. G.; Stevens, R. K.; Gordon, G. E.; Olmez, I.; Sheffield, A. E.; Courtney, W. J. (1988) A composite
40 receptor method applied to Philadelphia aerosol. Environ. Sci. Technol. 22: 46-52.
41 Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the United
42 States of America. Atmos. Environ. 28: 1009-1019.
43 Fitz-Simons, T. S.; Mathias, M.; Rizzo, M. (2000) Analyses of 1999 PM data for the PM NAAQS review. U.S.
44 Environmental Protection Agency, Office of Air Quality Planning and Standards; September 29.
45 Garratt, J. R. (1994) The atmospheric boundary layer. Cambridge, United Kingdom: Cambridge University Press.
46 (Cambridge atmospheric and space science series).
47 Gatz, D. F.; Prospero, J. M. (1996) A large silicon-aluminum aerosol plume in central Illinois: North African Desert
48 dust? Atmos. Environ. 30: 3789-3799.
49 Gillette, D. A.; Hanson, K. J. (1989) Spatial and temporal variability of dust production caused by wind erosion in
50 the United States. J. Geophys. Res. [Atmos.] 94: 2197-2206.
51 Gillies, J. A.; Gertler, A. W.; Lowenthal, D.; Abu-Allaban, M.; Jennison, B.; Goodrich, A. (2000) Enhanced PM25
52 source apportionment using chemical mass balance receptor modeling and scanning electron microscopy.
53 Aerosol Sci. Technol.: submitted.
March 2001 3-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 Glover, D. M; Hopke, P. K.; Vermette, S. J.; Landsberger, S.; D'Auben, D. R. (1991) Source apportionment with
2 site specific source profiles. J. Air Waste Manage. Assoc. 41: 294-305.
3 Gray, H. A.; Landry, B.; Liu, C. S.; Henry, R. C.; Cooper, J. A.; Sherman, J. R. (1988) Receptor modeling for PM10
4 source apportionment in the South Coast Air Basin of California. In: Mathai, C. V.; Stonefield, D. H., eds.
5 PM.10: implementation of standards, an APCA/EPA international specialty conference; February; San
6 Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 399-418. (APCA transactions series
7 TR-13).
8 Hanna, A.; Mathur, R.; Alapaty, K.; Pinto, J. (1999) Modeling the episodic transport of air pollutants from Asia to
9 North America. http://envpro.ncsc.org/projects/MITP/.
10 Hopke, P. K.; Wlaschin, W.; Landsberger, S.; Sweet, C.; Vermette, S. J. (1988) The source apportionment of PM10
11 in south Chicago. In: Mathai, C. V.; Stonefield, D. H., eds. PM.10: implementation of standards, an
12 APCA/EPA international specialty conference; February; San Francisco, CA. Pittsburgh, PA: Air Pollution
13 Control Association; pp. 484-494. (APCA transactions series TR-13).
14 Houck, J. E.; Rau, J. A.; Body, S.; Chow, J. C. (1992) Source apportionment: Pocatello, Idaho PM10 nonattainment
15 area. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source controls, an
16 A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air &
17 Waste Management Association; pp. 219-230. (A&WMA transactions series no. 22).
18 Huang, X.; Olmez, I.; Aras, N. K.; Gordon, G. E. (1994) Emissions of trace elements from motor vehicles: potential
19 marker elements and source composition profile. Atmos. Environ. 28: 1385-1391.
20 Husar, R. B.; Schichtel, B. A.; Falke, S. R.; Li, F.; Wilson, W. E.; Pinto, J.; Malm, W. C.; Fox, D. G.; Feldman,
21 G. C.; McClain, C.; Kuring, N.; Holben, B. N.; Vermote, E. F.; Herman, J. R.; Elvidge, C. D. (2000) The
22 impact of the 1998 Central American smoke on the atmospheric environment of eastern North America.
23 J. Geophys. Res.: submitted.
24 IMPROVE: interagency monitoring of protected visual environments [database]. (1998) [Data on gross chemical
25 composition of PM25 particles in rural areas of the eastern and western United States as of 1998]. Fort
26 Collins, CO: National Park Service (NFS); Cooperative Institute for Research in the Atmosphere (CIRA).
27 Available at: http://vista.cira.colostate.edu/improve/ [2001, January 26].
28 Inhalable Paniculate Network. (1985) [Inhalable Paniculate Network Database.] Research Triangle Park, NC:
29 U.S. Environmental Protection Agency, National Exposure Research Laboratory; Jose M. Sune, contact.
30 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
31 segregated mass concentration data and ultrafine particle number concentrations in urban areas. Atmos.
32 Environ. 33:2907-2913.
33 Kim, B. M.; Zeldin, M. D.; Liu, C. S. (1992) Source apportionment study for state implementation plan
34 development in the Coachella Valley. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional
35 paniculate source controls: an A&WMA/EPA international specialty conference, v. II; January; Scottsdale,
36 AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 979-991. (A&WMA transactions series
37 no. 22).
38 Koutrakis,P. (1995) [PM2 5 and PM10 data for Philadelphia, 1992-1995]. Boston, MA: Harvard School of Public
39 Health.
40 Lawson, D. R.; Smith, R. E. (1998) The northern front range air quality study: a report to the Governor and General
41 Assembly. Fort Collins, CO: Colorado State University; December.
42 Magliano, K. L. (1988) Level 1 PM10 assessment in a California air basin. In: Mathai, C. V; Stonefield, D. H., eds.
43 PM.10: implementation of standards, an APCA/EPA international specialty conference; February;
44 San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 508-517. (APCA transactions
45 series TR-13).
46 Magliano, K. L. (1998) Chemical mass balance modeling of data from the 1995 integrated monitoring study.
47 Sacramento, CA: California Air Resources Board.
48 McMurry, P. H.; Woo, K. S.; Weber, R.; Chen, D.-R.; Pui, D. Y. H. (2000) Size distributions of 3 to 10 nm
49 atmospheric particles: implications for nucleation mechanisms. Philos. Trans. R. Soc. London Ser. A 358:
50 2625-2642.
51 Motallebi, N. (1999) Wintertime PM2 5 and PM10 source apportionment at Sacramento, California. J. Air Waste
52 Manage. Assoc. 49: PM-25-34.
53 Neas, L. M. (1996) [Data provided for 1986 from Harvard six cities data base.] Boston, MA: Harvard School of
54 Public Health.
March 2001 3-62 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ott, S.-T.; Ott, A.; Martin, D. W.; Young, J. A. (1991) Analysis of a trans-Atlantic Saharan dust outbreak based on
2 satellite and GATE data. Mon. Weather Rev. 119: 1832-1850.
3 Paatero, P.; Tapper, U. (1993) Analysis of different modes of factor analysis as least squares fit problems. Chemom.
4 Intelligent Lab. Syst. 18: 183-194.
5 Paatero, P.; Tapper, U. (1994) Positive matrix factorization: a non-negative factor model with optimal utilization of
6 error estimates of data values. Environmetrics 5: 111-126.
7 Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Dutcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
8 dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
9 Pinto, J. P.; Pleasant, M; Kellogg, R.; Torok, Sz.; Stiles, D.; Willis, R. D.; Koutrakis, P.; Allen, G. A. (1995)
10 Chemical characterization of the ambient aerosol in Philadelphia. Presented at: Paniculate matter: health and
11 regulatory issues: an A&WMA international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA:
12 Air & Waste Management Association; submitted abstract.
13 Placet, M.; Battye, R. E.; Fehsenfeld, F. C.; Bassett, G. W. (1991) Emissions involved in acidic deposition
14 processes. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume I: emissions,
15 atmospheric processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment
16 Program. (State of science and technology report no. 1).
17 Prospero, J. M. (1996) The atmospheric transport of particles to the ocean. In: Ittekkot, V.; Honjo, S.; Depetris, P.
18 J., eds. Particle flux in the ocean. New York, NY: John Wiley & Sons, Ltd.; pp. 19-52. (SCOPE report no.
19 57).
20 Prospero, J. M. (1999) Long-range transport of mineral dust in the global atmosphere: impact of African dust on the
21 environment of the southeastern United States. Proc. Natl. Acad. Sci. U.S.A. 96: 3396-3403.
22 Prospero, J. M.; Nees, R. T.; Uematsu, M. (1987) Deposition rate of paniculate and dissolved aluminum derived
23 from Saharan dust in precipitation at Miami, Florida. J. Geophys. Res. [Atmos.] 92: 14,723-14,731.
24 Ramadan, Z.; Song, X.-H.; Hopke, P. K. (2000) Identification of sources of Phoenix aerosol by positive matrix
25 factorization. J. Air Waste Manage. Assoc. 50: 1308-1320.
26 Rodes, C. E.; Evans, E. G. (1985) Preliminary assessment of 10 um paniculate sampling at eight locations in the
27 United States. Atmos. Environ. 19: 293-303.
28 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993) Sources of fine organic
29 aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci.
30 Technol. 27:636-651.
31 Ryan, W. M.; Badgett-West, C. R.; Holtz, D. R.; Peters, T. A.; Cooper, J. A.; Ono, D. (1988) Reconciliation of
32 receptor and dispersion modeling impacts of PM10 in Hayden, Arizona. In: Mathai, C. V.; Stonefield, D. H.,
33 eds. PM.10: implementation of standards, an APCA/EPA international specialty conference; February;
34 San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 419-429. (APCA transactions
35 series TR-13).
36 Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurik, M. A.; Cass, G. R. (1996) Source apportionment of
37 airborne paniculate matter using organic compounds as tracers. Atmos. Environ. 30: 3837-3855.
38 Skidmore, L. W.; Chow, J. C.; Tucker, T. T. (1992) PM10 air quality assessment forthe Jefferson County, Ohio air
39 quality control region. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source
40 controls, an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ. Pittsburgh, PA:
41 Air and Waste Management Association; pp. 1016-1031. (A&WMA transactions series no. 22).
42 Spengler, J. D.; Briggs, S. L. K.; Ozkaynak, H. (1986) Relationship between TSP measurements and
43 size-fractionated particle mass measurements in six cities participating in the Harvard Air Pollution Health
44 Study. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
45 and Standards; EPA cooperative agreement no. CR-812667-01-0; December 5.
46 Suh, H. H.; Allen, G. A.; Koutrakis, P.; Burton, R. M. (1995) Spatial variation in acidic sulfate and ammonia
47 concentrations within metropolitan Philadelphia. J. Air Waste Manage. Assoc. 45: 442-452.
48 Tanner, R. L.; Parkhurst, W. J. (2000) Chemical composition of fine particles in the Tennessee Valley region. J. Air
49 Waste Manage. Assoc. 50: 1299-1307.
50 Taylor, C. A., Jr.; Stover, C. A.; Westerdahl, F. D. (1998) Speciated fine particle (<2.5 um aerodynamic diameter)
51 and vapor-phase acid concentrations in southern California. Presented at: 91st annual meeting & exhibition
52 of the Air & Waste Management Association; June; San Diego, CA. Pittsburgh, PA: Air & Waste
5 3 Management Association; paper no. WA7401.
54
March 2001 3-63 DRAFT-DO NOT QUOTE OR CITE
-------
1 Thanukos, L. C.; Miller, T.; Mathai, C. V.; Reinholt, D.; Bennett, J. (1992) Intercomparison of PM10 samplers and
2 source apportionment of ambient PM10 concentrations at Rillito, Arizona. In: Chow, J. C.; Ono, D. M, eds.
3 PM10 standards and nontraditional paniculate source controls, an A&WMA/EPA international specialty
4 conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association;
5 pp. 244-261. (A&WMA transactions series no. 22).
6 U.S. Environmental Protection Agency. (1995) Compilation of air pollutant emission factors. Volume I: stationary
7 point and area sources. 5th ed. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
8 report no. AP-42-ED-5. Available from: NTIS, Springfield, VA; PB95-196028.
9 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
10 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
11 U.S. Environmental Protection Agency. (2000a) Aerometric Information Retrieval System (AIRS). Research
12 Triangle Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/ [2000,
13 October 6].
14 U.S. Environmental Protection Agency. (2000b) National air pollutant emission trends, 1900-1998. Research
15 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-00-002. Available:
16 http://www.epa.gov/ttn/chief/trends/trends98/ [2001, January 30].
17 U.S. Environmental Protection Agency. (2000c) Air quality criteria for carbon monoxide. Research Triangle Park,
18 NC: National Center for Environmental Assessment; report no. EPA/600/P-99/001F. Available:
19 www.epa.gov/ncea/co/ [2000, October 6].
20 U.S. Environmental Protection Agency. (2000d) National air quality and emissions trends report, 1998. Research
21 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/454/R-00-003. Available
22 from: NTIS, Springfield, VA; PB2000-104960.
23 U.S. Environmental Protection Agency. (2000e) Guidance for demonstrating attainment of air quality goals for
24 PM2 5 and regional haze. March.
25 Vermette, S. J.; Williams, A. L.; Landsberger, S. (1992) PM10 source apportionment using local surface dust
26 profiles: examples from Chicago. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional
27 paniculate source controls, an A&WMA/EPA international specialty conference, v. I; January; Scottsdale,
28 AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 262-278. (A&WMA transactions series
29 no. 22).
30 Watson, J. G.; Chow, J. C. (1999) Reconciling urban fugitive dust emissions inventory and ambient source
31 contribution estimates: summary of current knowledge and needed research [draft]. Reno, NV: Desert
32 Research Institute; DRI document no. 6110.4D2.
33 Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
34 (1990a) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
35 Watson, J. G.; Chow, J. C.; Pritchett, L. C.; Houck, J. A.; Ragazzi, R. A.; Burns, S. (1990b) Chemical source
3 6 profiles for paniculate motor vehicle exhaust under cold and high altitude operating conditions. Sci. Total
37 Environ. 93: 183-190.
38 Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M.; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994)
3 9 Chemical mass balance source apportionment of PM10 during the Southern California Air Quality Study.
40 Aerosol Sci. Technol. 21: 1-36.
41 Wilson, W. E.; Stockburger, L. (1990) Diurnal variations in aerosol composition and concentration. In: Masuda, S.;
42 Takahashi, K., eds. Aerosols: science, industry, health and environment: proceedings of the third
43 international aerosol conference, v. 2; September; Kyoto, Japan. Oxford, United Kingdom: Pergamon Press
44 pic; pp. 962-965.
45 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
46 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
47 Wongphatarakul, V.; Friedlander, S. K.; Pinto, J. P. (1998) A comparative study of PM25 ambient aerosol chemical
48 databases. Environ. Sci. Technol. 32: 3926-3934.
49 Woo, K. S.; Chen, D. R.; Pui, D. Y. H.; McMurry, P. H. (2000) Measurement of Atlanta aerosol size distributions:
50 observations of ultrafine particle events. Aerosol Sci. Technol.: accepted.
March 2001 3-64 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 3A
2
3 Organic Composition of Particulate Matter
4
5
6 Although organic compounds typically constitute approximately 10 to 70% of the total dry
7 fine particle mass in the atmosphere, organic PM concentrations, composition, and formation
8 mechanisms are poorly understood. This is because paniculate organic matter is an aggregate of
9 hundreds of individual compounds spanning a wide range of chemical and thermodynamic
10 properties (Saxena and Hildemann, 1996). The presence of multiphase or "semi-volatile"
11 compounds complicates collection of organic particulate matter. Furthermore, no single
12 analytical technique currently is capable of analyzing the entire range of compounds present.
13 Rigorous analytical methods frequently identify only 10 to 20% of the organic mass on the
14 molecular level (Rogge et al., 1993). The data shown in Appendix 3A are meant to complement
15 the data given for the inorganic components of particles in Appendix 6 A of the 1996 PM AQCD
16 (U. S. Environmental Protection Agency, 1996). Table 3A-1 lists a number of recent urban and
17 some rural measurements of parti culate organic and elemental carbon in //g of carbon/m3.
18 Emphasis was placed on measurements published after 1995. The analysis method and artifact
19 correction procedure, if any, are indicated. Table 3A-2 presents information on recent (post
20 1990) studies concerning concentrations of parti culate organic compounds found at selected U.S.
21 sites.
March 2001 3 A-1 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3A-1. PARTICIPATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in ^ C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
to
o
C2
OJ
to
o
£
H
I
0
o
o
H
O
o
H
W
O
O
H
W
Reference
URBAN PM2 5
Offenberg and Baker
(2000)
Allen etal. (1999)
Pedersen etal. (1999)
IMPROVE (2000)
Lewtas etal. (2001)
Khwaja (1995)
Christoforou et al.
(2000)
Turpin and Huntzicker
(1995)
Location Dates
Chicago, IL July 1994;
Jan 1995
UniontownPA July-Aug 1990
Boston, MA Jan-Dec 1995
Reading, MA (suburban)
Quabbin, MA (rural)
Rochester, NY (urban)
Brockport, NY (rural)
Washington, DC 1994-1998
Seattle, WA
Seattle, WA Apr-May 1999
Schenectady, NY Oct 24-26, 1991
Azusa, CA Jan-Dec 1993
Long Beach, CA
Central, LA
Rubidoux, LA
San Nicolas, LA
Claremont, CA Jun-Sept
Long Beach, CA Nov-Dec 1987
OCMean
(Max)
(0.8-8.4)a
5.8
4.0
2.8
3.3
2.7
3.4
1.8
8.0
9.4
8.9
12.3
9.7
1.6
na (29.4)
na (62.6)
EC Mean TC Mean
(Max) (Max)
2.2 (3.8)
1.7
(0.4-3. 5)a
1.3(3.1)
1.7
0.7
0.5
0.7
0.5
1.1
0.3
1.4
23.2 (49.9)
1.3
1.8
2.7
1.5
1.5
na (9.0)
na (24.6)
Avg.
Time Notes
12 h PM12; Imp; TOT
PMM; Imp; TOT
3 h PM2 5; DQQ; TORb
10 min Aeth
24 h PM20; Q; TOT
24 h PM25; QQ; TOR
23 h PM2 5; DQA; EGAC
6 h PM1.0; Q; Th
24 h PM2 1, Q; TOR
2h PM25;Q+TQ;TOTd
2-6 h
-------
TABLE 3A-1 (cont'd). PARTICULATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in ^ C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
to
o
o
OJ
>
OJ
o
g
2
6
o
0
H
O
o
H
W
O
O
H
W
Reference Location Dates
RURAL PM2 5
Klinedinst and Currie Welby, CO Dec 1996-Jan
(1999) Brighton, CO 1997
Andrews et al. (2000) Look Rock, Smoky July-Aug 1995
Mountains, TN
Malm and Gebhart Tahoma Wood, WA June-Aug 1990
(1996)
IMPROVE (2000) Three Sisters Wilderness, OR 1994-1998
Rocky Mountains, CO
Brigantine, NJ
Acadia, MA
Jefferson: James River Face
Wilderness, VA
Glacier, MT
Hegg et al. ( 1 997) 1 50 km East of Mid-Atlantic July 1 996
Coast
(0.02-4 km altitude)
Cui et al. (1997) Meadview, AZ Aug 6-15, 1992
OCMean
(Max)
5.6(13.4)
3.6 (6.4)
2.2
2.7
1.2
2.6 (7.4)
0.9
1.0
2.0
1.2
3.8
2.4
EC Mean TC Mean Avg.
(Max) (Max) Time Notes
3.3(8.1) 6h PM25;Q;TOR
1.9 (3.6)
0.4 12 h (day) PM2 ^, QQ; TOR6
0.1 PM21;Q+TQ;TORd
0.2 PM; 8; Imp; TMO
0.7 (2.2) 12 h PM2 5; QQ; TORf
0.2 24 h PM25;QQ;TOR
0.2
0.5
0.2
0.7
0.4
2.9 (5.4) PMj 0; QQ ; EGA6
3.0 12 h PM25;VDQA;EGAC
-------
TABLE 3A-1 (cont'd). PARTICULATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in ^ C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
to
o
>
-^
a
%
H
6
o
0
H
/O
o
H
W
O
Q
H
W
Reference Location Dates
RURAL PM2 5
Chow et al. ( 1 996) Point Reyes, CA July-Aug 1 990
Altamont Pass, CA
Pacheco Pass, CA
Crows Landing, CA
Academy, CA
Button- Willow, CA
Edison, CA
Caliente, CA
Sequoia, CA
Yosemite, CA
Malm and Day (2000) Grand Canyon, AZ July-Aug 1998
PM10
Omar etal. (1999) Bondville, IL Jan-Dec 1994
Gertler et al. (1995) Bullhead City, AZ Sept 1988-Oct 1989
Chow et al. ( 1 996) Santa Barbara, CA (urban) Jan-Dec 1989
Santa Maria, CA (urban)
Santa Ynez, CA (airport)
Gaviota, CA (rural SB)
Watt Road, CA (rural SB)
Anacapa Island, CA
OC Mean
(Max)
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)
1.1(1.6)
2.6
6.0 (16.0)
EC Mean TC Mean Avg.
(Max) (Max) Time Notes
0.4 (0.6) 5-7 h PM2 5; Q+TQ; TORg
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)
0. 10 (0.3) 24 h PM2 5; QQ; TORf
0.2 24-48 h PM10; Q; TOR
1.9(4.0) 24 h PM10;Q;TOR
8.8 24 h PM10; Q; TOR
4.6
3.5
3.4
2.1
3.1
-------
TABLE 3A-1 (cont'd). PARTICULATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in ^g C/m3)
3
to
2 Reference
Lioy and Daisey
(1987)
BASED ON STUDIES PUBLISHED AFTER 1995
Location
Newark, NJ
Elizabeth, NJ
Camden, NJ
Dates
1982:
Summer
Winter
Summer
Winter
Summer
Winter
OCMean
(Max)
4.1
5.9
2.1
7.1
2.2
5.2
EC Mean TC Mean Avg.
(Max) (Max) Time
3.0
3.3
1.7
2.3
1.3
2.0
Notes
PM15; Q
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
A limited amount of rural data is presented. In some cases, total carbon (TC = OC + EC) is reported. OC concentrations must be multiplied by the average molecular weight per carbon weight to
convert to mass of particulate organic compounds. The location and dates over which sampling occurred are provided. Averaging time refers to the sampling duration. Sampling method: Q - quartz
fiber filter; QQ - two quartz fiber filters in series; Q+TQ - a quartz fiber filter in one port and a Teflon followed by a quartz filter in a parallel port; Imp - cascade impactor; DQQ - denuder followed
by two quartz fiber filters; DQA - denuder followed by quartz fiber filter and adsorbent; VDQA - virtual impactor inlet followed by denuder, quartz filter, and adsorbent. Analysis method is reported
as follows: TOR - thermal optical reflectance; TOT - thermal optical transmittance; TMO - thermal MnO2 oxidation; EGA - evolved gas analysis; Th - Thermal analysis; Aeth - Aethalometer.
na - data not available.
"Range is provided. It should be noted that samples were collected only during elevated pollution episodes and are not representative of average concentrations.
bParticulate OC was considered to be the sum of front and back quartz fiber filters.
°Sum of adsorbent and filter after correction for inlet losses and denuder efficiency.
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.
-------
g.
to
o
o
TABLE 3A-2. PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
n-Alkanes
n-tricosane
n-tetracosane
n-pentacosane
n-hexacosane
n-heptacosane
n-octacosane
n-nonacosane
n-triacontane
n-hentriacontane
n-dotriacontane
n-tritriacontane
n-tetratriacontane
Total n-alkanes
n-Alkanoic Acids
n-nonanoic acid
n-decanoic acid
n-undecanoic acid
n-dodecanoic acid
n-tridecanoic acid
n-tetradecanoic acid
n-pentadecanoic acid
n-hexadecanoic acid
(palmitic acid)
n-heptadecanoic acid
n-octadecanoic acid
(stearic acid)
n-nonadecanoic acid
n-eicosanoic acid
n-heneicosanoic acid
n-docosanoic acid
n-tricosanoic acid
n-tetracosanoic acid
Los Angeles,
CA
6.7
6.4
11.2
8.2
6.7
3.1
7.1
2.7
12.6
1.5
2.1
0.58
68.9
6.6
2.0
2.8
5.3
4.3
19.7
5.3
140.5
4.7
59.2
1.1
5.1
2.1
8.7
2.0
11.8
Pasadena,
CA
5.4
4.7
9.5
4.3
5.6
2.5
4.7
2.5
9.6
1.5
2.3
0.68
53.3
5.3
2.4
6.0
7.0
4.9
22.2
6.1
127.4
5.2
50.0
1.1
6.1
2.3
9.9
2.5
16.5
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
42.3
41.2
29.9
25.0
12.3
33.8
7.39
16.1
2.61
5.02
215.6
0.711
0.905
6.17
9.42
33.7
166
13.6
60.0
10.7
41.2
20.8
160
32.1
205
Bakersfleld,
CA
12.7
14.2
10.7
10.8
5.24
23.6
4.27
9.66
3.50
3.31
98.0
0.164
0.803
1.78
4.01
5.63
54.4
3.77
24.1
2.58
10.4
6.46
43.1
9.71
78.0
Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CO NY Boston, MA CA
19.23 (57.7)
6.04(21.1)
7.77(21.3)
2.08(12.7)
5.62(15.1)
1.26(9.0)
7.70 (20.6)
0.76 (4.6)
5.24(17.9)
0.41 (2.1)
1.49(5.5)
57.9
-------
g.
to
o
o
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
TABLE 3A-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
n-Alkanoic Acids
(cont'd)
n-pentacosanoic acid
n-hexacosanoic acid
n-heptacosanoic acid
n-octacosanoic acid
n-nonacosanoic acid
n-triacontanoic acid
Total n-alcanoic acids
n-Alkenoic Acids
n-9-hexadecenoic acid
n-9-octadecenoic acid
n-9,12-octadecane-
Los Angeles,
CA
1.3
5.6
0.49
2.7
0.33
1.0
292.6
24.8
Pasadena,
CA
1.6
9.3
0.81
4.9
0.57
2.2
294.3
26.0
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
15.4
174
2.56
21.3
1.46
4.32
979.3
18.8
27.1
13.6
Bakersfleld, Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CA CO NY Boston, MA CA
6.59
81.3
2.38
9.65
2.11
5.79
352.7
3.96
3.96
1.83
dienoic acid
Total n-alkenoic acids
24.8
26.0
59.5
9.75
n-Alkanals
1-octanal
n-nonanal
n-decanal
n-dodecanal
n-tridecanal
n-tetradecanal
n-pentadecanal
n-hexadecanal
n-heptadecanal
n-octadecanal
Total n-alkanals
3.26
5.7 9.5 19.4 3.01 29.01
23.58
6.01
6.50
9.62
12.47
17.45
24.09
1.84
5.7 9.5 19.4 3.01 133.8
(14.4)
(62.8)
(71.2)
(16.4)
(25.8)
(30.7)
(113.6)
(49.3)
(88.9)
(11.7)
-------
g.
to
o
o
TABLE 3A-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
oo
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
Rogge et al. (1993)a Schauer and Cass (2000)
Jan-Dec 1 982 Dec 26-28, 1995
(annual average) (pollution episode)
PM2, PM2.5
Los Angeles, Pasadena, Fresno, Bakersfleld,
CA CA CA CA
n-Alkanols
1-decanol
1-dodecanol
1-tetradecanol
1-pentadecanol
1-hexadecanol
Total n-alkanols
Aliphatic Dicarboxylic
Acids
oxalic acid (C2)
malonic acid 32.7 44.4
(propanedioic)
methylmalonic acid 2.13 nd
(methylpropanedioic)
malonic acid 0.66 1.3
(2-butenedioic)
succinic acid 66.5 51.2
(butanedioic)
methylsuccinic acid 18.0 15.0 24.0 8.80
(methylbutanedioic)
glutaricacid 32.3 28.3 21.3 10.5
(pentanedioic)
methylglutaric acid 19.3 16.6
(methylpentanedioic)
hydroxy butanedioic 14.3 16.0
acid
adipicacid 14.1 14.1 3.39 3.07
(hexanedioic)
pimelicacid 2.22 1.03
(heptanedioic)
suberic acid 3.4 4.1 4.41 13.4
(octanedioic)
axelaicacid 29.0 22.8 19.9 8.22
(nonanedioic)
Total aliphatic 230.3 213.8 77.4 45.0
(ttcarboxvlic acids
Khwaja (1995) Allen et al. (1997)
Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
July24-Aug4, 1989 (semiurban) (urban) Sept 8-9, 1993
no precut no precut PM^ 9 (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 3A-2
PARTICIPATE ORGANIC COMPOUND CONCENTRATIONS
C/m3) BASED ON STUDIES
g.
to
o
o
PUBLISHED AFTER 1990 AT SELECTED
Ketocarboxylic Acids
pyruvic acid (C3)
glyoxylic acid (C2)
Total ketocarboxylic
acids
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
Los Angeles, Pasadena,
CA CA
Schauer and Cass (2000)
Dec 26-28, 1995
(pollution episode)
PM2.5
Fresno, Bakersfleld,
CA CA
Veltkamp et al. (1996)
July24-Aug4, 1989
no precut
Niwot Ridge,
CO
SITES
Khwaja (1995)
October 1991
(semiurban)
no precut
Schenectady,
NY
59(103)
44 (68)
103
Allen et al. (1997)
Summer 1994
(urban)
Kenmore Square,
Boston, MA
Fraser et al. (1998)
Sept 8-9, 1993
(urban)
Los Angeles Basin,
CA
Diterpenoid/Resin Acids
dehydroabietic acid 23.6 22.6 98.5 8.01
abietic acid 30.4 0.784
13-isopropyl-5a- 0.63 1.2
podocarpa-6,8,11,13-
tetraen-16-oic acid
8,15-pimaradien-18-oic 0.44 0.57 0.48 0.03
acid
o
£
H
i
O
o
2|
0
H
O
O
H
W
O
O
HH
H
W
pimaric acid
isopimaric acid
7-oxodehydroabietic acid
abieta-6,8,11,13,15-
pentaen-1 8-oic acid
abieta-8,1 1,13,15-tetraen-
1 8-oic acid
sandaracopimaric acid
Total diterpenoid acids
Aromatic Polycarboxylic
Acids
1 ,2-benzene-dicarboxylic
acid (phthalic acid)
1 ,3-benzene-dicarboxylic
acid
2.3
1.3
3.4
1.6
33.3
60.0
3.4
4.8
2.3
4.1
2.2
37.6
55.7
2.9
9.97
127
6.68
11.8
2.62
8.91
296.4
9.16
3.41
0.735
7.95
1.43
2.43
0.251
0.525
22.15
6.78
1.98
-------
g.
to
o
o
TABLE 3A-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM,,
Schauer and Cass (2000)
Dec 26-28, 1995
(pollution episode)
Veltkamp et al. (1996)
July24-Aug4, 1989
no precut
Khwaja (1995)
October 1991
(semiurban)
no precut
Allen et al. (1997)
Summer 1994
(urban)
PM,g
Fraser et al. (1998)
Sept 8-9, 1993
(urban)
Los Angeles, Pasadena, Fresno, Bakersfleld, Niwot Ridge, Schenectady,
CA CA CA CA CO NY
Kenmore Square, Los Angeles Basin,
Boston, MA CA
>
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
Aromatic Polycarboxylic
Acids (cont'd)
1,4-benzene-dicarboxylic
acid
benzene tricarboxylic
acids
1.5
5.16
14.4
4.48
3.77
4-methy 1-1,2-
benzenedicarboxylic acid
1,2,4-benzene-
tricarboxylic acid
(trimellitic acid)
1,3,5-benzene-
tricarboxylic acid
(trimesic acid)
1,2,4,5-benzene-
tetracarboxylic acid
(pyromellitic acid)
Total aromatic
polycarboxylic acids
Poly cyclic Aromatic
Hydrocarbons
retene
fluoranthene
acephenanthrylene
pyrene
Cr202MWPAH
C2-202 MW PAH
benz [a] anthracene
cyclopenta[c
-------
g.
to
o
o
TABLE 3A-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
>
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
Los Angeles, Pasadena,
CA CA
Poly cyclic Aromatic
Hydrocarbons (cont'd)
Cr228MWPAH
C2-228 MW PAH
benz [e] acephen-
anthrylene
benzo[fc]fluoranthene 1.15 1.20
benzo[&]fluoranthene 1.23 0.85
benzo [/] fluoranthene
benzo[e]pyrene 0.97 0.93
benzo[a]pyrene 0.42 0.44
perylene
methyl-substituted 252
MWPAH
mdeno[l,2,3-cd]-pyrene 0.37 0.42
mdeno[l,2,3-cd]- 1.05 1.09
fluoranthene
benzo[g/n']perylene 4.47 4.43
anthanthrene
coronene
Total poly cyclic aromatic 11.66 11.10
hydrocarbons
Oxygenated PAHs/
Poly cyclic Aromatic
Ketones/Quinones
1 ,4-naphthoquinone
1 -acenaphthenone
9-fluorenone
1,8-naphthalic anhydride
phenanthrenequinone
phenalen-9-one
anthracene-9, 1 0-dione
methy lanthracene- 9,10-
dione
1 lH-benzo[a]fluoren-l 1-
one
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
17.6
8.69
10.7
3.62
7.20
8.23
1.50
6.84
1.36
9.75
0.180
139.57
Bakersfleld, Niwot Ridge,
CA CO
5.35
2.13
2.48
0.499
1.98
1.77
0.246
2.56
0.764
3.49
0.131
34.40
Schenectady, Kenmore Square, Los Angeles Basin,
NY Boston, MA CA
0.34(2.16)
0.09 (0.46)
0.20(1.00)
0.22(1.07)
0.02(0.10)
0.22(1.00)
0.14(0.80)
0.05(0.51)
0.10(0.88)
0.29(1.38)
0.10(0.46)
0.77(4.23)
3.77
0.26
2.07 0.29(1.04)
1.77 0.41(1.65)
0.43
0.53(2.23)
0.36(1.14)
0.09 (0.24)
1.03
-------
g.
to
o
o
to
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
TABLE 3A-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
Los Angeles, Pasadena,
CA CA
Schauer and Cass (2000)
Dec 26-28, 1995
(pollution episode)
PM2.5
Fresno, Bakersfleld,
CA CA
Veltkamp et al. (1996)
July24-Aug4, 1989
no precut
Niwot Ridge,
CO
Khwaja (1995)
October 1991
(semiurban)
no precut
Schenectady,
NY
Allen et al. (1997)
Summer 1994
(urban)
PML9
Kenmore Square,
Boston, MA
Fraser et al. (1998)
Sept 8-9, 1993
(urban)
Los Angeles Basin,
CA
Oxygenated PAHs/
Poly cyclic Aromatic
Ketones/Quinones
(cont'd)
7H-benzo [c] fluoren-7-one
1 lH-benzo[6]fluoren-l 1-
one
1 H-phenalen-1 -one
benzanthrone
5,12-naphthacene-quinone
7H-benz[rfe]-anthracen-7- 0.81
one
benz [de] anthracene-7-
dione
benz[a]anthracene-7,12- 0.21
dione
cyclopenta[rfe/|phen-
anthrone
benzo[«/]pyren-6-one 0.80
6H-benzo[crf]pyrene-6-
one
benzo [a]pyrene-6,12-
dione
Total poly cyclic aromatic 1.82
ketones/quinones
7.96
0.588
0.37
0.85
1.18
0.32
0.84
0.25
1.24
2.33
15.76
2.07
1.34
0.096
9.72
0.20(1.00)
0.09(0.31)
0.05(0.14)
0.54 (2.47)
2.56
Steroids
cholesterol nd
Substituted Phenols
/>-benzenediol
OT-benzenediol
hydroxybenzaldehydes
Total substituted phenols
1.9
3.46
7.59
2.64
13.69
nd
nd
0.604
0.604
-------
g.
to
o
o
TABLE 3A-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
>
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
Los Angeles, Pasadena,
CA CA
Guaiacol and Substituted
Guaiacols
guaiacol
4-methylguaiacol
trans-isoeugenol
vanillin
acetovanillone
guaiacyl acetone
coniferyl aldehyde
Total guaiacol and
substituted guaiacols
Syringol and Substituted
Syringols
syringol
4-methylsyringol
4-ethylsyringol
4-propylsyringol
4-propenylsyringol
syringaldehyde
acetosyringone
acetonylsyringol
propionylsyringol
butyrylsyringol
sinapyl aldehyde
Total syringol and
substituted syringols
Sugars
levoglucosan
other sugars
Total sugars
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
0.889
0.606
1.45
26.8
3.23
10.8
47.0
90.78
1.16
1.72
2.28
0.871
4.38
135
171
406
32.1
15.3
15.9
785.7
7590
1070
8660
Bakersfleld, Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CA CO NY Boston, MA CA
0.832
0.387
1.04
6.05
0.705
4.29
nd
13.30
0.845
1.77
2.39
nd
1.40
44.5
55.7
68.1
16.2
6.18
197.1
1100
171
1271
-------
g.
to
o
o
TABLE 3A-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
>
£
fe
H
6
o
o
H
O
O
H
W
O
O
HH
H
W
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
Other Compounds
divanillyl
divanillyl methane
vanillylmethylguaiacol
Total other
N-Containing Compounds
3 -methoxypyridine
isoquinoline
1 -methoxypyridine
1 ,2-dimethoxy-4-nitro-
benzene
dihydroxynitrobenzene
Total N-containing
compounds
Total Quantified Organic
Compound Mass
Total Organic Compound
Mass
Percent of Organic Mass
Quantified
Percent of Organic Mass
Extractable and Elutable
Los Angeles,
CA
0.86
1.1
0.27
1.8
4.03
789
8-15% (a)
45-60% (a)
Pasadena,
CA
1.4
1.1
0.24
3.9
6.64
764
8-15% (a)
45-60% (a)
Schauer and Cass (2000) Veltkamp et al. Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
19.4
2.39
3.24
25.0
11410
55700
20%
30%
Bakersfield, Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CA CO NY Boston, MA CA
3.18
nd
0.568
3.75
1.62(10.52)
1.62
2075 267 487 10 8
18700
11% <3%
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. 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.
March 2001 3 A-15 DRAFT-DO NOT QUOTE OR CITE
-------
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.
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.
U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-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.
March 2001 3 A-16 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 3B
2
3 Composition of Particulate Matter Source Emissions
4
5
6 This appendix includes discussions of the elemental composition of emissions from various
7 source categories discussed in Table 3-7. Discussions in this appendix incorporate material
8 dealing with the inorganic components of source emissions from Chapter 5 of the 1996 PM
9 AQCD (U. S. Environmental Protection Agency, 1996), updates to that material, and material
10 describing the composition of organic components in source emissions. The primary emphasis in
11 the discussions is on the composition of PM2 5 particle sources.
12
13 Soil and Fugitive Dust
14 The compositions of soils and average crustal material are shown in Table 3B-1 (adapted
15 from Warneck, 1988). Two entries are shown as representations of average crustal material.
16 Differences from the mean soil composition shown can result from local geology and climate
17 conditions. Major elements in both soil and crustal profiles are Si, Al, and Fe, which are found in
18 the form of various minerals. In addition, organic matter constitutes a few percent, on average, of
19 soils. In general, the soil profile is similar to the crustal profiles, except for the depletion of
20 soluble elements such as Ca, Mg, Na, and K. It should be noted that the composition of soils from
21 specific locations can vary considerably from these global averages, especially for elements like
22 Ca, Mg, Na, and K.
23 Fugitive dust emissions arise from paved and unpaved roads, building construction and
24 demolition, parking lots, mining operations, storage piles, and agricultural tilling in addition to
25 wind erosion. Figure 3B-1 shows examples of size distributions in dust from paved and unpaved
26 roads, agricultural soil, sand and gravel, and alkaline lake bed sediments, which were measured in
27 a laboratory resuspension chamber as part of a study in California (Chow et al., 1994). This figure
28 shows substantial variation in particle size among some of these fugitive dust sources. The PMX 0
29 abundance (6.9%) in the total suspended PM (TSP) from alkaline lake bed dust is twice its
30 abundance in paved and unpaved road dust. Approximately 10% of the TSP is in the PM25
31 fraction and approximately 50% of TSP is in the PM10 fraction. The sand/gravel dust sample
32 shows that 65% of the mass is in particles larger than the PM10 fraction. The PM2 5 fraction of
March 2001 3B-1 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-1. AVERAGE ABUNDANCES OF MAJOR ELEMENTS IN
SOIL AND CRUSTAL ROCK
Element
Si
Al
Fe
Ca
Mg
Na
K
Ti
Mn
Cr
V
Co
Elemental
Soil
(1)
330,000
71,300
38,000
13,700
6,300
6,300
13,600
4,600
850
200
100
8
Abundances (ppmw)
(2)
277,200
81,300
50,000
36,300
20,900
28,300
25,900
4,400
950
100
135
25
Crustal Rock
(3)
311,000
77,400
34,300
25,700
33,000
31,900
29,500
4,400
670
48
98
12
Source: (1) Vinogradov (1959); (2) Mason (1966); (3) Turekian (1971), Model A; as quoted in Warneck (1988).
1 TSP is approximately 30 to 40% higher in alkaline lake beds and sand/gravel than in the other soil
2 types. The tests were performed after sieving and with a short (<1 min) waiting period prior to
3 sampling. It is expected that the fraction of PMl 0 and PM25 would increase with distance from a
4 fugitive dust emitter as the larger particles deposit to the surface faster than do the smaller
5 particles.
6 The size distribution of samples of paved road dust obtained from a source characterization
7 study in California is shown in Figure 3B-2. As might be expected, most of the emissions are in
8 the coarse size mode. The chemical composition of paved road dust obtained in Denver, CO,
9 during the winter of 1987-1988 is shown in Figure 3B-3. The chemical composition of paved
10 road dust consists of a complex mixture of particulate matter from a wide variety of sources.
11 Hopke et al. (1980) found that the inorganic composition of urban roadway dust in samples from
March 2001 3B-2 DRAFT-DO NOT QUOTE OR CITE
-------
100
80
60
40
20
Paved Unpaved
Road Dust Road Dust
Agricultural Soil/Gravel Alkaline
Soil Lake Bed
Figure 3B-1. Size distribution of particles generated in a laboratory resuspension
chamber.
Source: Chow etal. (1994).
1 Urbana, IL, could be described in terms of contributions from natural soil, automobile exhaust,
2 rust, tire wear, and salt. Automobile contributions arose from exhaust emissions enriched in Pb;
3 from rust as Fe; tire wear particles enriched in Zn; brake linings enriched in Cr, Ba, and Mn; and
4 cement particles derived from roadways by abrasion. In addition to organic compounds from
5 combustion and secondary sources, road dust also contains biological material such as pollen and
6 fungal spores.
7 Very limited data exist for characterizing the composition in organic compounds
8 resuspended paved road dust and soil dust. The only reported measurements are from Rogge et al.
9 (1993a) and Schauer and Cass (2000), which consist of data for the fine particle fraction. The
10 resuspended road dust sample analyzed Rogge et al. (1993a) was collected in Pasadena, California
during May of 1988. The sample analyzed by Schauer and Cass (2000) is a composite sample
March 2001
3B-3
DRAFT-DO NOT QUOTE OR CITE
-------
100
80
60
0.
CO
s.
40
20
92.8%
82.7%
(<2.5M)
81.6%
52.3%
10.7%
(<2.5u)
4.5%
95.8%
93.1%
(<2.5|J
92.4%
96.2%
(<10M)
92.3%
(<2.5M)
91.8%
99.2%
97.4%
(<2.5u)
87.4%
34.9%
Road and Agricultural Residential Diesel
Soil Dust Burning Wood Truck
Combustion Exhaust
Crude Oil Construction
Combustion Dust
Code:
>10u
2.5u-10u
1u-2.5u
Figure 3B-2. Size distribution of California source emissions, 1986.
Source: Houcketal. (1989, 1990).
March 2001 3B-4 DRAFT-DO NOT QUOTE OR CITE
-------
10'
10
10'
.£>.
'^^&^^«$f<&<$i&tff4f&fS
" O O V0y C^ *s?^ Cj X^1* "C* ^ j*S^> ^^Cj t>^ £& sifc^ ».
-------
TABLE 3B-2. SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
PRESENT IN FINE PARTICLE ROAD DUST SAMPLE
Source
Pasadena Road Dust
(Rogge et al., 1993a)
San Joaquin Valley
Road Dust (Schauer
and Cass, 2000)
Contribution to Dominant Contributors to
Compound Class Paniculate Mass (%) Emissions of Compound Class
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Petroleum biomarkers
PAH
n-Alkanals
n-Alkanols
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
0.13
0.37
0.028
0.017
0.0059
0.046
0.021
0.023
0.23
0.095
CIY, C19, C21
Palmitic acid and stearic acid
Oleic acid and linoleic Acid
Hopanes and steranes
No dominant compounds
Octacosanol and triacontanal
Hexacosanol and octacosanol
No dominant compounds
Palmitic acid and stearic acid
Oleic acid, linoleic acid, and
hexadecenoic acid
1 Stationary Sources
2 The elemental composition of primary particulate matter emitted in the fine fraction from a
3 variety of power plants and industries in the Philadelphia area is shown in Table 3B-3 as a
4 representative example of emissions from stationary fossil combustion sources (Olmez et al.,
5 1988). Entries for the coal fired power plant show that Si and Al followed by sulfate are the
6 major primary constituents produced by coal combustion, whereas fractional abundances of
7 elemental carbon were much lower and organic carbon species were not detected. Sulfate is the
8 major parti culate constituent released by the oil fired power plants examined in this study, and,
9 again, elemental and organic carbon are not among the major species emitted. Olmez et al. (1988)
10 also compared their results to a number of similar studies and concluded that their data could have
11 much wider applicability to receptor model studies in other areas with some of the same source
12 types. The high temperature of combustion in power plants results in the almost complete
13 oxidation of the carbon in the fuel to CO2 and very small amounts of CO. Combustion conditions
14 in smaller boilers and furnaces allow the emission of unburned carbon and sulfur in
March 2001
3B-6
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-3. COMPOSITION OF FINE PARTICLES RELEASED BY VARIOUS STATIONARY
g.
to
o
o
OJ
td
O
?•
H
1
0
o
o
H
0
o
1
w
k-LJ
0
/-\
SOURCES IN THE PHILADELPHIA AREA
Species Eddystone Coal-
(Units) Fired Power Plant
C-v (%)
C-e (%)
NH4 (%)
Na (%)
Al (%)
Si (%)
P (%)
S (%)
c/~\ /o/ \
O\-/4 ^ /oj
Cl (%)
K(%)
Ca (%)
Sc (ppm)
Ti (%)
V (ppm)
Cr (ppm)
Mn (ppm)
Fe (%)
Co (ppm)
Ni (ppm)
ND
0.89 ±0.12
1.89 ±0.19
0.31 ±0.03
14 ±2
21.8±1.6
0.62 ±0.10
3.4 ±0.6
11.9±1.2
0.022 ±0.11
1.20 ±0.09
1.4 ±0.5
42 ±2
1.1 ±0.2
550 ± 170
390 ± 120
290 ± 15
7.6 ±0.4
93 ±10
380 ±50
Oil-Fired Power Plants
; N
3
3
3
3
9
9
9
3
3
9
3
3
3
3
3
3
3
3
9
Eddystone
2.7 ±1.2
7.7 ±1.5
3.5 ±1.6
3.0 ±0.8
0.45 ±0.09
1.9 ±0.6
1.5 ±0.4
11±2
40 ±4
0.019 ±0.009
0.16 ±0.05
3.6 ±1.0
0.17 ±0.02
0.040 ± 0.044
11500 ±3000
235 ±10
380 ± 40
1.6 ±0.2
790 ± 150
15000 ± 5000
N
3
3
3
3
3
9
9
9
3
2
9
3
3
9
3
3
3
3
3
9
Schuylkill
0.75 ±0.63
0.22 ±0.17
3.7 ± 1.7
3. 3 ±0.8
0.94 ±0.08
2.6 ±0.4
1.0 ±0.2
13 ±1
45 ±7
ND
0.21 ±0.03
2.3 ±1.0
0.47 ±0.02
0.12 ±0.02
20,000 ± 3000
230 ± 70
210 ±50
1.7 ±0.4
1100 ±200
19000 ± 2000
N
4
4
4
3
3
11
11
11
4
11
3
3
11
3
3
3
3
3
11
Secondary
Al Plant
1.6 ± 1.5
0.18±0.10
2.2 ±0.9
16.3 ±0.8
1.74 ±0.09
3.1 ±2.2
0.45 ± 0.27
3±4
5.9 ±2
21±4
10.9 ±1.5
0.12 ±0.09
0.092 ±0.039
0.024 ± 0.003
36 ±7
410 ±20
120 ±15
0.31 ±0.02
13 ±2
300 ± 100
N
2
2
2
1
1
2
2
2
2
1
2
2
1
2
1
1
1
1
1
2
Fluid Cat.
Cracker
ND
0.16 ±0.05
0.43 ± 0.22
0.38 ±0.05
6.8 ± 1.2
9.8 ±20.0
ND
4.2 ± 12.6
38 ±4
ND
0.031 ±0.005
0.030 ±0.004
2.7 ±0.4
0.38 ±0.1
250 ± 70
59 ±8
14 ±3
0.20 ±0.03
15 ±2
220 ± 30
N
3
3
3
3
9
9
3
9
9
3
3
3
3
3
9
3
9
Municipal
Incinerator
0.57 ± 0.26
3.5 ±0.2
0.36 ±0.07
6.6 ±3. 5
0.25 ±0.10
1.7 ±0.3
0.63 ±0.12
2.9 ±0.8
6.8 ±2.3
29 ±5
7.6 ±2.3
0.23 ±0.10
0.11 ±0.02
0.030 ±0.015
8.6 ±5. 3
99 ±31
165 ± 40
0.22 ±0.05
3.7 ±0.8
290 ± 40
N
4
4
4
3
3
10
10
10
4
3
10
10
1
10
2
3
3
3
3
10
-------
TABLE 3B-3 (cont'd). COMPOSITION OF FINE PARTICLES RELEASED BY VARIOUS STATIONARY
g.
to
o
OJ
td
oo
o
§
2
o
o
H
0
O
0
o
SOURCES IN THE PHILADELPHIA AREA
Species
(units)
Cu (ppm)
Zn (%)
As (ppm)
Se (ppm)
Br (ppm)
Rb (ppm)
Sr (ppm)
Zr (ppm)
Mo (ppm)
Ag (ppm)
Cd (ppm)
In (ppm)
Sn (ppm)
Sb (ppm)
Cs (ppm)
Ba (ppm)
La (ppm)
Ce (ppm)
Nd (ppm)
Sm (ppm)
Eddy stone
Coal-Fired
Power Plant
290 ± 20
0.041 ±0.005
640 ± 80
250 ± 20
35 ±8
190 ± 80
1290 ± 60
490 ± 190
170 ± 60
ND
ND
0.71 ±0.04
ND
a
9.2 ±0.9
ND
120 ± 10
180 ±10
80 ±26
23 ±2
Oil-Fired Power Plants
N
9
3
3
3
3
1
9
9
2
2
2
3
2
3
3
Eddystone
980 ± 320
1.3 ±0.3
33 ±6
26 ±9
90 ±60
ND
160 ± 50
140 ± 180
930 ±210
ND
ND
ND
320 ±230
370 ±410
ND
1960 ± 100
130 ±30
89 ±23
28 ±5
3.7 ±0.7
N
9
3
1
3
9
9
9
3
9
3
3
3
3
2
3
Schuylkill
1100 ±500
0.78 ±0.30
50 ±16
23 ±7
45 ±17
ND
280 ± 70
100 ± 120
1500 ±300
ND
ND
ND
200 ± 80
1020 ± 90
ND
2000 ± 500
450 ±30
360 ± 20
230 ± 20
20.5 ± 1.5
N
11
3
3
3
11
11
11
3
11
3
3
3
3
3
3
Secondary
Al Plant
450 ± 200
0.079 ± 0.006
15 ±6
66 ±3
630 ± 70
97 ±38
ND
ND
ND
ND
ND
ND
550 ± 540
6100 ±300
ND
ND
19 ±2
ND
ND
ND
N Fluid Cat. Cracker
2 14 ±8
1 0.0026 ± 0.0007
1 ND
1 15± 1
2 5.6 ±1.8
1 ND
36 ±6
130 ±50
ND
ND
ND
ND
2 ND
1 7.7 ±1.5
ND
290 ± 90
1 3300 ±500
2700 ± 400
1800 ± 250
170 ± 20
N
9
3
3
9
9
2
3
2
3
3
3
3
Municipal
Incinerator
1300 ±500
10.4 ±0.5
64 ±34
42 ± 16
2300 ± 800
230 ±50
87 ±14
ND
240 ± 130
71 ±15
1200 ± 700
4.9 ±1.4
6700 ± 1900
1300 ± 1000
5.9 ±3.0
ND
1.1±0.5
ND
ND
ND
N
3
3
3
3
10
2
10
10
3
3
3
10
3
3
1
-------
TABLE 3B-3 (cont'd). COMPOSITION OF FINE PARTICLES RELEASED BY VARIOUS
3
to
2 Species
(units)
Eu (ppm)
Gd (ppm)
Tb (ppm)
Yb (ppm)
Lu (ppm)
Hf(ppm)
Ta (ppm)
W (ppm)
Au (ppm)
w Pb (%)
td
Th (ppm)
% mass
STATIONARY SOURCES IN THE PHILADELPHIA AREA
Eddystone Coal-
Fired Power
Plant
5.1 ±0.5
ND
3. 3 ±0.3
10.3 ±0.5
ND
5.8 ±0.8
ND
20 ±8
ND
0.041 ±0.004
24 ±2
24 ±2
Oil-Fired Power Plants
N
3
3
1
3
1
9
3
6
Eddystone N
ND
ND
ND
ND
ND
0.39 ±0.07 1
ND
60 ±5 2
0.054 ±0.017 2
1.8 ±0.6 9
1.9 ±0.5 2
93. 5 ±2.5 6
Schuylkill
0.65 ±0.23
ND
0.90 ±0.29
ND
ND
ND
ND
ND
ND
1.0 ±0.2
ND
96 ±2
Secondary
N Al Plant
3 ND
ND
3 ND
ND
ND
ND
ND
ND
ND
11 0.081 ±0.014
ND
6 81 ±10
Fluid Cat.
N Cracker
4.9 ±0.7
71 ±10
8.9 ±1.3
3.7 ±0.4
0.59 ±0.17
0.99 ±0.08
0.56 ±0.10
ND
ND
2 0.0091 ±0.0021
6.2 ±0.7
2 97 ±2
N
3
3
3
3
3
3
3
9
3
7
Municipal
Incinerator
ND
ND
ND
ND
ND
ND
ND
ND
0.56 ±0.27
5.8 ±1.2
ND
89 ±2
N
3
10
7
H
6
o
*
o
H
O
aOmitted because of sample contamination.
N = Number of samples.
ND = Not detected.
The "% mass" entries give the average percentage of the total emitted mass found in the fine fraction.
Source: Adapted from Olmez et al. (1988).
o
&
o
-------
1 more reduced forms such as thiophenes and inorganic sulfides. A number of trace elements are
2 greatly enriched over crustal abundances in different fuels, such as Se in coal and V, Zn, and Ni in
3 oil. In fact, the higher V content of the fuel oil than in coal could help account for the higher
4 sulfate seen in the profiles from the oil-fired power plant compared to the coal-fired power plant
5 because V at combustion temperatures found in power plants is known to catalyze the oxidation
6 of reduced sulfur species. During combustion at lower temperatures, the emission of reduced
7 sulfur species also occurs. For example, Huffman et al. (2000) identified sulfur species emitted
8 by the combustion of several residual fuels oils (RFO) in a fire tube package boiler, which is
9 meant to simulate conditions in small institutional and industrial boilers. They found that sulfur
10 was emitted not only as sulfate (26 to 84%), but as thiophenes (13 to 39%) with smaller amounts
11 of sulfides and elemental S. They also found that Ni, V, Fe, Cu, Zn, and Pb are present mainly as
12 sulfates in emissions. Linak et al. (2000) found, when burning RFO, that the fire tube package
13 boiler produced particles with a bimodal size distribution in which about 0.2% of the mass was
14 associated with particles smaller than 0.1-//m AD, with the rest of the mass lying between 0.5 and
15 100 //m. Miller et al. (1998) found that larger particles consisted mainly of cenospheric carbon,
16 whereas trace metals and sulfates were found concentrated in the smaller particles in a fire tube
17 package boiler. In contrast, when RFO was burning in a refractory-lined combustor, which is
18 meant to simulate combustion conditions in a large utility residual oil fired boiler, Linak et al.
19 (2000) found that particles were distributed essentially unimodally, with a mean diameter of about
20 0.1//m.
21 Apart from emissions in the combustion of fossil fuels, trace elements are emitted as the
22 result of various industrial processes such as steel and iron manufacturing and nonferrous metal
23 production (e.g., for Pb, Cu, Ni, Zn, and Cd). As may be expected, emissions factors for the
24 various trace elements are highly source-specific (Nriagu and Pacyna, 1988). Inspection of
25 Table 3B-3 reveals that the emissions from the catalytic cracker and the oil-fired power plant are
26 greatly enriched in rare-earth elements such as La compared to other sources.
27 Emissions from municipal waste incinerators are heavily enriched in Cl arising mainly from
28 the combustion of plastics and metals that form volatile chlorides. The metals can originate from
29 cans or other metallic objects and some metals such as Zn and Cd are also additives in plastics or
30 rubber. Many elements such as S, Cl, Zn, Br, Ag, Cd, Sn, In, and Sb are enormously enriched
31 compared to their crustal abundances. A comparison of the trace elemental composition of
March 2001 3B-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 incinerator emissions in Philadelphia, PA (shown in Table 3B-3), with the composition of
2 incinerator emissions in Washington DC, and Chicago, IL (Olmez et al., 1988), shows agreement
3 for most constituents to better than a factor of two.
4 Very limited data exist for characterizing the chemical composition of organic compounds
5 present in particulate emissions from industrial-scale stationary fuel combustion. Oros and
6 Simoneit (2000) have presented the abundance and distribution of organic constituents in coal
7 smokes that have been burned under laboratory conditions. This work provides the basis for
8 further investigation addressing the emissions of coal fired boilers.
9 Rogge et al. (1997a) measured the composition of the organic constituents in the particulate
10 matter emissions from a 50 billion kj/h boiler that was operating at 60% capacity and was burning
11 number 2 distillate fuel oil. The fine carbon particulate matter emissions from this boiler over
12 five tests were composed of an average of 14% organic carbon and 86% elemental carbon
13 (Hildemann et al., 1991). Significant variability in the distribution of organic compounds present
14 in the emissions from two separate tests was observed. Most of the identified organic mass
15 consisted of n-alkanonic acids, aromatic acids, n-alkanes, PAH, oxygeanted PAH, and chlorinated
16 compounds. It is unclear if these emissions are representative of typical fuel oil combustion units
17 in the United States. Rogge et al. (1997b) measured the composition of hot asphalt roofing tar
18 pots, and Rogge et al. (1993b) measured the composition of emissions from home appliances that
19 use natural gas.
20
21 Motor Vehicles
22 Exhaust emissions of particulate matter from gasoline powered motor vehicles and diesel
23 powered vehicles have changed significantly over the past 25 years (Sawyer and Johnson, 1995;
24 Cadle et al., 1999). These changes have resulted from reformulation of fuels, the wide application
25 of exhaust gas treatment in gasoline-powered motor vehicles, and changes in engine design and
26 operation. Because of these evolving tailpipe emissions, along with the wide variability of
27 emissions between vehicles of the same class (Hildemann et al., 1991; Cadle et al., 1997; Sagebiel
28 et al., 1997; Yanowitz et al., 2000), well-defined average emissions profiles for the major classes
29 of motor vehicles have not been established. Two sampling strategies have been employed to
30 obtain motor vehicle emissions profiles: (1) the measurement of exhaust emissions from vehicles
31 operating on dynamometers and (2) the measurement of integrated emissions of motor vehicles
March 2001 3B-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 driving through roadway tunnels. Dynamometer testing can be used to measure vehicle emissions
2 operating over an integrated driving cycle and allows the measurement of emissions from
3 individual vehicles. However, dynamometer testing requires considerable resources and usually
4 precludes testing a very large number of vehicles. In contrast, a large number of vehicles can be
5 readily sampled in tunnels, but vehicles driving through tunnels operate over limited driving
6 conditions and the measurements represent contributions from a large number of vehicle types.
7 As a result, except in a few cases, tunnel tests have not been effective at developing chemically
8 speciated particulate matter emissions profiles for individual motor vehicle classes. As a result,
9 several studies have measured the contribution of both organic and elemental carbon to the
10 particulate matter emissions from different classes of motor vehicles operating on chassis
11 dynamometers.
12 The principal components emitted by diesel and gasoline fueled vehicles are organic carbon
13 (OC) and elemental carbon (EC) as shown in Tables 3B-4a and 4b. As can be seen, the variability
14 among entries for an individual fuel type is large and overlaps that found between different fuel
15 types. On average, the abundance of elemental carbon is larger than that of organic carbon in the
16 exhaust of diesel vehicles, whereas organic carbon is the dominant species in the exhaust of
17 gasoline fueled vehicles. Per vehicle, total carbon emissions from light and heavy duty diesel
18 vehicles can range from 1 to 2 orders of magnitude higher than those from gasoline vehicles.
19 There appears to be a tendency for emissions of elemental carbon to increase relative to emissions
20 of organic carbon for gasoline fueled vehicles as simulated driving conditions are changed from a
21 steady 55 km/h to the various load conditions specified in the Federal Test Procedures (FTPs).
22 Also shown are the results of sampling from mixed vehicle types along roadsides and in tunnels.
23 As might be expected, most of the PM emitted by motor vehicles is in the PM2 5 size range.
24 Particles in diesel exhaust are typically trimodal consisting of a nuclei mode, an accumulation
25 mode and a coarse mode and are lognormal in form (Kittelson, 1998). More than 90% of the total
26 number of particles are in the nuclei mode, which contains only about 1 to 20% of the particle
27 mass with a mass median diameter of about 0.02 //m, whereas the accumulation mode (with a
28 mass median diameter of about 0.25 //m) contains most of the mass with a smaller fraction (5 to
29 20%) contained in the coarse mode. Kerminin et al. (1997), Bagley et al. (1998), and Kleeman
30 et al. (2000) also have shown that gasoline and diesel fueled vehicles produce particles that are
31 mostly less than 2.0 //m in diameter. Cadle et al. (1999) found that 91% of PM emitted by in-use
March 2001 3B-12 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-4a. ORGANIC AND ELEMENTAL CARBON FRACTIONS OF DIESEL
AND GASOLINE ENGINE PARTICIPATE MATTER EXHAUST
Heavy-duty diesel engines3
Heavy-duty diesel engines (SPECIATE)b
Light-duty diesel engines0
Light-duty diesel engines (SPECIATE)b
Gasoline engines (hot stabilized)8
Gasoline engines ("smoker" and "high emitter")"'0
Gasoline engines (cold start)"
Organic Carbon
19 ± 8%
21-36%
30 ± 9%
22 - 43%
56 ± 11%
76 ± 10%
46 ± 14%
Elemental Carbon
75 ± 10%
52 - 54%
61 ± 16%
51-64%
25 ± 15%
7 ± 6%
42 ± 14%
Tujita et al. (1998) and Watson et al. (1998).
bU.S. EPA SPECIATE database.
°Norbeck et al. (1998).
Source: U.S. Environmental Protection Agency (1999).
1 gasoline vehicles in the Denver area was in the PM2 5 size range, which increased to 97% for
2 "smokers" (i.e., light-duty vehicles with visible smoke emitted from their tailpipes) and 98% for
3 diesels. Durbin et al. (1999) found that about 92% of the PM was smaller than 2.5 //m for
4 smokers and diesels. The mass median diameter of the PM emitted by the gasoline vehicles
5 sampled by Cadle et al. (1999) was about 0.12 //m, which increased to 0.18 //m for smokers and
6 diesels. Corresponding average emissions rates of PM2 5 found by Cadle et al. (1999) for diesels
7 were 552 mg/mi; for smokers they were 222 mg/mi; and, for gasoline vehicles, they were
8 38 mg/mi. The values for smokers and for diesels appear to be somewhat lower than those given
9 in Table 3B-5, whereas the value for gasoline vehicles falls in the range given for low and
10 medium gasoline vehicle emissions.
11 Examples of data for the trace elemental composition of the emissions from a number of
12 vehicle classes obtained as part of the North Frontal Range Air Quality Study (NFRAQS), which
13 took place in December 1997 in Colorado are shown in Table 3B-5. As can be seen from
14 Table 3B-5, emissions of total carbon (TC), which is equal to the sum of organic carbon (OC) and
15 elemental carbon (EC), from gasoline vehicles are highly variable. Gillies and Gertler (2000)
16 point out that there is greater variability in the concentrations of trace elements and ionic
March 2001
3B-13
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-4b. CONTRIBUTION OF ORGANIC CARBON TO PARTICIPATE
MATTER CARBON EMISSIONS IN MOTOR VEHICLE EXHAUST COLLECTED
FROM VEHICLES OPERATED ON CHASSIS DYNAMOMETERS
GASOLINE POWERED VEHICLES
Light-duty vehicles
High-CO/VOC-emitting smokers
High-CO/VOC-emitting nonsmokers
Catalyst-equipped vehicles
Noncatalyst vehicles
DIESEL VEHICLES
Light-duty diesel vehicles
Medium-duty diesel vehicles
Heavy-duty diesel vehicles
Heavy-duty diesel vehicles
Year of Tests
1996-97
1994
1994
Mid-1980s
Mid-1980s
1996-1997
1996
1992
Mid-1980s
Test Cycle
FTP
IM-240
IM-240
FTP
FTP
FTP
FTP
c
c
Number of
Vehicles
195a
7
15
7
6
195a
2
6
2
OC % of
Total Carbon
70
91
76
69
89
40
50b
42
45
Notes
A
B
B
C
C
A
D
E
C
Notes:
A. From(Cadle etal., 1999). Average of summer and winter cold start emissions.
B. From (Sagebiel etal., 1997). Hot start testing of vehicles identified as either high emitters of carbon
monoxide or volatile organic compounds (VOCs).
C. From (Hildemann et al., 1991). Cold start tests.
D. From (Schauer et al., 1999). Hot start tests of medium duty vehicles operating on an FTP cycle.
E. From (Lowenthal et al., 1994). Only includes measurement of vehicles powered by diesel fuel operated
without an exhaust paniculate trap.
aA total of 195 light duty vehicles were tested that include both gasoline powered vehicles and diesel powered
vehicles.
bFraction of paniculate matter consisting of organic carbon was measured with and without an organics denuder
upstream of paniculate filter. Results reported here represent measurement without an organics denuder for
consistency with other measurements. Using an organics denuder, the organic carbon comprised 39% of the
paniculate matter carbon.
Thriving cycle comprised of multiple idle, steady acceleration, constant speed, deceleration steps (see reference
for more details).
1 species than for OC and EC among different source profiles (e.g., SPECIATE, Lawson and Smith
2 (1998), Norbeck et al. (1998)). They suggest that this may arise because their emissions are not
3 related only to the combustion process, but also to their abundances in different fuels and
March 2001 3B-14 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-5. EMISSION RATES (mg/mi) FOR CONSTITUENTS OF PARTICIPATE
MATTER FROM GASOLINE AND DIESEL VEHICLES
Gasoline Vehicles
TC
OC
EC
NO3
SO4=
Na
Mg
Al
Si
P
S
Cl
K
Ca
Fe
Ni
Cu
Zn
Br
Ba
Pb
Low
9.07 ±0.75
6.35 ±0.54
2.72 ±0.52
0.039 ± 0.027
0.158 ±0.036
0.060 ± 0.063
0.036 ±0.022
0.083 ±0.016
0.066 ± 0.008
0.035 ±0.004
0.085 ±0.006
0.024 ±0.012
0.010 ±0.009
0.060 ±0.010
0.143 ±0.004
0.001 ±0.004
0.002 ± 0.004
0.048 ±0.003
0.001 ±0.002
0.013 ±0.136
0.007 ± 0.006
Medium
41. 30 ±1.68
26.02 ±1.31
15.28 ±0.99
0.057 ±0.028
0.518 ±0.043
0.023 ±0.111
0.068 ± 0.027
0.078 ±0.016
0.279 ±0.011
0.152 ±0.007
0.442 ± 0.009
0.038 ±0.012
0.019 ±0.009
0.212 ±0.011
0.756 ±0.005
0.005 ± 0.004
0.016 ±0.003
0.251 ±0.004
0.016 ±0.002
0.009 ±0.138
0.085 ± 0.005
High
207.44 ± 7.29
95.25 ±4.28
112.19 ±5. 82
0.141 ±0.031
0.651 ±0.052
0.052 ±0.092
0.041 ±0.033
0.057 ±0.014
0.714 ±0.012
0.1 13 ±0.007
0.822 ± 0.022
0.081 ±0.020
0.031 ±0.035
0.210 ±0.030
1.047 ±0.010
0.011 ±0.005
0.021 ±0.005
0.265 ± 0.023
0.079 ±0.003
0.011 ±0.299
0.255 ±0.008
Smoker
456.38 ±16.80
350.24 ±15.27
106. 14 ±5.42
0.964 ±0.051
2.160 ±0.137
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
2.515 ±0.116
0.140±0.117
0.033 ±0.386
0.362 ±0.250
2.438 ±0.054
0.008 ±0.017
0.071 ±0.018
0.188 ±0.272
0.047 ±0.012
0.380 ±2.175
0.345 ±0.032
Diesel
Light Duty
373.43 ±13.75
132.01 ±5.82
241.42 ±12.11
1.474 ±0.071
2.902 ±0.165
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
2.458 ±0.124
0.228 ±0.114
0.000 ± 0.426
0.150 ±0.304
0.515 ±0.057
0.014 ±0.018
0.024 ±0.021
0.000 ±0.299
0.003 ±0.014
0.428 ±2.390
0.153 ±0.033
Vehicles
Heavy Duty
1570.69 ±58.24
253. 94 ±16.12
1316.75 ±55. 33
1.833 ±1.285
3.830 ±1.286
1.288 ±2.160
1.061 ±0.729
0.321 ±0.543
8.018 ±0.221
0.407 ±0.136
3.717±0.111
0.881 ±0.221
0.064 ± 0.248
0.716 ±0.107
0.376 ±0.055
0.002 ±0.057
0.001 ±0.062
0.707 ±0.032
0.012 ±0.050
0.493 ±3. 108
0.008 ±0.154
Source: Lawson and Smith (1998).
1 lubricants and also to wear and tear during vehicle operation. Emissions from smokers are
2 comparable to those from diesel vehicles. Thus, older, poorly maintained gasoline vehicles could
3 be significant sources of PM25 (Sagebiel et al., 1997; Lawson and Smith, 1998), in addition to
March 2001
3B-15
DRAFT-DO NOT QUOTE OR CITE
-------
1 being significant sources of gaseous pollutants (e.g., Calvert et al., 1993). Durbin et al. (1999)
2 point out that although smokers constitute only 1.1 to 1.7% of the light-duty fleet in the South
3 Coast Air Quality Management District in California, they contribute roughly 20% of the total PM
4 emissions from the light-duty fleet. In general, motor vehicles that are high emitters of
5 hydrocarbons and carbon monoxide also will tend to be high emitters of PM (Sagebiel et al.,1997;
6 Cadle et al., 1997). Particle emission rates also are correlated with vehicle acceleration and
7 emissions occur predominantly during periods of heavy acceleration, even in newer vehicles
8 (Maricq et al., 1999).
9 Although the data shown in Table 3B-5 indicate that S (mainly in the form of sulfate) is a
10 minor component of PM2 5 emissions, S may be the major component of the ultrafme particles that
11 are emitted by either diesel or internal combustion engines (Gertler et al., 2000). It is not clear
12 what the source of the small amount of Pb seen in the auto exhaust profile is. It is extremely
13 difficult to find suitable tracers for automotive exhaust because Pb has been removed from
14 gasoline. However, it also should be remembered that restrictions in the use of leaded gasoline
15 have resulted in a dramatic lowering of ambient Pb levels.
16 Several tunnel studies have measured the distribution of organic and elemental carbon in the
17 integrated exhaust of motor vehicle fleets comprising several classes of motor vehicles (Pierson
18 and Brachaczek, 1983; Weingartner et al., 1997a; Fraser et al., 1998a). The study by Fraser et al.
19 (1998a) found that organic carbon constituted 46% of the carbonaceous particulate matter
20 emissions from the vehicles operating in the Van Nuys tunnel in Southern California in the
21 Summer of 1993. Although diesel vehicles constituted only 2.8% of the vehicles measured by
22 Fraser et al. (1998a), the contribution of the organic carbon to the total parti culate carbon
23 emissions obtained in the Van Nuys tunnels is in reasonable agreement with the dynamometer
24 measurements shown in Table 3B-4b.
25 Very few studies have reported comprehensive analyses of the organic composition of motor
26 vehicle exhaust. The measurements by Rogge et al. (1993c) are the most comprehensive, but are
27 not expected to be the best representation of current motor vehicle emissions because these
28 measurements were made in the mid-1980s. Measurements reported by Fraser et al. (1999) were
29 made in a tunnel study conducted in 1993 and represent integrated diesel and gasoline powered
30 vehicle emissions. In addition, exhaust emissions from two medium-duty diesel vehicles
31 operating over an FTP cycle were analyzed by Schauer et al. (1999). A unique feature of both the
March 2001 3B-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 measurements by Faser et al. (1999) and Schauer et al. (1999) is that they include the
2 quantification of unresolved complex mixture (UCM), which comprises aliphatic and cyclic
3 hydrocarbons that cannot be resolved by gas chromatography (Schauer etal., 1999). Schauer
4 et al. (1999) have shown that all of the organic compound mass in their diesel exhaust samples
5 could be extracted and eluted by CG/MS techniques, even though not all of the organic compound
6 mass can identified on a single compound basis. Table 3B-6 summarizes the composition of
7 motor vehicle exhaust measured by Fraser et al. (1999) and Schauer et al. (1999).
TABLE 3B-6. SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
EMITTED FROM MOTOR VEHICLES
Source
Gasoline and diesel-
powered vehicles
driving through the
Van Nuys Tunnel
(Fraser etal., 1999)a
Medium-duty diesel
vehicles operated over
an FTP Cycle
(Schauer etal., 1999)
Contribution to Dominant Contributors to
Compound Class Paniculate Mass (%) Emissions of Compound Class
n-Alkanes
Petroleum biomarkers
PAH
Aromatic acids
Aliphatic acids
Substituted aromatic
UCMb
n-Alkanes
Petroleum biomarkers
PAH
Aliphatic acids
Aromatic acids
Saturated cycloalkanes
UCMb
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
C21 through C29
Hopanes and steranes
No dominant compound
Benzenedicarboxylic acids
Palmitic and stearic acids
No dominant compound
C20 through C28
Hopanes and steranes
No dominant compound
n-Octadecanoic acid
Methylbenzoic acid
C21 through C25
includes emissions of brake wear, tire wear, and resuspension of road dust associated with motor vehicle traffic.
bUnresolved complex mixture.
March 2001
3B-17
DRAFT-DO NOT QUOTE OR CITE
-------
1 Several studies have measured the distribution of polycyclic aromatic hydrocarbons (PAHs)
2 in motor vehicles exhaust from on-road vehicles (Westerholm et al., 1991; Lowenthal et al., 1994;
3 Venkataraman et al., 1994; Westerholm and Egeback, 1994; Reilly et al., 1998; Cadle et al., 1999,
4 Weingartner et al., 1997b; Marr et al., 1999). Cadle et al. (1999) found high molecular weight
5 PAHs (PAHs with molecular weights greater than or equal to 202 g/mole) to make up from 0.1 to
6 7.0% of the particulate matter emissions from gasoline powered and diesel powered light duty
7 vehicles. It is important to note, however, that PAHs with molecular weights of
8 202 (fluoranthene, acephenanthrylene, and pyrene), 226 (benzo[ghi]fluoranthene and
9 cyclopenta[cd]pyrene), and 228 (benz[a]anthracene, chrysene, and triphenylene) exist in both the
10 gas-phase and particle-phase at atmospheric conditions (Fraser et al., 1998b). Excluding these
11 semi-volatile PAHs, the contribution of nonvolatile PAHs to the parti culate matter emitted from
12 the light-duty vehicles sampled by Cadle et al. (1999) ranges from 0.013 to 0.18%. These
13 measurements are in good agreement with the tunnel study conducted by Fraser et al. (1999) and
14 the heavy-duty diesel truck and bus exhaust measurements by Lowenthal et al. (1994), except that
15 the nonvolatile PAH emissions from the heavy duty diesel vehicles tested by Lowenthal et al.
16 (1994) were moderately higher, making up approximately 0.30% of the particulate matter mass
17 emissions.
18
19 Biomass Burning
20 In contrast to the mobile and stationary sources discussed earlier, emissions from biomass
21 burning in woodstoves and forest fires are strongly seasonal and can be highly episodic within
22 their peak emissions seasons. The burning of fuelwood is confined mainly to the winter months
23 and is acknowledged to be a major source of ambient air particulate matter in the northwestern
24 United States during the heating season. Forest fires occur primarily during the driest seasons of
25 the year in different areas of the country and are especially prevalent during prolonged droughts.
26 PM produced by biomass burning outside the United States (e.g., in Central America during the
27 spring of 1988) also can affect ambient air quality in the United States.
28 An example of the composition of fine particles (PM25) produced by woodstoves is shown
29 in Figure 3B-4. These data were obtained in Denver during the winter of 1987-1988 (Watson and
30 Chow, 1994). As was the case for motor vehicle emissions, organic and elemental carbon are the
31 major components of particulate emissions from wood burning. It should be remembered
March 2001 3B-18 DRAFT-DO NOT QUOTE OR CITE
-------
10"
10
10"
^^fg^^'^^S^ '^ ^^^
Chemical Species
Figure 3B-4. Chemical abundances for PM2 5 emissions from wood burning in Denver, CO.
Solid bars represent fractional abundances, and the error bars represent
variability in species abundances. Error bars represent detection limits when
there are no solid bars.
Source: Watson and Chow (1994).
1 that the relative amounts shown for organic carbon and elemental carbon vary with the type of
2 stove, the stage of combustion and the type and condition of the fuelwood. Fine particles are
3 dominant in studies of wood burning emissions. For instance, the mass median diameter of wood-
4 smoke particles was found to be about 0.17 //m in a study of the emissions from burning
5 hardwood, softwood, and synthetic logs (Dasch, 1982).
6 Kleeman et al. (1999) showed that the particles emitted by the combustion of wood in
7 fireplaces are predominately less than 1.0 //m in diameter, such that the composition of fine
8 particulate matter (PM25) emitted from fireplace combustion of wood is representative of the total
9 particulate matter emissions from this source. Hildemann et al. (1991) and McDonald et al.
10 (2000) reported that smoke from fireplace and wood stove combustion consists of 48% to
11 71% OC and 2.9% to 15% EC. Average elemental and organic carbon contents for these
12 measurements are shown in Table 3B-7. It should be noted that the two methods used for the
March 2001
3B-19
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-7. MASS EMISSIONS, ORGANIC CARBON, AND ELEMENTAL
CARBON EMISSIONS FROM RESIDENTIAL COMBUSTION OF WOOD
Wood Type
Softwood
Softwood
Hardwood
Hardwood
Hardwood
Combustion
Type
Fireplace
Fireplace
Fireplace
Fireplace
Wood Stove
Average Mass
Emission Rate
(gkg"1 of wood
burned)
13.0
5.14
5.28
5.66
3.96
Number
of Tests
2
5
3
5
8
Percent
Organic
Carbona
48.4
58.5
48.4
63.2
71.2
Percent
Elemental
Carbon8
5.2
15.0
2.9
7.0
9.0
References
Hildemann et al. (1991)
McDonald et al. (2000)
Hildemann etal. (1991)
McDonald et al. (2000)
McDonald et al. (2000)
"Hildemann et al. (1991) used the method described by Birch and Gary (1996) to measure EC and McDonald
et al. (2000) used the method reported by Chow et al. (1993) to measure OC.
1 measurements shown in Table 3B-7 have been reported to produce different relative amounts of
2 OC and EC for wood smoke samples, but show good agreement for total carbon (OC + EC)
3 measurements (Chow etal., 1993).
4 Hawthorne et al. (1988) and Hawthorne et al. (1989) measured gas-phase and particle-phase
5 derivatives of guaiacol (2-methoxyphenol), syringol (2,6-dimethoxyphenol), phenol, and catechol
6 (1,2-benzenediol) in the downwind plume of 28 residential wood stoves and fireplaces. Rogge
7 et al. (1998) reported a broad range of particle-phase organic compounds in the wood smoke
8 samples collected by Hildemann et al. (1991), which include n-alkanes, n-alkanoic acids,
9 n-alkenoic acids, dicarboxylic acids, resin acids, phytosterols, polycyclic aromatic hydrocarbons
10 (PAH), and the compounds reported by Hawthorne et al . (1989). Supplementing these
11 measurements, McDonald et al. (2000) reported the combined gas-phase and particle-phase
12 emissions of PAH and the compounds quantified by Hawthorne et al. (1989). The measurements
13 by Rogge et al. (1998), which represent a comprehensive data set of the organic compounds
14 present in wood smoke aerosol, are summarized in Table 3B-8. It should be noted, however, that
15 these nearly 200 compounds account for only approximately 15 to 25% of the organic carbon
16 particle mass emitted from the residential combustion of wood. Simoneit et al. (1999) have
17 shown that levoglucosan constitutes a noticeable portion of the organic compound mass not
18 identified by Rogge et al. (1998). In addition, Elias et al. (1999) used high-temperature gas
March 2001
3B-20
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-8. SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
EMITTED FROM THE COMBUSTION OF WOOD IN FIREPLACES
(Measurements were made using a dilution sampler and no
semi-volatile organic compound sorbent.)
Biomass Type
Fireplace
combustion of
softwood
Fireplace
combustion of
hardwood
Contribution to Paniculate Dominant Contributors to Emissions
Compound Class Mass (%) of Compound Class
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Dicarboxylic acids
Resin acids
Substituted phenols
Phytosterols
PAH
Oxygenated PAH
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Dicarboxylic acids
Resin acids
Substituted phenols
Phytosterols
PAH
Oxygenated PAH
0.039
0.45
0.12
0.36
1.28
3.30
0.37
0.092
0.019
0.044
1.33
0.049
0.42
0.11
8.23
0.21
0.13
0.020
C2i through C31
c c c c c c
M6> *"18! *^20> ^215 *^22> ^24
Oleic and linoleic acid
Malonic acid
Abietic, dehydroabietic, isopimaric,
pimaric, and sandaracopimaric acids
Benzenediols and guaiacols
p-Sitosterol
Fluoranthene and pyrene
IH-phenalen-l-one
C21 through C29
Cie, C22, C24, C26
Oleic and linoleic acid
Succinic acid
Dehydroabietic acid
Benzediols, guaiacols, and syringols
p-sitosterol
No dominant compounds
IH-phenalen-l-one
Source: Rogge et al. (1998).
1 chromatography/mass spectrometry (HTGC-MS) to measure high-molecular-weight organic
2 compounds in smoke from South American leaf and steam litter biomass burning. These
3 compounds cannot be measured by the analytical techniques employed by Rogge et al. (1998) and,
4 therefore, are strong candidates to make up some of the unidentified organic mass in the wood
March 2001
3B-21
DRAFT-DO NOT QUOTE OR CITE
-------
1 smoke samples analyzed by Rogge et al. (1998). These compounds, which include triterpenyl
2 fatty acid esters, wax esters, triglycerides, and high-molecular-weight n-alkan-2-ones, are
3 expected to be present in North American biomass smoke originating from agricultural burning,
4 forest fires, grassland fires, and wood smoke.
5 Measurements of aerosol composition, size distributions, and aerosol emissions factors have
6 been made in biomass burning plumes either on towers (Susott et al., 1991) or aloft on fixed-wing
7 aircraft (e.g., Radke et al., 1991) or on helicopters (e.g., Cofer et al., 1988). As was found for
8 woodstove emissions, the composition of biomass burning emissions is strongly dependent on the
9 stage of combustion (i.e., flaming, smoldering, or mixed), and the type of vegetation (e.g., forest,
10 grassland, scrub). Over 90% of the dry mass in particulate biomass burning emissions is
11 composed of organic carbon (Mazurek et al., 1991). Ratios of organic carbon to elemental carbon
12 are highly variable, ranging from 10:1 to 95:1, with the highest ratio found for smoldering
13 conditions and the lowest for flaming conditions. Emissions factors for total particulate emissions
14 increase by factors of two to four in going from flaming to smoldering stages in the individual
15 fires studied by Susott et al. (1991).
16 Particles in biomass burning plumes from a number of different fires were found to have
17 three distinguishable size modes, (1) a nucleation mode, (2) an accumulation mode, and
18 (3) a coarse mode (Radke et al., 1991). Based on an average of 81 samples, approximately 70%
19 of the mass was found in particles <3.5 //m in aerodynamic diameter. The fine particle
20 composition was found to be dominated by tarlike, condensed hydrocarbons and the particles were
21 usually spherical in shape. Additional information for the size distribution of particles produced
22 by vegetation burning was shown in Figure 3B-2.
23 An example of ambient data for the composition of PM2 5 collected at a tropical site that was
24 heavily affected by biomass burning is shown in Table 3B-9. The samples were collected during
25 November of 1997 on the campus of Sriwijaya University, which is located in a rural setting on
26 the island of Sumatra in Indonesia (Pinto et al., 1998). The site was subjected routinely to levels
27 of PM25 well in excess of the U.S. NAAQS as a result of the Indonesian biomass fires from the
28 summer of 1997 through the spring of 1988. As can be seen from a comparison of the data shown
29 in Table 3B-9 with those shown in Figure 3B-4, there are a number of similarities and differences
30 (especially with regard to the heavy metal content) in the abundances of many species. The
March 2001 3B-22 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-9. MEAN AEROSOL COMPOSITION AT TROPICAL SITE
(SRIWIJAYA UNIVERSITY, SUMATRA, INDONESIA) AFFECTED
HEAVILY BY BIOMASS BURNING EMISSIONS3
Component
OC
EC
scv
Al
Si
Cl
K
Ca
Ti
V
Abundance (%)
76
1.2
11
BDb
9.3 x 1C'2
4.4
0.7
4.5 x 1C'2
4.2 x 1C'3
BDb
Component
Cr
Mn
Fe
Ni
Cu
Zn
As
Se
Br
Pb
Abundance (%)
BDb
BDb
3.9 x 1C'2
<3.8 x 1Q-5
4.8 x 1C'4
3.1 x 1Q-3
6.4 x 1C'4
2.8 x 1C'4
3.6 x 1C'2
3.1 x 1C'3
The mean PM2 5 concentration during the sampling period (November 5 through 11, 1997) was 264
bBeneath detection limit.
Source: Pinto etal. (1998).
1
2
3
4
5
9
10
11
abundances of some crustal elements (e.g., Si, Fe) are higher in Table 3B-9 than in Figure 3B-4,
perhaps reflecting additional contributions of entrained soil dust.
Limited emissions data that includes 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) have been obtained.
March 2001
3B-23
DRAFT-DO NOT QUOTE OR CITE
-------
1 Natural Sources
2 Although sea-salt aerosol production is confined to salt water bodies, it is included here
3 because many marine aerosols can exert a strong influence on the composition of the ambient
4 aerosol in coastal areas. In some respects, the production of sea-salt aerosols is like that of
5 windblown dust in that both are produced by wind agitation of the surface. The difference
6 between the two categories arises because sea-salt particles are produced from the bursting of air
7 bubbles rising to the sea surface. Air bubbles are formed by the entrainment of air into the water
8 by breaking waves. The surface energy of a collapsing bubble is converted to kinetic energy in
9 the form of a j et of water that can ej ect drops above the sea surface. The mean diameter of the j et
10 drops is about 15% of the bubble diameter (Wu, 1979). Bubbles in breaking waves range in size
11 from a few //m to several mm in diameter. Field measurements by Johnson and Cooke (1979) of
12 bubble size spectra show maxima in diameters at around 100 //m, with the bubble size distribution
13 varying as (d/d0)"5 with d0 = 100 //m.
14 Because sea-salt particles receive water from the surface layer, which is enriched in organic
15 compounds, the aerosol drops are composed of this organic material in addition to sea salt (about
16 3.5% by weight in sea water). Na+ (30.7%), Cl' (55.0%), SO4= (7.7%), Mg2+ (3.6%), Ca2+ (1.2%),
17 K+ (1.1%), HCO3" (0.4%), and Br" (0.2%) are the major ionic species by mass in sea water
18 (Wilson, 1975). The composition of the marine aerosol also reflects the occurrence of
19 displacement reactions that enrich sea-salt particles in SO4" and NO3", while depleting them of Cl"
20 and Br.
21 Seasalt is concentrated in the coarse size mode with a mass median diameter of about 7 //m
22 for samples collected in Florida, the Canary Islands, and Barbados (Savoie and Prospero, 1982).
23 The size distribution of sulfate is distinctly bimodal. Sulfate in the coarse mode is derived from
24 sea water but sulfate in the submicron aerosol arises from the oxidation of dimethyl sulfide
25 (CH3SCH3) or DMS. DMS is produced during the decomposition of marine micro-organisms.
26 DMS is oxidized to methane sulfonic acid (MSA), a large fraction of which is oxidized to sulfate
27 (e.g., Herteletal., 1994).
28 Apart from sea spray, other natural sources of particles include the suspension of organic
29 debris and volcanism. Profiles of organic compounds in vegetative detritus have been obtained by
30 Rogge et al. (1993d). Particles are released from plants in the form of seeds, pollen, spores, leaf
31 waxes, and resins, ranging in size from 1 to 250 //m (Warneck, 1988). Fungal spores and animal
March 2001 3B-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 debris, such as insect fragments, also are to be found in ambient aerosol samples in this size range.
2 Although material from all the foregoing categories may exist as individual particles, bacteria
3 usually are found attached to other dust particles (Warneck, 1988). Smaller bioaerosol particles
4 include viruses, individual bacteria, protozoa, and algae (Matthias-Maser and Jaenicke, 1994).
5 In addition to natural sources, other sources of bioaerosol include industry (e.g., textile mills),
6 agriculture, and municipal waste disposal (Spendlove, 1974). The size distribution of bioaerosols
7 has not been characterized as well as it has for other categories.
8 Trace metals are emitted to the atmosphere from a variety of sources such as sea spray,
9 wind-blown dust, volcanoes, wildfires and biotic sources (Nriagu, 1989). Biologically mediated
10 volatilization processes (e.g., biomethylation) are estimated to account for 30 to 50% of the
11 worldwide total Hg, As, and Se emitted annually, whereas other metals are derived principally
12 from pollens, spores, waxes, plant fragments, fungi, and algae. It is not clear, however, how much
13 of the biomethylated species are remobilized from anthropogenic inputs. Median ratios of the
14 natural contribution to globally averaged total sources for trace metals are estimated to be
15 0.39 (As), 0.15 (Cd), 0.59 (Cr), 0.44 (Cu), 0.41 (Hg), 0.35 (Ni), 0.04 (Pb), 0.41 (Sb), 0.58 (Se),
16 0.25 (V), and 0.34 (Zn), suggesting a not insignificant natural source for many trace elements.
17 It should be noted though that these estimates are based on emissions estimates that have
18 uncertainty ranges of an order of magnitude.
March 2001 3B-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Bagley, S. T.; Gratz, L. D.; Johnson, J. H.; McDonald J. F. (1998) Effects of an oxidation catalytic converter and a
3 biodiesel fuel on the chemical, mutagenic, and particle size characteristics of emissions from a diesel engine.
4 Environ. Sci. Technol. 32: 1183-1191.
5 Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of paniculate diesel
6 exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
7 Cadle, S. H.; Mulawa, P. H.; Ball, J.; Donase, C; Weibel, A.; Sagebiel, J. C; Knapp, K. T.; Snow, R. (1997)
8 Paniculate emission rates from in-use high-emitting vehicles recruited in Orange county, California. Environ.
9 Sci. Technol. 31:3405-3412.
10 Cadle, S. H.; Mulawa, P.; Hunsanger, E. C.; Nelson, K.; Ragazzi, R. A.; Barrett, R.; Gallagher, G. L.; Lawson, D. R.;
11 Knapp, K. T.; Snow, R. (1999) Light-duty motor vehicle exhaust paniculate matter measurement in the
12 Denver, Colorado, area. J. Air Waste Manage. Assoc. 49: PM-164-174.
13 Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some
14 reflections on controlling vehicle emissions. Science (Washington, DC) 261: 37-45.
15 Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
16 thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
17 studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere;
18 April 1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
19 Chow, J. C.; Watson, J. G.; Houck, J. E.; Pritchett, L. C.; Rogers, C. F.; Frazier, C. A.; Egami, R. T.; Ball, B. M.
20 (1994) A laboratory resuspension chamber to measure fugitive dust size distributions and chemical
21 compositions. Atmos. Environ. 28: 3463-3481.
22 Cofer, W. R., Ill; Levine, J. S.; Sebacher, D. I.; Winstead, E. L.; Riggin, P. J.; Brass, J. A.; Ambrosia, V. G. (1988)
23 Paniculate emissions from a mid-latitude prescribed chaparral fire. J. Geophys. Res. [Atmos.] 93: 5207-5212.
24 Dasch, J. M. (1982) Paniculate and gaseous emissions from wood-burning fireplaces. Environ. Sci. Technol.
25 16:639-645.
26 Durbin, T. D.; Smith, M. R.; Norbeck, J. M.; Truex, T. J. (1999) Population density, paniculate emission
27 characterization, and impact on the paniculate inventory of smoking vehicles in the South Coast Air Quality
28 Management District. J. Air Waste Manage. Asssoc. 49: 28-38.
29 Elias, V. O.; Simoneit, B. R. T.; Pereira, A. S.; Cabral, J. A.; Cardoso, J. N. (1999) Detection of high molecular
3 0 weight organic tracers in vegetation smoke samples by high-temperature gas chromatography-mass
31 spectrometry. Environ. Sci. Technol. 33: 2369-2376.
32 Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. (1998a) Gas-phase and particle-phase organic compounds emitted from
33 motor vehicle traffic in a Los Angeles roadway tunnel. Environ. Sci. Technol. 32: 2051-2060.
34 Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T.; Rasmussen, R. A. (1998) Air quality model evaluation data for
35 organics. 5. C6-C22 nonpolar and semipolar aromatic compounds. Environ. Sci. Technol. 32: 1760-1770.
36 Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. (1999) Paniculate organic compounds emitted from motor vehicle
37 exhaust and in the urban atmosphere. Atmos. Environ. 33: 2715-2724.
38 Fujita, E.; Watson, J. G.; Chow, J. C.; Robinson, N. F.; Richards, L. W.; Kumar, N. (1998) Northern front range air
39 quality study, volume C: source apportionment and simulation methods and evaluation. Reno, NV: Desert
40 Research Institute; prepared for Colorado State University, Cooperative Institute for Research in the
41 Atmosphere.
42 Gertler, A. W.; Abu-Allaban, M.; Coulombe, W.; Gilles, J. A.; Pierson, W. R.; Rogers, C. F.; Sagebiel, J. C.; Tarnay,
43 L. (2000) Preliminary results of a tunnel study to characterize mobile source paniculate emissions. Presented
44 at: PM2000: paniculate matter and health—the scientific basis for regulatory decision-making, specialty
45 conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
46 Gillies, J. A.; Gertler, A. W. (2000) Comparison and evaluation of chemically speciated mobile source PM2 5
47 paniculate matter profiles. J. Air Waste Manage. Assoc. 50: 1459-1480.
48 Hawthorne, S. B.; Miller, D. J.; Barkley, R. M.; Krieger, M. S. (1988) Identification of methoxylated phenols as
49 candidate tracers for atmospheric wood smoke pollution. Environ. Sci. Technol. 22: 1191-1196.
50 Hawthorne, S. B.; Krieger, M. S.; Miller, D. J.; Mathiason, M. B. (1989) Collection and quantitation of methoxylated
51 phenol tracers for atmospheric pollution from residential wood stoves. Environ. Sci. Technol. 23: 470-475.
52 Hertel, O.; Christensen, J.; Hov, 0. (1994) Modelling of the end products of the chemical decomposition of DMS in
53 the marine boundary layer. Atmos. Environ. 28: 2431-2449.
March 2001 3B-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hildemann, L. M; Markowski, G. R.; Jones, M. C.; Cass, G. R. (1991) Submicrometer aerosol mass distributions of
2 emissions from boilers, fireplaces, automobiles, diesel trucks, and meat-cooking operations. Aerosol Sci.
3 Technol. 14: 138-152.
4 Hopke, P. K.; Lamb, R. E.; Natusch, D. F. S. (1980) Multielemental characterization of urban roadway dust. Environ.
5 Sci. Technol. 14: 164-172.
6 Houck, J. E.; Chow, J. C.; Watson, J. G.; Simons, C. A.; Pritchett, L. C.; Goulet, J. M.; Frazier, C. A. (1989)
7 Determination of particle size distribution and chemical composition of paniculate matter from selected
8 sources in California: volume I and executive summary (final report). Sacramento, CA: California Air
9 Resources Board.
10 Houck, J. E.; Goulet, J. M.; Chow, J. C.; Watson, J. G.; Pritchett, L. C. (1990) Chemical characterization of emission
11 sources contributing to light extinction. In: Mathai, C. V., ed. Visibility and fine particles: an A&WMA/EPA
12 international specialty conference; October 1989; Estes Park, CO. Pittsburgh, PA: Air & Waste Management
13 Association; pp. 437-446. (A&WMA transactions series no. TR-17).
14 Huffman, G. P.; Huggins, F. E.; Shah, N.; Huggins, R.; Linak, W. P.; Miller, C. A.; Pugmire, R. J.; Meuzelaar, H. L.
15 C.; Seehra, M. S.; Manivannan, A. (2000) Characterization of fine paniculate matter produced by combustion
16 of residual fuel oil. J. Air Waste Manage. Assoc. 50: 1106-1114.
17 Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
18 and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning.
19 Atmos. Environ. 30: 3825-3835.
20 Johnson, B. D.; Cooke, R. C. (1979) Bubble populations and spectra in coastal waters: a photographic approach.
21 J. Geophys. Res. C: Oceans Atmos. 84: 3761-3766.
22 Kerminen, V. M.; Makela, T. E.; Ojanen, C. H.; Hillamo, R. E.; Vilhunen, J. K.; Rantanen, L.; Havers, N.;
23 Von Bohlen, A.; Klockow, D. (1997) Characterization of the paniculate phase in the exhaust from a diesel
24 car. Environ. Sci. Technol. 31: 1883-1889.
25 Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
26 Kleeman, M. J.; Schauer, J. J.; Cass, G. R. (1999) Size and composition distribution of fine paniculate matter emitted
27 from wood burning, meat charbroiling, and cigarettes. Environ. Sci. Technol. 33: 3516-3523.
28 Kleeman, M. J.; Schauer, J. J.; Cass, G. R. (2000) Size and composition distribution of fine paniculate matter emitted
29 from motor vehicles. Environ. Sci. Technol. 34: 1132-1142.
30 Lawson, D. R.; Smith, R. E. (1998) The northern front range air quality study: a report to the Governor and General
31 Assembly. Fort Collins, CO: Colorado State University; December.
32 Linak, W. P.; Miller, C. A.; Wendt, J. O. L. (2000) Comparison of particle size distributions and elemental
3 3 partitioning from the combustion of pulverized coal and residual fuel oil. J. Air Waste Manage. Assoc.
34 50: 1532-1544.
35 Lowenthal, D. H.; Zielinska, B.; Chow, J. C.; Watson, J. G.; Gautam, M.; Ferguson, D. H.; Neuroth, G. R.; Stevens,
36 K. D. (1994) Characterization of heavy-duty diesel vehicle emissions. Atmos. Environ. 28: 731-743.
37 Maricq, M. M.; Podsiadlik, D. H.; Chase, R. E. (1999) Gasoline vehicle particle size distributions: comparison of
38 steady state, FTP, and US06 measurements. Environ. Sci. Technol. 33: 2007-2015.
39 Marr, L. C.; Kirchstetter, T. W.; Harley, R. A.; Miguel, A. H.; Hering, S. V; Hammond, S. K. (1999)
40 Characterization of polycyclic aromatic hydrocarbons in motor vehicle fuels and exhaust emissions. Environ.
41 Sci. Technol. 33:3091-3099.
42 Mason, B. (1966) Principles of geochemistry. 3rd ed. New York, NY: John Wiley & Sons, Inc.
43 Matthias-Maser, S.; Jaenicke, R. (1994) Examination of atmospheric bioaerosol particles with radii > 0.2 ^m.
44 J. Aerosol Sci. 25: 1605-1613.
45 Mazurek, M. A.; Cofer, W. R., Ill; Levine, J. S. (1991) Carbonaceous aerosols from prescribed burning of a boreal
46 forest ecosystem. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
47 implications. Cambridge, MA: MIT Press; pp. 258-263.
48 McDonald, J. D.; Zielinska, B.; Fujita, E. M.; Sagebiel, J. C.; Chow, J. C.; Watson, J. G. (2000) Fine particle and
49 gaseous emission rates from residential wood combustion. Environ. Sci. Technol. 34: 2080-2091.
50 Miller, C. A.; Linak, W. P.; King, C.; Wendt, J. O. L. (1998) Fine particle emissions from heavy fuel oil combustion
51 in a firetube package boiler. Combust. Sci. Technol. 134: 477-502.
52 Norbeck, J. M.; Durbin, T. D.; Truex, T. J. (1998) Measurement of primary paniculate matter emissions from
53 light-duty motor vehicles. Riverside, CA: University of California, College of Engineering, Center for
54 Environmental Research and Technology; prepared for Coordinating Research Council, Inc. and South Coast
5 5 Air Quality Management District.
March 2001 3B-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nriagu, J. O. (1989) A global assessment of natural sources of atmospheric trace metals. Nature (London)
2 338:47-49.
3 Nriagu, J. O.; Pacyna, J. M. (1988) Quantitative assessment of worldwide contamination of air, water and soils by
4 trace metals. Nature (London) 333: 134-139.
5 Olmez, I.; Sheffield, A. E.; Gordon, G. E.; Houck, J. E.; Pritchett, L. C.; Cooper, J. A.; Dzubay, T. G.; Bennett, R. L.
6 (1988) Compositions of particles from selected sources in Philadelphia for receptor modeling applications.
7 JAPCA 38: 1392-1402.
8 Oros, D. R.; Simoneit, B. R. T. (2000) Identification and emission rates of molecular tracers in coal smoke
9 paniculate matter. Fuel 79: 515-536.
10 Pierson, W. R.; Brachaczek, W. W. (1983) Paniculate matter associated with vehicles on the road. II. Aerosol Sci.
11 Technol. 2: 1-40.
12 Pinto, J. P.; Grant, L. D.; Hartlage, T. A. (1998) Report on U.S. EPA air monitoring of haze from S.E. Asia biomass
13 fires. Research Triangle Park, NC: U.S. Environmental Protection Agency, National Center for Environmental
14 Assessment; report no. EPA/600/R-98/071.
15 Radke, L. F.; Hegg, D. A.; Hobbs, P. V.; Nance, J. D.; Lyons, J. H.; Laursen, K. K.; Weiss, R. E.; Riggan, P. J.;
16 Ward, D. E. (1991) Paniculate and trace gas emissions from large biomass fires in North America. In: Levine,
17 J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric implications. Cambridge, MA: MIT
18 Press; pp. 209-224.
19 Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. (1998) Real-time characterization of the organic
20 composition and size of individual diesel engine smoke particles. Environ. Sci. Technol. 32: 2672-2679.
21 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine organic
22 aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
23 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
24 aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks. Environ.
25 Sci. Technol. 27: 1892-1904.
26 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
27 aerosol. 5. Natural gas home appliances. Environ. Sci. Technol. 27: 2736-2744.
28 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
29 aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci.
30 Technol. 27:636-651.
31 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993d) Sources of fine organic
32 aerosol. 4. Paniculate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol.
33 27:2700-2711.
34 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1994) Sources of fine organic
35 aerosol. 6. Cigarette smoke in the urban atmosphere. Environ. Sci. Technol. 28: 1375-1388.
36 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997a) Sources of fine organic
37 aerosol. 8. Boilers burning no. 2 distillate fuel oil. Environ. Sci. Technol. 31: 2731-2737.
38 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997b) Sources of fine organic
3 9 aerosol. 7. Hot asphalt roofing tar pot fumes. Environ. Sci. Technol. 31: 2726-2730.
40 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1998) Sources of fine organic
41 aerosol. 9. Pine, oak, and synthetic log combustion in residential fireplaces. Environ. Sci. Technol. 32: 13-22.
42 Sagebiel, J. C.; Zielinska, B.; Walsh, P. A.; Chow, J. C.; Cadle, S. H.; Mulawa, P. A.; Knapp, K. T.; Zweidinger,
43 R. B.; Snow, R. (1997) PM-10 exhaust samples collected during IM-240 dynamometer tests of in-service
44 vehicles in Nevada. Environ. Sci. Technol. 31: 75-83.
45 Savoie, D. L.; Prospero, J. M. (1982) Particle size distribution of nitrate and sulfate in the marine atmosphere.
46 Geophys. Res. Lett. 9: 1207-1210.
47 Sawyer, R. F.; Johnson, J. H. (1995) Diesel emissions and control technology. Diesel exhaust: a critical analysis of
48 emissions, exposure, and health effects. Cambridge, MA: Health Effects Institute.
49 Schauer, J. J.; Cass, G. R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants
50 using organic compounds as tracers. Environ. Sci. Technol. 34: 1821-1832.
51 Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. (1999) Measurement of emissions from air pollution
52 sources. 2. Q through C30 organic compounds from medium duty diesel trucks. Environ. Sci. Technol.
53 33: 1578-1587.
54 Simoneit, B. R. T. (1985) Application of molecular marker analysis to vehicular exhaust for source reconciliations.
5 5 Int. J. Environ. Anal. Chem. 22: 203 -23 3.
March 2001 3B-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 Simoneit, B. R. T.; Schauer, J. I; Nolle, C. G.; Oros, D. R.; Elias, V. O.; Fraser, M. P.; Rogge, W. F.; Cass, G. R.
2 (1999) Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles. Atmos. Environ. 33:
3 173-182.
4 Spendlove, J. C. (1974) Industrial, agricultural, and municipal microbial aerosol problems. In: Current problems in
5 aerobiology: proceedings of the thirtieth general meeting of the Society for Industrial Microbiology; August
6 1973; Evanston, IL. Dev. Ind. Microbiol. 15: 20-27.
7 Standley, L. J.; Simoneit, B. R. T. (1987) Characterization of extractable plant wax, resin, and thermally matured
8 components in smoke particles from prescribed burns. Environ. Sci. Technol. 21: 163-169.
9 Susott, R. A.; Ward, D. E.; Babbitt, R. E.; Latham, D. J. (1991) The measurement of trace emissions and combustion
10 characteristics for a mass fire. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and
11 biospheric implications. Cambridge, MA: MIT Press; pp, 245-257.
12 Turekian, K. K. (1971) Elements, geochemical distribution of. In: Encyclopedia of science and technology, v. 4. New
13 York, NY: McGraw-Hill Book Company; pp. 627-630.
14 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
15 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
16 U.S. Environmental Protection Agency. (1999) Health assessment document for diesel emissions, SAB review draft.
17 Washington, DC: Office of Research and Development; report no. EPA/600/8-90/057C.
18 Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
19 and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci. Technol.
20 28: 555-562.
21 Vinogradov, A. P. (1959) The geochemistry of rare and dispersed chemical elements in soils. 2nd ed. New York,
22 NY: Consultants Bureau, Inc.
23 Warneck, P. (1988) Chemistry of the natural atmosphere. New York, NY: Academic Press, Inc.
24 Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ.
25 19:241-266.
26 Watson, J. G.; Fujita, E. M.; Chow, J. C.; Zielinska, B.; Richards, L. W.; Neff, W.; Dietrich, D. (1998) Northern fron
27 range air quality study final report. Reno, NV: Desert Research Institute; prepared for Colorado State
28 University, Cooperative Institute for Research in the Atmosphere.
29 Weingartner, E.; Burtscher, H.; Baltensperger, U. (1997a) Hygroscopic properties of carbon and diesel soot particles.
30 Atmos. Environ. 31: 2311-2327.
31 Weingartner, E.; Keller, C.; Stahel, W. A.; Burtscher, H.; Baltensperger, U. (1997b) Aerosol emission in a road
32 tunnel. Atmos. Environ. 31: 451-462.
33 Westerholm, R.; Egeback, K.-E. (1994) Exhaust emissions from light- and heavy-duty vehicles: chemical
34 composition, impact of exhaust after treatment, and fuel parameters. In: Symposium of risk assessment of
35 urban air: emissions, risk identification, and risk quantitation; May-June 1992; Stockholm, Sweden. Environ.
36 HealthPerspect. 102(suppl. 4) 13-23.
37 Westerholm, R. N.; Almen, J.; Li, H.; Rannug, J. U.; Egeback, K.-E.; Gragg, K. (1991) Chemical and biological
3 8 characterization of paniculate-, semivolatile-, and gas-phase-associated compounds in diluted heavy-duty
39 diesel exhausts: a companson of three different semivolatile-phase samplers. Environ. Sci. Technol.
40 25: 332-338.
41 Wilson, T. R. S. (1975) Salinity and the major elements of sea water. In: Riley, J. P.; Skirrow, G., eds. Chemical
42 oceanography: v. 1. 2nd ed. London, United Kingdom: Academic Press Inc.; pp. 365-413.
43 Wu, J. (1979) Spray in the atmospheric surface layer: review and analysis of laboratory and oceanic results.
44 J. Geophys. Res. C: Oceans Atmos. 84: 1693-1704.
45 Yanowitz, J.; McCormick, R. L.; Grabowski, M. S. (2000) In-use emissions from heavy-duty diesel vehicles.
46 Environ. Sci. Technol. 34: 729-740.
March 2001 3B-29 DRAFT-DO NOT QUOTE OR CITE
-------
i 4. ENVIRONMENTAL EFFECTS OF
2 PARTICULATE MATTER
3
4
5 4.1 INTRODUCTION
6 Several later chapters (Chapters 5 through 8) of this document assess the latest available
7 information on determinants of human exposures to particulate matter (PM); dosimetry of
8 particle deposition, clearance, and retention in human respiratory tract; epidemiologic analyses of
9 health effects associated with human exposures to ambient PM; and toxicologic evaluations of
10 pathophysiologic effects of PM and underlying mechanisms of action. The human exposure and
11 health-related findings assessed in those chapters provide key elements of the scientific bases to
12 support upcoming decision making regarding potential retention or revision of the primary PM
13 National Ambient Air Quality Standards (NAAQS). This chapter, in contrast, assesses
14 information pertinent to decision making regarding secondary standards aimed at protecting
15 against welfare effects of PM. More specifically, this chapter assesses environmental effects of
16 atmospheric PM, including discussion of PM effects on vegetation and ecosystems, PM effects
17 on visibility, PM effects on man-made materials, and relationships of ambient PM to global
18 climate change processes.
19
20
21 4.2 EFFECTS ON VEGETATION AND ECOSYSTEMS
22 The Parti culate Matter National Ambient Air Quality Standards (PM NAAQS) set in 1971
23 were specified in terms of total suspended particulates (TSP), which included both fine and
24 coarse mode particles (the latter ranging up to 25 to 40 //m in size). The 1987 revision of the
25 TSP NAAQS to PM10 standards focused attention on those particles (< 10 //m mean aerometic
26 diameter) capable of being deposited in lower (thoracic) portions of the human respiratory tract.
27 The subsequent 1997 PM NAAQS revisions retained the PM10 standards and added fine particle
28 (PM25) standards (both specified in terms of mass concentrations of particles undifferentiated in
29 terms of their specific chemical composition). The effects of PM on vegetation and ecosystems
30 as a basis for a secondary standard were not considered as part of the 1997 PM NAAQS
March 2001 4-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 revisions. Vegetation and ecosystem effects of ambient PM evaluated in this chapter are
2 dependent not so much just on PM size-related mass concentration, but rather on exposure of
3 plants to PM components differentiated by chemical composition as well.
4 Particulate matter deposition to vegetation is not well understood. Unlike gaseous dry
5 deposition, neither the solubility of the particles nor the physiological activity of the surface are
6 likely to be of first order importance in determining deposition velocity (Vd). Factors that
7 contribute to surface wetness and stickiness may be critical determinants of sticking efficiency.
8 Available tabulations of deposition velocities are highly variable and suspect. High-elevation
9 forests receive larger particulate deposition loadings than equivalent lower elevation sites,
10 because of higher wind speeds and enhanced rates of aerosol impaction; orographic effects on
11 rainfall intensity and composition; increased duration of occult deposition; and, in many areas,
12 the dominance of coniferous species with needle-shaped leaves (Lovett, 1984). Recent evidence
13 indicates that all three modes of deposition, (1) wet, (2) occult, and (3) dry, must be considered in
14 determining inputs to watersheds or ecosystems because each may dominate over specific
15 intervals of time or space.
16 Exposure to a given mass concentration of airborne PM may lead to widely differing
17 phytotoxic responses, depending on the particular mix of deposited particles. The most common
18 and useful subdivision of PM, derived from the typical bimodal distribution of atmospheric
19 particles, is into fine and coarse particles (Wilson and Suh, 1997). The smallest particle at or
20 near 1.0 to 2.5 //m generally is taken as the division between fine and coarse, although this is not
21 an absolute and is subject to some shift (e.g., with changing ambient humidity). However, the
22 typical the rule of thumb, as previously used in the 1996 PM Air Quality Criteria Document or
23 "PM AQCD" (U.S. Environmental Protection Agency, 1996a), is that fine PM nominally falls in
24 the range of 0 to 2.5 //m and coarse-mode PM, 2.5 to 10.0 //m.
25 In general, fine-mode PM is secondary in nature, having condensed from the vapor phase or
26 been formed by chemical reaction from gaseous precursors in the atmosphere. These particles
27 exist in a nucleation mode (having a mass median aerodynamic diameter or MMAD of about
28 0.06 //m) and may grow by coagulation of existing particles or by condensation of additional
29 gases onto existing particles into an accumulation mode (about 0.5 //m). Sulfur and nitrogen
30 oxides (SOX and NOX), as well as volatile organic gases, are common precursors for fine PM,
31 often neutralized with ammonium cations as particulate salts. Condensation of volatilized metals
March 2001 4-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 and products of incomplete combustion also are common precursors. Reactions of many of these
2 materials with an oxidizing atmosphere lead to high secondary PM concentrations during
3 summer months in many parts of the United States.
4 In general, coarse-mode particles are primary in nature, having been produced and emitted
5 from a point or area source as a fully formed particle. They range in size from ca. 2.5 to 100 //m.
6 This material is created by abrasion and subsequent suspension by wind or mechanical means.
7 Suspended geologic material contains the chemical and, potentially, the biological signature of
8 the soil from which it derives (dominated by iron, silica, aluminum, and calcium). Additional
9 anthropogenically derived coarse-mode PM derives from fly ash, automobile tires and brake
10 linings, and industrial effluent associated with crushing and grinding operations. Coarse-mode
11 particles also include biogenically derived organic materials (e.g., fragments of plants and
12 insects, pollen, fungal spores, bacteria and viruses included in marine aerosols).
13 Atmospheric deposition of particles to ecosystems takes place via both wet and dry
14 processes through three major routes: (1) precipitation scavenging in which particles are
15 deposited in rain and snow; (2) fog, cloud-water, and mist interception; and (3) dry deposition,
16 a much slower, yet more continuous removal to surfaces (Hicks, 1986).
17 Precipitation scavenging includes rainout involving within-cloud nucleation phenomena
18 and washout involving below-cloud scavenging by impaction. Total inputs from wet deposition
19 to vegetative canopies can be significant (Table 4-1), although not all wet deposition involves
20 particle scavenging because gaseous pollutants also dissolve during precipitation.
21 Wet deposition is not affected by surface properties as much as is dry or occult deposition.
22 However, forested hillsides may receive much (four- to sixfold) greater precipitation than short
23 vegetation in nearby valleys because of a variety of orographic effects (Unsworth and Wilshaw,
24 1989). Additionally, closer aerodynamic coupling to the atmosphere of the tall forest canopy
25 than of the shorter canopies in the valleys leads to more rapid foliar drying, reduced residence
26 time of solubilized particulate materials available for foliar uptake, and, consequently, more rapid
27 and more extreme concentration of such materials on the cuticular surface. The results of direct
28 physical effects on leaves are not known.
29 Most of wet deposited particulate material passes through the plant canopy to the soil by
30 throughfall and stemflow, causing soil-mediated ecosystem-level responses. Rainfall also
31 removes much of the dry-deposited PM resident on foliar surfaces, reducing direct foliar effects
March 2001 4-3 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-1. RELATIVE IMPORTANCE OF WET, DRY, PARTICIPATE,
AND TOTAL DEPOSITION TO THREE FOREST SITES3
Deposition
Total Nitrogenb
Site
Duke Forest
Gary Forest
Austin Forest
Wet
(%)
75
71
71
Dry
(%)
25
20
29
Particle
(%)
0.11
0.94
0.58
Total
(kg ha1)
9.87
5.80
6.57
Wet
(%)
64
76
83
Total Sulfuf
Dry
(%)
33
20
13
Particle
(%)
2.7
4.2
4.3
Total
(kg ha1)
17.20
7.60
7.79
aData from Allen et al. (1994). Sampling was by triple filter pack, so that fine-mode particles could be sampled
preferentially. An average particle deposition velocity of 0.9 cm s"1 was derived, as in Hicks et al. (1987).
bWet nitrogen consists of NO3" and NH4+, dry nitrogen consists of vapor phase HNO3 and NO2, and paniculate
nitrogen consists of NO3".
°Wet sulfur consists of SO4~, dry sulfur consists of vapor phase SO2, and paniculate sulfur consists of pSO4".
1 (Lovett and Lindberg, 1984). This washing effect, combined with differential foliar uptake and
2 foliar leaching (both of which depend on the physiological status of the vegetation), alters the
3 composition of rainwater that reaches the soil. Dry deposition onto foliage and subsequent wet
4 removal by runoff enhances soil-mediated effects of paniculate deposition, both by enhancing
5 total dry deposition relative to unvegetated surfaces nearby and by accelerating passage of
6 deposited particles to the soil. The most significant effects of wet deposition occur through soil-
7 mediated processes involving biogeochemcial cycling of major and minor nutrients and trace
8 elements.
9 Dry deposition is more effective for coarse particles of natural origin and elements such as
10 iron and manganese, whereas wet deposition generally is more effective for fine PM of
11 atmospheric origin and elements such as cadmium, chromium, lead, nickel, and vanadium
12 (Smith, 1990a). The actual importance of wet versus dry deposition, however, is highly variable,
13 depending on ecosystem type, location, and elevation. For the Walker Branch Watershed, a
14 deciduous forest in rural eastern Tennessee, dry deposition constituted a major fraction of total
15 annual atmospheric input of cadmium and zinc (=20%), lead (=55%), and manganese (=90%),
16 but wet deposition rates for single precipitation events exceeded dry deposition rates by one to
17 four orders of magnitude (Lindberg and Harriss, 1981). Miller et al. (1993) emphasized that
18 immersion of high-elevation forests in cloudwater for 10% or more of the year can enhance
19 significantly overall efficiency of transfer of atmospheric particles and gases to a forest canopy.
March 2001 4-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 Dry deposition of particles occurs to all vegetational surfaces exposed to the atmosphere
2 (U.S. Environmental Protection Agency, 1982). The range of particle sizes, the diversity of
3 canopy surfaces, and the variety of chemical constituents in airborne PM have slowed progress in
4 both prediction and measurement of dry particulate deposition. Wet deposition generally is
5 confounded by fewer factors and has been easier to quantify (Chapter 2).
6 Emphasis in this and the next section is placed on discussion of PM effects on individual
7 plants in natural habitats and terrestrial ecosystems. Except for the deposition of nitrogen and
8 sulfur-containing compounds and their effects exerted via acidic precipitation, information
9 concerning the effects of deposition of other specific substances as PM on crops is not readily
10 available. The U.S. National Acid Precipitation Assessment Program (NAPAP) Biennial Report
11 to Congress: An Integrated Assessment presents an extensive overall discussion of the effects of
12 acidic deposition (National Science and Technology Council, 1998). The effects of gaseous
13 sulfur oxides and nitrogen oxides on crops are discussed in detail in EPA criteria documents for
14 those substances (U.S. Environmental Protection Agency, 1982, 1993). A detailed discussion of
15 the ecological effects of acidic precipitation and nitrate deposition on aquatic ecosystems also
16 can be found in the EPA Nitrogen Oxides Air Quality Criteria Document (U.S. Environmental
17 Protection Agency, 1993). Neither nitrate or sulfate deposition on crops is discussed in this
18 chapter, as they are added frequently in fertilizers. Also, the effects of lead on crops, vegetation,
19 and ecosystems are discussed in the EPA document, Air Quality Criteria for Lead (U.S.
20 Environmental Protection Agency, 1986).
21 The effects of deposited PM may be direct or indirect. Indirect effects are chiefly
22 nutritional responses mediated through the soil and result from the effects of PM components on
23 soil processes. In the following sections, the direct effects on individual plants are discussed
24 first, followed by effects on plant species and their interactions in ecosystems.
25 4.2.1 Direct Effects of Particulate Matter on Individual Plant Species
26 Particulate matter in the atmosphere may affect vegetation directly following deposition on
27 foliar surfaces, indirectly by changing the soil chemistry, or through changes in the amount of
28 radiation reaching the Earth's surface through PM-induced climate change processes. Indirect
29 impacts, however, are usually the most significant because they can alter nutrient cycling and
March 2001 4-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 inhibit plant nutrient uptake. The possible direct responses to PM deposition are considered in
2 this section, and the indirect responses in the sections on ecosystems.
3 Particles transferred from the atmosphere to foliar surfaces may reside on the leaf, twig or
4 bark surface for extended periods; be taken up through the leaf surface; or be removed from the
5 plant via resuspension to the atmosphere, washing by rainfall, or litter-fall with subsequent
6 transfer to the soil. Any PM deposited on above-ground plant parts may exert physical or
7 chemical impacts. The effects of "inert" PM are mainly physical, whereas those of toxic particles
8 are both chemical and physical. The chemical effects of dust deposited on plant surfaces or soil
9 are more likely to be associated with their chemistry than simply with the mass of deposited
10 particles and may be more important than any physical effects (Farmer, 1993).
11 Studies of the direct effects of chemical additions to foliage in particulate deposition have
12 found little or no effects of PM on foliar processes unless exposure levels were significantly
13 higher than typically would be experienced in the ambient environment. Interpretation of the
14 effects of atmospheric chemical deposition at the level of individual plants and ecosystems is
15 difficult because of the complex interactions that exist among biological, physicochemical, and
16 climatic factors. The majority of the easily identifiable direct and indirect effects, other than
17 climate, occur in severely polluted areas around heavily industrialized point sources, such as
18 limestone quarries, cement kilns, and smelting facilities for iron, lead, or various other metals.
19 The diverse chemical nature and size characteristics of ambient airborne particles and the lack of
20 any clear distinction between effects attributed to phytotoxic particles and to other forms of air
21 pollutants confound the direct effects of PM on foliar surfaces. Most documented toxic effects of
22 particles on vegetation reflect their acidity, trace metal content, nutrient content, surfactant
23 properties, or salinity. These materials typically elicit similar biological effects, whether
24 deposited as coarse or fine particles, in wet, dry, or occult form, and, frequently, whether
25 deposited to foliage or to the soil. Studies of direct effects of particles on vegetation have not yet
26 advanced to the stage of reproducible exposure experiments. Experimental difficulties in
27 application of ambient particles to vegetation have been discussed by Olszyk et al. (1989).
28 4.2.1.1 Effects of Coarse Particles
29 Coarse-mode particles, ranging in size from 2.5 to 100 //m, are chemically diverse, are
30 dominated by local sources, and are typically deposited near the source because of their
March 2001 4-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 sedimentation velocities. Airborne coarse particles are derived from road, cement kiln, and
2 foundry dust; fly ash; tire particles and brake linings; soot and cooking oil droplets; biogenic
3 materials (e.g., plant pollen, fragments of plants, fungal spores, bacteria and viruses) and sea salt.
4 Most coarse particles in rural and some urban areas are composed of silicon, aluminum, calcium,
5 and iron, suggesting that their main source is fugitive dust from disturbed land, roadways,
6 agricultural tillage, or construction. Rapid sedimentation of these particles tends to restrict their
7 direct effects on vegetation largely to roadsides and forest edges.
8 Physical Effects—Radiation. Dust can have both a physical and chemical impact.
9 Deposition of inert PM on above-ground plant organs may result in an increase in radiation
10 received, in leaf temperature and blockage of stomata. Increased leaf temperature, heat stress,
11 reduced net photosynthesis, and leaf chlorosis, necrosis, and abscission were reported by
12 Guderian (1986). Road dust decreased the leaf temperature on Rhododendron catawbiense by
13 ca. 4 °C (Eller, 1977), whereas foundry dust caused an 8.7 °C increase in leaf temperature of
14 black poplar (Populus nigrd) (Guderian, 1986) under the conditions of the experiment.
15 Broad-leaved plants exhibited greater temperature increases because of particle loading than did
16 the needle-like leaves of conifers. Deciduous (broad) leaves exhibited larger temperature
17 increases because of particle loading than did conifer (needle) leaves, a function of poorer
18 coupling to the atmosphere. Inert road dust caused a three- to fourfold increase in the absorption
19 coefficient of leaves ofHedera helix (Eller, 1977; Guderian, 1986) for near infrared radiation
20 (NIR; 750 to 1350 nm). Little change occurred in absorption for photosynthetically active
21 radiation (PAR; 400 to 700 nm). The increase in NIR absorption was equally at the expense of
22 reflectance and transmission in these wavelengths. The net energy budget increased by ca. 30%
23 in the dust-affected leaves. Deposition of coarse particles increased leaf temperature and
24 contributed to heat stress, reduced net photosynthesis, and caused leaf chlorosis, necrosis, and
25 abscission (Dassler et al., 1972; Parish, 1910; Guderian, 1986; Spinka, 1971).
26 Starch storage in dust-affected leaves increased with dust loading under high (possibly
27 excessive) radiation, but decreased following dust deposition when radiation was limiting. These
28 modifications of the radiation environment had a large impact on single-leaf utilization of light.
29 The boundary layer properties, determined by leaf morphology and environmental conditions,
March 2001 4-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 strongly influenced the direct effects of particle deposition on radiation heating (Eller, 1977;
2 Guderian, 1986) and on gas exchange as well.
3 Brandt and Rhoades (1973) attributed the reduction in growth of trees because of crust
4 formation from limestone dust on the leaves. Crust formation reduced photosynthesis and
5 formation of carbohydrate needed for normal growth, induced premature leaf-fall, destruction of
6 leaf tissues, inhibited growth of new tissue, and reduced storage. Dust may affect
7 photosynthesis, respiration, and transpiration, and it may allow penetration of phytotoxic gaseous
8 pollutants, thereby causing visible injury symptoms and decreased productivity. Permeability of
9 leaves to ammonia increased with increasing dust concentrations and decreasing particle size
10 (Farmer, 1993).
11 Dust also has been reported to physically block stomata (Krajickova and Mejstfik, 1984).
12 Stomatal clogging by particulate matter from automobiles, stone quarries, and cement plants was
13 also studied by Abdullah and Iqbal (1991). The percentage of clogging was low in young leaves
14 when compared with old and mature leaves and the amount of clogging varied with species and
15 locality. The maximum clogging of stomata observed was about 25%. The authors cited no
16 evidence that stomatal clogging inhibited plant functioning. The heaviest deposit of dust is
17 usually on the upper surface of broad-leaved plants, however, whereas the majority of the
18 stomata are on the lower surface where stomatal clogging would be less likely.
19 Chemical Effects. The chemical composition of PM is usually the key phytotoxic factor
20 leading to plant injury. Cement-kiln dust on hydration liberates calcium hydroxide, which can
21 penetrate the epidermis and enter the mesophyll, and, in some cases, the leaf surface alkalinity
22 may reach to pH 12. Lipid hydrolysis coagulation of the protein compounds and ultimately
23 plasmolysis of the leaf tissue result in reduction in growth and quality of plants (Guderian, 1986).
24 In experimental studies, application of cement kiln dust of known composition for 2 to 3 days
25 yielded dose-response curves between net photosynthetic inhibition or foliar injury and dust
26 application rate (Darley, 1966). Lerman and Darley (1975) determined that leaves must be
27 misted regularly to produce large effects. Alkalinity was probably the essential phytotoxic
28 property of the applied dusts.
29 Particulate matter in the form of sea salt enters the atmosphere from oceans following
30 mixing of air into the water and subsequent bursting of bubbles at the surface. This process can
March 2001 4-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 be a significant source of sulfate, sodium chloride, and trace elements in the atmosphere over
2 coastal vegetation, resulting in the formation of the maritime forest, a specialized ecosystem.
3 Sea-salt particles can serve as nuclei for the adsorption and subsequent reaction of other gaseous
4 and paniculate pollutants. Both nitrate and sulfate from the atmosphere have been found to be
5 associated with coarse and fine sea-salt particles (Wu and Okada, 1994). Direct effects on
6 vegetation reflect these inputs, as well as classical salt injury caused by the sodium and chloride
7 that constitute the bulk of these particles. Salt pruning is a common phenomenon near the ocean
8 (i.e., salt spray kills the buds on the windward side of trees and shrubs).
9 4.2.1.2 Effects of Fine Particles
10 Fine PM is generally secondary in nature, having condensed from the vapor phase or been
11 formed by chemical reaction from gaseous precursors in the atmosphere and is generally smaller
12 than 1 to 2.5 //m. Nitrogen and sulfur oxides, volatile organic gases, condensation of volatilized
13 metals, and products of incomplete combustion are common precursors for fine PM. Reactions
14 of many of these materials with an oxidizing atmosphere contribute to high secondary PM
15 concentrations during summer months in many U.S. areas. The conclusion reached in the 1982
16 PM AQCD (U.S. Environmental Protection Agency, 1982) that sufficient data were not available
17 for adequate quantification of dose-response functions for direct effects of fine aerosols on
18 vegetation continues to be true today. Only a few studies have been completed on the direct
19 effects of acid aerosols (U. S. Environmental Protection Agency, 1982). The major effects are
20 indirect and occur through the soil (Section 4.3).
21 Nitrogen. Nitrate is observed in both fine and coarse particles. Nitrates from atmospheric
22 deposition represent a substantial fraction of total nitrogen inputs to southeastern forests (e.g.,
23 Lovett and Lindberg, 1986). However, much of this is contributed by gaseous nitric acid vapor,
24 and a considerable amount of the particulate nitrate is taken up indirectly, through the soil.
25 Garner et al. (1989) estimated deposition of nitrogen to forested landscapes in eastern North
26 America at 10 to 55 kg/ha/year for nitrate and 2 to 10 kg/ha/year for ammonium. About half of
27 these values were ascribed to dry deposition.
28 Atmospheric additions of particulate nitrogen in excess of vegetation needs are lost from
29 the system, mostly as leachate from the soil as nitrate. Managed agricultural ecosystems may be
March 2001 4-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 able to utilize deposited participate nitrogen more efficiently than native ecosystems, although
2 many cultivated systems also lose considerable nitrogen as nitrate in runoff, deep drainage, or tail
3 water. It has proven difficult to quantify direct foliar fertilization by uptake of nitrogen from
4 ambient particles.
5 There is no doubt that foliar uptake of nitrate can occur, as clearly shown by the efficacy of
6 foliar fertilization in horticultural systems. Potassium nitrate was taken up by leaves of
7 deciduous fruit trees (Weinbaum and Neumann, 1977) and resulted in increased foliar nitrogen
8 concentrations. Not all forms of nitrogen are absorbed equally, nor are all equally benign.
9 Following foliar application of 2600 ppm of nitrogen as Ca(NO3)2, (NH4)2SO4, or (NH2)2CO to
10 apple canopies (Rodney, 1952; Norton and Childers, 1954), leaf nitrogen levels were observed to
11 increase to similar levels, but calcium nitrate and ammonium sulfate caused visible foliar
12 damage, whereas urea did not. Urea is generally the recommended horticultural foliar fertilizer.
13 The mechanism of uptake of foliarly deposited nitrate is not well established. Nitrate
14 reductase is generally a root-localized enzyme. It is generally not present in leaves, but is
15 inducible there. This typically occurs when the soil is heavily enriched in NO3". As the root
16 complement of nitrate reductase becomes overloaded, unreduced nitrate reaches the leaves
17 through the transpiration stream. Nitrate metabolism has been demonstrated in leaf tissue
18 (Weinbaum and Neumann, 1977) following foliar fertilization. Residual nitrate reductase
19 activity in leaves may be adequate to assimilate typical rates of paniculate nitrate deposition.
20 Uptake of nitrate may be facilitated by codeposited sulfur (Karmoker et al., 1991; Turner and
21 Lambert, 1980).
22 Nitrate reductase is feedback-inhibited by its reaction product, NH4+. The common
23 atmospheric aerosol, NH4NO3, therefore may be metabolized in two distinct biochemical steps,
24 first the ammonium (probably leaving nitric acid) and then the nitrate. Volatilization losses of
25 nitric acid during this process, if they occur, have not been characterized.
26 Direct foliar effects of particulate nitrogen have not been documented. Application of a
27 variety of fine nitrogenous aerosol particles (0.25 //m) ranging from 109 to 244 //g/m3 nitrogen,
28 with or without 637 //g/m3 sulfur, caused no consistent short-term (2- to 5-h) effect on gas
29 exchange in oak, maize, or soybean leaves (Martin et al., 1992).
30 Although no evidence exists for direct transfer of nutrient particulate aerosols into foliage,
31 a few studies give insights into the potential for ammonium and nitrate transfer into leaves.
March 2001 4-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 Fluxes of both NO3" and NH4+, measured in wet deposition and in throughfall plus stemflow in
2 forests, commonly indicate higher fluxes of nitrogen above the canopy (Parker, 1983; Lindberg
3 et al., 1987; Sievering et al., 1996), indicating net foliar uptake. Lovett and Lindberg (1993)
4 reported a linear relationship between inorganic nitrogen fluxes in deposition and throughfall,
5 suggesting that uptake may be considered passive to some extent.
6 Garten and Hanson (1990) studied the movement of 15N-labeled nitrate and ammonium
7 across the cuticles of red maple (Acer rubrum) and white oak (Quercus alba) leaves when
8 applied as an artificial rain mixture. Brumme et al. (1992), Bowden et al. (1989), and Vose and
9 Swank (1990) have published similar data for conifers. These studies show the potential for
10 nitrate and ammonium to move into leaves, where it may contribute to normal physiological
11 processes (e.g., amino acid production; Wellburn, 1990). Garten (1988) showed that internally
12 translocated 35S was not leached readily from tree leaves of yellow poplar (Liriodendron
13 tulipiferd) and red maple (Acer rubrum), suggesting that SO42" would not be as mobile as the
14 nitrogen-containing ions discussed by Garten and Hanson (1990). Further, when the foliar
15 extraction method is used it is not possible to distinguish sources of chemical deposited as gases
16 or particles (e.g., nitric acid [HNO3], nitrogen dioxides [NO2], nitrate [NO3"], or sources of
17 ammonium deposited as ammonia [NH3] or ammonium ion [NH4+]) (Garten and Hanson, 1990).
18 Particle deposition contributes only a portion of the total atmospheric nitrogen deposition
19 reaching vegetation but, when combined with gaseous and precipitation-derived sources, total
20 nitrogen deposition to ecosystems has been identified as a possible causal factor leading to
21 changes in natural ecosystems (See Section 4.3).
22 Sulfur. Anthropogenic sulfur emissions are >90% as SO2. Most of the remaining emission
23 of sulfur is directly as sulfate (U.S. Environmental Protection Agency, 1996a). Sulfur dioxide is
24 hydrophilic and is rapidly hydrated and oxidized to sulfite and bisulfite and then to sulfate, which
25 is approximately 30-fold less phytotoxic. The ratio of sulfate/SO2 increases with aging of the air
26 mass and, therefore, with distance from the source. Sulfate is sufficiently hygroscopic that, in
27 humid air, it may exist significantly in the coarse particulate fraction. As dilution of both SO2
28 and particulate SO42" occurs with distance from the source, it is unusual for damaging levels of
29 particulate sulfate to be deposited. Gas to particle conversion in this case is of benefit to
30 vegetation.
March 2001 4-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sulfur is an essential plant nutrient. Low dosages of sulfur serve as a fertilizer, particularly
2 for plants growing in sulfur-deficient soil (Hogan et al., 1998). Current levels of sulfate
3 deposition reportedly exceed the capacity of most vegetative canopies to immobilize the sulfur
4 (Johnson, 1984). Nitrogen uptake in forests may be regulated loosely by sulfur availability, but
5 sulfate additions in excess of needs do not typically lead to injury (Turner and Lambert, 1980).
6 There are few field demonstrations of foliar sulfate uptake (Krupa and Legge, 1986, 1998).
7 Sulfate in throughfall is often enriched above levels in precipitation. The relative importance of
8 foliar leachate and prior dry-deposited sulfate particles remains difficult to quantify (Cape et al.,
9 1992). Leaching rates are not constant and may respond to levels of other pollutants, including
10 acids. Uptake and foliar retention of gaseous and particulate sulfur are confounded by variable
11 rates of translocation and accessibility of deposited materials to removal and quantification by
12 leaf washing. Following soil enrichment with 35SO42" in a Scots pine forest, the apparent
13 contribution of leachate to throughfall was only a few percent, following an initial burst of over
14 90% because of extreme disequilibrium in labeling of tissue sulfate pools (Cape et al., 1992).
15 Olszyk et al. (1989) provide information on the impacts of multiple pollutant exposures
16 including particles (NO3; 142 //g/m3; NH4+, 101 //g/m3; SO42; 107 //g/m3). They found that only
17 gaseous pollutants produced direct (harmful) effects on vegetation for the concentrations
18 documented, but the authors hypothesized that long-term accumulation of the nitrogen and sulfur
19 compounds contributed from particle deposition might have effects on plant nutrition over long
20 periods of time. Martin et al. (1992) exposed oak (Quercus macrocarpa\ soybean (Glycine
21 max), and maize (Zea mays) plants to acute exposures (2 to 5 h) of aerosols (0.25 //m) containing
22 only nitrate (109 //g/m3), ammonium and nitrate (244 and 199 //g/m3, respectively), or
23 ammonium and sulfate (179 and 637 //g/m3, respectively). They found that these exposures,
24 which exceeded the range of naturally occurring aerosol concentrations, had little effect on foliar
25 photosynthesis and conductance. Martin et al. (1992) concluded that future investigations should
26 focus on the effects of particles on physiological characteristics of plants following chronic
27 exposures.
28 Acidic Deposition. The effects of acidic deposition have been accorded wide attention in
29 the media and elsewhere (Altshuller and Linthurst, 1984; Hogan et al., 1998). Probably the most
30 extensive assessment of acidic deposition processes and effects is the NAPAP Biennial Report to
March 2001 4-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 Congress: An Integrated Assessment (National Science and Technology Council, 1998).
2 Concern regarding the effects of acidic deposition on crops and forest trees has resulted in
3 extensive monitoring and research. Exposures to acidic rain or clouds can be divided into
4 "acute" exposures to higher ionic concentrations (several //mol/L), and "chronic" long-term
5 repeated exposures to lower concentrations (Cape, 1993). Pollutant concentrations in rainfall
6 have been shown to have little capacity for producing direct effects on vegetation (Altshuller and
7 Linthurst, 1984; Hogan et al., 1998); however, fog and clouds, which may contain solute
8 concentrations up to 10 times those found in rain, have the potential for direct effects. More than
9 80% of the ionic composition of most cloud water is made up of four major pollutant ions: H+,
10 NH4+, NO3", and SO42". Ratios of hydrogen to ammonium and sulfate to nitrate vary from site to
11 site with all four ions usually present in approximately equal concentrations. Available data from
12 plant effect studies suggest that hydrogen and sulfate ions are more likely to cause injury than
13 ions containing nitrogen (Cape, 1993).
14 The possible direct effects of acidic precipitation on forest trees have been evaluated by
15 experiments on seedlings and young trees. The size of mature trees makes experimental
16 exposure difficult, therefore necessitating extrapolations from experiments on seedlings and
17 saplings; however, such extrapolations must be used with caution (Cape, 1993). Both conifers
18 and deciduous species have shown significant effects on leaf surface structures after exposure to
19 simulated acid rain or acid mist at pH 3.5. Some species have shown subtle effects at pH 4 and
20 above. Visible lesions have been observed on many species at pH 3 and on sensitive species at
21 pH 3.5 (Cape, 1993). The relative sensitivities of forest vegetation to acidic precipitation based
22 on macroscopic injury have been ranked as follows: herbaceous dicots > woody dicots >
23 monocots > conifers (Percy 1991).
24 Huttunen (1994) described the direct effects of acid rain or acidic mist on epicuticular
25 waxes whose ultrastructure is affected by plant genotype and phenotype. The effects of air
26 pollutants on epicuticular waxes of conifers have received greater study than the waxes of other
27 species. Leaf age and shorter life span of broad-leaved trees make them less indicative of the
28 effects of acid precipitation. Many experimental studies indicate that epicuticular waxes that
29 function to prevent water loss from plant leaves can be destroyed by acid rain in a few weeks
30 (Huttunen, 1994). This function is crucial in conifers because of their longevity and evergreen
31 foliage. Microscopic observations of epicuticular wax structures have, for a long time, suggested
March 2001 4-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 links between acidic deposition and aging. In Norway spruce (Picea abies), acid rain causes not
2 only the aging of needles (which in northern conditions normally last from 11 to 14 years) to be
3 shortened but also accelerates the erosion rate of the waxes as the needles age.
4 The effects of acidic precipitation and fog on red spruce (Picea rubens) have been studied
5 extensively (Schier and Jensen, 1992). Visible foliar injury of the needles in the form of a
6 reddish-brown discoloration has been observed on red spruce seedlings experimentally exposed
7 to acidic mist, but this visible symptom has not been observed in the field. Ultrastructural
8 changes in the epicuticular wax were observed both experimentally and on spruce growing at
9 high elevations. Laboratory studies indicate that visible injury usually does not occur unless the
10 pH is 3 or less (Schier and Jensen, 1992). Cape (1993) reported that, when compared with other
11 species, red spruce seedlings appeared to be more sensitive to acid mist. Huttunen (1994)
12 concluded that his studies of conifers and review of the literature suggest that acidic precipitation
13 causes direct injury to tree foliage and, also, indirect effects through the soil. The indirect effects
14 of acidic precipitation are discussed in Section 4.3.
15 Based on his review of the many studies involving field and controlled laboratory
16 experiments on crops in the literature, Cape (1993) drew a number of conclusions concerning the
17 direct effects of acidic precipitation on crops:
18 • foliar injury and growth reduction occurs below pH 3;
19 • allocation of photosynthate is altered, with increased shoot to root ratios;
20 • expanded and recently expanded leaves are most susceptible, and injury occurs first to
21 epidermal cells;
22 • leaf surface characteristics such as wettability, buffering capacity, and transport of
23 material across the leaf surface contribute to susceptibility and differ among species;
24 • data obtained from experiments in greenhouses or controlled environmental chambers
25 cannot be used to predict effects on plants grown in the field;
26 • quantitative data from experimental exposures cannot be extrapolated to field exposures
27 because of differences and fluctuations in concentrations, durations, and frequency of
28 exposure;
29 • there are large differences in response within species;
30 • timing of exposure in relation to phenology is of utmost importance;
31 • plants may be able to recover from or adapt to injurious exposures; and
March 2001 4-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 • sequential exposure to acidic precipitation and gaseous pollutants is unlikely to be more
2 injurious than exposure to individual pollutants.
3 Studies by Chevone et al. (1986), Krupa and Legge (1986), and Blaschke (1990) differ with
4 the last conclusion of Cape listed above. Their studies indicate that interactions between acidic
5 deposition and gaseous pollutants do occur. Acidity affects plant responses to both O3 and SO2.
6 Chevone et al. (1986) observed increased visible injury on soybean and pinto bean when acid
7 aerosol exposure preceded O3 exposure, whereas linear decreases in dry root weight of yellow
8 poplar occurred as acidity increased with simultaneous exposures to O3 and simulated acid rain.
9 Krupa and Legge (1986) also noted increased visible injury to pinto bean when aerosol exposure
10 preceded O3 exposure. In none of the studies cited above did acid rain per se produce significant
11 growth changes. Blaschke (1990) observed a decrease in ectomycorrhizal frequency and short
12 root distribution caused by acid rain exposure in combination with either SO2 or O3.
13 Trace Elements. All but 10 of the 90 elements that comprise the inorganic fraction of the
14 soil occur at concentrations of less than 0.1% (1000 //g/g) and are termed "trace" elements.
15 Trace elements with a density greater than 6 g.cm"3, referred to as "heavy metals", are of
16 particular interest because of their potential toxicity for plant and animals. Although some trace
17 metals are essential for vegetative growth or animal health, they are all toxic in large quantities.
18 Combustion processes produce metal chlorides that tend to be volatile and metal oxides that tend
19 to be nonvolatile in the vapor phase (McGowan et al., 1993). Most trace elements exist in the
20 atmosphere in particulate form as metal oxides (Ormrod, 1984). Aerosols containing trace
21 elements derive predominantly from industrial activities (Ormrod, 1984). Generally, only
22 cadmium, chromium, nickel, and mercury are released from stacks in the vapor phase (McGowan
23 etal., 1993). Concentrations of heavy metals in incinerator fly ash increase with decreasing
24 particle size.
25 Vegetational surfaces, especially the foliage, present a major reaction and filtration surface
26 to the atmosphere and act to accumulate particles deposited via wet and dry processes described
27 in Chapter 2 (Tong, 1991; Youngs et al., 1993). Particles deposited on foliar surfaces may be
28 taken up through the leaf surface. The greatest particle loading is usually on the adaxial (upper)
29 leaf surface where particles accumulate in the mid-vein, center portion of the leaves. The
March 2001 4-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 mycelium of fungi becomes particularly abundant on leaf surfaces as the growing season
2 progresses and is in intimate association with deposited particles (Smith, 1990b).
3 Investigations of trace elements present along roadsides and in industrial and urban
4 environments indicate that impressive burdens of particulate heavy metals can accumulate on
5 vegetative surfaces. Foliar uptake of available metals could result in metabolic impact in above-
6 ground tissues. Only a few metals, however, have been documented to cause direct phytotoxicity
7 in field conditions. Copper, zinc, and nickel toxicities have been observed most frequently. Low
8 solubility, however, limits foliar uptake and direct heavy metal toxicity. A trace metal must be
9 brought into solution before it can enter into leaves or bark of vascular plants. In those instances
10 when trace metals are absorbed, they are frequently bound in leaf tissue and are lost when the leaf
11 drops off (Hughes, 1981). Trace metals in mixtures may interact to cause a different plant
12 response when compared with a single element; however, there has been little research on this
13 aspect (Ormrod, 1984). In experiments using chambers, Marchwinska and Kucharski (1987)
14 studied the effects of SO2 alone and in combination with PM components (Pb, Cd, Zn, Fe, Cu,
15 and Mn) obtained from a zinc smelter bag filter. The combined effects of SO2 and PM further
16 increased the reduction in yield of beans caused by SO2, whereas the combination, though
17 severely injuring the foliage, produced little effect on carrots and parsley roots, except after
18 long-term exposures (when there was a decrease in root weight).
19 Trace metal toxicity of lichens has been demonstrated in relatively few cases. Nash (1975)
20 documented zinc toxicity in the vicinity of a zinc smelter near Palmerton, PA. Lichen species
21 richness and abundance were reduced by approximately 90% in lichen communities at Lehigh
22 Water Gap near the zinc smelter when compared with those at Delaware Water Gap. Zinc,
23 cadmium, and sulfur dioxide were present in concentrations toxic to some species near the
24 smelter; however, toxic zinc concentrations extended beyond the detectable limits of sulfur
25 dioxide (Nash, 1975). Experimental data suggest that lichen tolerance to Zn and Cd falls
26 between 200 and 600 ppm (Nash, 1975).
27 Though there has been no direct evidence of a physiological association between tree injury
28 and exposure to metals, heavy metals have been implicated because their deposition pattern is
29 correlated with forest decline. The role of heavy metals has been indicated by phytochelatin
30 measurements. Phytochelatins are intracellular metal-binding peptides that act as specific
31 indicators of metal stress. Because they are produced by plants as a response to sublethal
March 2001 4-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations of heavy metals, they can be used to indicate that heavy metals play a role in
2 forest decline (Gawel et al., 1996). Concentrations of heavy metals increased with altitude, as
3 did forest decline, and increased concentrations across the region showing increased levels of
4 forest injury, as well.
5 Phytochelatin concentrations were measured in red spruce and balsam fir (Abies balsamea)
6 needles throughout the 1993 growing season at 1000 m on Whiteface Mountain in New York.
7 Mean foliar concentrations in red spruce were consistently higher than in balsam fir from June
8 until August, with the greatest and most significant difference occurring at the peak of the
9 growing season in mid-July. In July, the phytochelatin concentrations were significantly higher
10 than at any other time measured. Balsam fir did exhibit this peak, but maintained a consistently
11 low level throughout the season. Both the number of dead red spruce trees and phytochelatin
12 concentrations increased sharply with elevation (Gawel et al., 1996). The relationship between
13 heavy metals and the decline of forests in northeastern United States was further tested by
14 sampling red spruce stands showing varying degrees of decline at 1000 m on nine mountains
15 spanning New Hampshire, Vermont, and New York. The collected samples indicated a
16 systematic and significant increase in phytochelatin concentrations associated with the extent of
17 tree injury. The highest phytochelatin concentrations were measured during 1994 from sites
18 most severely affected by forest decline in the Green Mountains, VT, and the Adirondack
19 Mountains, NY. These data strongly imply that metal stress is a cause of tree injury and,
20 therefore, contributes to forest decline in the northeastern United States (Gawel et al., 1996).
21 One potential direct impact of heavy metals is on the activity of microorganisms and
22 arthropods resident on and in the leaf surface ecosystem. The fungi and bacteria living on and in
23 the surfaces of leaves play an important role in the microbial succession that prepares leaves for
24 decay and litter decomposition after their fall (U.S. Environmental Protection Agency, 1996b).
25 Numerous fungi were consistently isolated from foliar surfaces, at various crown positions,
26 from London plane trees growing in roadside environments in New Haven, CT. Those existing
27 primarily as parasites included Aureobasidium pullulans, Chaetomium sp., Cladosporium sp.,
28 Epicoccum sp., and Philaphora verrucosa. Those existing primarily as parasites included
29 Gnomoniaplatam, Pestalotiposis sp., and Pleurophomella sp. The following cations were tested
30 in vitro for their ability to influence the growth of these fungi: cadmium, copper, manganese,
31 aluminum, chromium, nickel, iron, lead, sodium, and zinc. Results indicated variable fungal
March 2001 4-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 response with no correlation between saprophytic or parasitic activity and sensitivity to heavy
2 metals. Both linear extension and dry weight data indicated that the saprophytic Chaetomum sp.
3 was very sensitive to numerous metals. Aureobasidiumpullulans, Epicoccum sp., and especially
4 P.verrucosa, on the other hand, appeared to be much more tolerant. Of the parasites, G. platani
5 appeared to be more tolerant than Pestalotiopsis sp. and Pleurophomella sp. Metals exhibiting
6 the broadest spectrum growth suppression were iron, aluminum, nickel, zinc, manganese, and
7 lead (Smith and Staskawicz, 1977; Smith, 1990c). These in vitro studies employed soluble
8 compounds containing heavy metals. Trace metals probably occur naturally on leaf surfaces as
9 low-solubility oxides, halides, sulfates, sulfides, or phosphates (Clevenger et al., 1991; Koslow
10 et al., 1977). In the event of sufficient solubility and dose, however, changes in microbial
11 community structure on leaf surfaces because of heavy metal accumulation are possible.
12 Organic Compounds. Fine particles in the atmosphere reacting with volatilized chemical
13 compounds are partitioned between the gas and particle phases, depending on the liquid phase
14 vapor pressure at the ambient atmospheric temperature, the surface area of the particles per unit
15 volume of air, the nature of the particles and of the chemical being adsorbed and can be removed
16 by wet and dry deposition (McLachlan, 1996a). Materials as diverse as DDT, polychlorinated
17 biphenyls (PCBs), and polynuclear aromatic hydrocarbons (PAHs) are being deposited from the
18 atmosphere on rural as well as urban landscapes (Kylin et al., 1994). Motor vehicles emit
19 particles to the atmosphere from several sources in addition to the tailpipe. Rogge et al. (1993)
20 inventoried the organic contaminants associated with fine particles (diameter <2.0 //m) in road
21 dust, brake lining wear particles, and tire tread debris. In excess of 100 organic compounds were
22 identified in these samples, including n-alkanols, benzoic acids, benaldehydes, polyalkylene
23 glycol ethers, PAHs, oxy-PAH, steranes, hopanes, natural resins, and other compound classes.
24 A large number of PAHs, ranging from naphthalene (C10H8) to 5- and 6-ring and higher PAHs,
25 their alkyl-substituted analogues, and their oxygen- and nitrogen-containing derivatives are
26 emitted from motor vehicle sources (Seinfeld, 1989).
27 Plants may be used as environmental monitors to compare the deposition of PAH, POPs, or
28 SOCs between sites (e.g., urban versus rural) (Wagrowski and Kites, 1997; Ockenden et al.,
29 1998; McLachlan, 1999). Vegetation can be used qualitatively to indicate organic pollutant
30 levels as long as the mechanism of accumulation is considered. The substance may enter the
March 2001 4-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 plant via the roots or, as mentioned above, deposited as a particle onto the waxy cuticle of leaves
2 or uptake thorough the stomata. The pathways are a function of the chemical and physical
3 properties of the pollutant, such as its lipophilicity, water solubility, vapor pressure (which
4 controls the vapro-particle partitioning) and Henry's law constant; environmental conditions,
5 such as ambient temperature and the organic content of the soil; and the plant species, which
6 controls the surface area and lipids available for accumulation (Simonich and Kites, 1995).
7 Ockenden et al. (1998) have observed that, for lipophilic POPs, atmospheric transfer to plant has
8 been the main avenue of accumulation. Plants can differentially accumulate POPs. Results have
9 shown differences between species with higher concentrations in the lichen (Hypogymnia
10 physiodes) than in pine needles (Pinus sylvestris). Even plants of the same species, because they
11 have different growth rates and different lipid contents (depending on the habitat in which they
12 are growing), have different rates of sequestering pollutants. These facts confound data
13 interpretations and must be taken into account when considering their use as passive samplers.
14 Vegetation itself is an important source of hydrocarbon aerosols. Terpenes, particularly
15 a-pinene, p-pinene, and limonene released from tree foliage, may react in the atmosphere to form
16 submicron particles. These naturally generated organic particles contribute significantly to the
17 blue haze aerosols formed naturally over forested areas (Smith, 1990d).
18 The low water solubility with high lipoaffmity of many of these organic xenobiotics
19 strongly control their interaction with the vegetative components of natural ecosystems. The
20 cuticles of foliar surfaces are covered with a wax layer that helps protect plants from moisture
21 and short-wave radiation stress. This epicuticular wax, consisting mainly of long-chain esters,
22 polyesters, and paraffins, has been demonstrated to accumulate lipophilic compounds. Organic
23 air contaminants, in the paniculate or vapor phase, are absorbed to and accumulate in the
24 epicuticular wax of vegetative surfaces (Gaggi et al., 1985; Kylin et al., 1994). Direct uptake of
25 organic contaminants through the cuticle or the vapor-phase uptake through the stomates are
26 characterized poorly for most trace organics.
27 The phytotoxicity and microbial toxicity of organic contaminants to soil microorganisms is
28 not well studied (Foster, 1991).
March 2001 4-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.2.2 Particulate Matter Effects on Natural Ecosystems
2 4.2.2.1 Introduction
3 Human existence on this planet depends on nature and the life-support services ecosystems
4 provide. Ecosystem services (Table 4-2) are the conditions and processes through which natural
5 ecosystems, and the species of which they are comprised, sustain and fulfill human life (Daily,
6 1997). Both ecosystem structure and function play an essential role in providing societal
7 benefits. Society derives two types of benefits from the structural aspects of an ecosystem:
8 (1) products with market value such as fish, minerals, forage, forest products, biomass fuels,
9 natural fiber, and many pharmaceuticals and the genetic resources of valuable species (e.g.,
10 plants for crops and timber, animals for domestication); and (2) ecosystem services (Table 4-2)
11 include the use and appreciation of ecosystems for recreation, aesthetic enjoyment, and study
12 (Westman, 1977; Daily, 1997). Economic benefits and values associated with ecosystem
13 functions and services and the need to preserve them because of their value to human life are
14 discussed by Costanza et al. (1997) and (Pimentel et al., 1997). Services usually are not
15 considered to be items with market value.
TABLE 4-2. ECOSYSTEM SERVICES
• Purification of air and water
• Mitigation of floods and droughts
• Detoxification and decomposition of wastes
• Generation and renewal of soil and soil fertility
• Pollination of crops and natural vegetation
• Control of the vast majority of potential agricultural pests
• Dispersal of seeds and translocation of nutrients
• Maintenance of biodiversity, from which humanity has derived key elements of its
agricultural, medicinal, and industrial enterprises
• Protection from the sun's harmful rays
• Partial stabilization of climate
• Moderation of temperature extremes and the force of winds and waves
• Support of diverse human cultures
• Providing of aesthetic beauty and intellectual stimulation that lift the human spirit
Source: Daily (1997).
March 2001 4-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ecosystems are structurally complex biotic communities consisting of populations of
2 plants, animals, insects, and microorganisms interacting with one another and with their abiotic
3 environment (Odum, 1993). They are dynamic, self-adjusting, self-maintaining, complex
4 adaptive systems in which patterns at higher levels of organization emerge from localized
5 interactions and selection processes. Macroscopic ecosystem properties such as structure,
6 diversity-productivity relationships and patterns of nutrient flux emerge from the interactions
7 among components and may feed back to influence subsequent development of those
8 interactions. The relationship between structure and function is a fundamental one in ecosystem
9 science (Levin, 1998). Structure refers to the species, their biodiversity, abundance, mass, and
10 arrangement within an ecosystem. Ecosystem functions, energy flow, nutrient flux, and water
11 and material flow, are characterized by the way in which ecosystem components interact.
12 Elucidating these interactions across scales is fundamental to understanding the relationships
13 between biodiversity and ecosystem functioning (Levin, 1998). To function properly and
14 maintain themselves, ecosystem components must have an adequate supply of energy, chemical
15 nutrients, and water. It is the flows of nutrients and energy, that provide the interconnectedness
16 between ecosystem parts and transforms the community from a random collection of species into
17 an integrated whole, an ecosystem in which the biotic and abiotic parts are interrelated (Levin,
18 1998).
19 Growth of new trees and other vegetation requires energy in the form of carbon
20 compounds. Plants accumulate, store, and use carbon compounds to build their structures and
21 maintain physiological processes. Plants, using energy from sunlight, in their leaves combine
22 carbon dioxide from the atmosphere and water from the soil to produce the carbon compounds
23 (sugars) that provide the energy required by vegetation for growth and maintenance (Waring and
24 Schlesinger, 1985). Energy is transferred through an ecosystem from organism to organism in
25 food webs and, finally, is dissipated into the atmosphere as heat (Odum, 1993). Chemical
26 nutrients, such as nitrogen, phosphorus, or sulfur, on the other hand, are taken up from the soil by
27 plants and are transferred to other species through the food webs. The process is cyclic with the
28 chemical nutrients eventually returning to the soil. This process is referred to as biogeochemical
29 cycling (Odum, 1993). The biogeochemistry of an ecosystem is influenced by vegetation growth
30 characteristics (Herbert et al., 1999).
March 2001 4-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ecosystem functions are characterized by the way components interact. These are the
2 functions that maintain clean water, pure air, a green earth, and a balance of organisms, the
3 functions that enable humans to survive. They are the dynamics of ecosystems. The benefits
4 they impart include absorption and breakdown of pollutants, cycling of nutrients, binding of soil,
5 degradation of organic waste, maintenance of a balance of gases in the air, regulation of radiation
6 balance, climate, and the fixation of solar energy (Table 4-2; Westman, 1977; Daily, 1997).
7 Concern has risen in recent years regarding the consequences of changing biological
8 diversity of ecosystems (Tilman, 2000; Ayensu et al., 1999; Wall, 1999; Hooper and Vitousek,
9 1997; Chapin et al., 1998). The concerns arise because human activities are creating
10 disturbances that are causing the loss of biodiversity and altering the complexity and stability of
11 ecosystems and producing changes in nutrient cycling (structure and function) (Pimm, 1984;
12 Levin, 1998; Chapin et al., 1998; Peterson et al., 1998; Tilman, 1996; Tilman and Downing,
13 1994; Wall, 1999; Daily and Ehrlich, 1999). There are few ecosystems on earth today that are
14 not influenced by humans (Freudenburg and Alario, 1999; Vitousek et al., 1997; Matson et al.
15 1997; Noble and Dirzo, 1997). The scientific literature is filled with references discussing the
16 importance of ecosystem structure and function. Ecorisk, complexity, stability, biodiversity,
17 resilience, sustainability, managing earth's ecosystems, and ecosystem health are frequently
18 discussed topics. There is a need, therefore, to understand how ecosystems respond to both
19 natural and anthropogenic stresses and, especially, the ways that anthropogenic stresses are
20 impacting ecosystem services and products.
21
22 4.2.2.2 Ecosystem Responses to Stress
23 Ecosystem responses to stresses begin at the population level. Population changes
24 however, begin with the response of individual plants or animals. Plant responses, both
25 structural and functional, must be scaled in both time and space and propagated from the
26 individual to the more complex levels of community interaction to produce observable changes
27 in an ecosystem (Figure 4-1). At least three levels of biological interaction are involved: (l)the
28 individual plant and its environment, (2) the population and its environment, and
29 (3) the biological community composed of many species and its environment (Billings, 1978).
30 The response of individual organisms within a population based on their genetic constitution
31 (genotype), stage of growth at time of exposure, and the microhabitats in which they are growing
March 2001 4-22 DRAFT-DO NOT QUOTE OR CITE
-------
^< Reactio
_evel of^
Organization
Leaf
(cm2)
Branch
(cm2)
Tree
(m2)
Stand
(ha)
n Time
Minute
•-
Day
*-^
Year
^-
^
k» •
N
^\ |
X-^J
•-^
•
Decade
1
3
-4
^•5
*-6
••7
f-8
»9
N
Vfc
V
1
Century
10
11
12
13
^14
^
16
Injury Symptom
Needle necrosis
and abscission
Branch length,
bifurcation ratio,
and ring-width
growth altered
Reduction in
diameter and death
of tree
Decreases in
stand productivity,
increases in mortality
and alterations in
regeneration patterns
Key Changes in Processes
Reduced carbon assimilation
because of reduced radiation
Reduced carbon available for foliage
replacement and branch growth/
export Synergistic interaction
between mistletoe and tephra
deposition
Reduced carbon available for
height, crown, and stem growth
Influence of crown class on initial
impact and subsequent recovery
Interaction between stand
composition and recovery
For a given level, the dot associated with a line begins with a process (e.g., photosynthesis for #1 under leaf) an
ends with the associated structure (e.g., the needle).
Evaluating Impacts Within a Level of Organization
Leaf Level
Branch Level
Carbon exchange-1
Carbon pools-2
Needle number and size-3
Needle retention/abscission-4
Carbon allocation-5
Branch growth-6
Branch morphology-7
Branch vigor-8
Branch retention-9
Tree Level Height and diameter growth-10
Crown shape and size-1 1
Tree vigor-12
Mortality-13
Stand Level Productivity-14
Mortality-15
Species composition-16
Evaluating Interactions Between Different Levels of Organization
The diagonal arrow indicates the interaction between any two levels of 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 4-1. 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 impacts 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).
March 2001
4-23
DRAFT-DO NOT QUOTE OR CITE
-------
1 vary in their ability to withstand the stress of environmental changes (Levin, 1998). Individual
2 organisms within a population vary in their ability to withstand the stress of environmental
3 changes. The range within which these organisms can exist and function determines the ability
4 of the population to survive. Those able to cope with the stresses survive and reproduce.
5 Competition among the different species results in succession (community change over time) and
6 ultimately produces ecosystems composed of populations of plant species that have the capability
7 to tolerate the stresses (Rapport and Whitford, 1999; Guderian, 1985).
8 The number of species in a community usually increases during succession in unpolluted
9 atmospheres. Productivity, biomass, community height, and structural complexity increase.
10 Severe stresses, on the other hand, divert energy from growth and reproduction to maintenance,
11 and return succession to an earlier stage (Waring and Schlesinger, 1985). Ecosystems are subject
12 to natural periodic stresses, such as drought, flooding, fire, and attacks by biotic pathogens (e.g.,
13 fungi, insects). Ecosystem perturbation by natural stresses can be only a temporary setback.
14 Extremely severe natural perturbations return succession to an earlier stage, reduce ecosystem
15 structure (scarcity of life forms and no symbiotic interactions) and functions, disrupt the plant
16 processes of photosynthesis and nutrient uptake, carbon allocation and transformation that are
17 directly related to energy flow and nutrient cycling, shorten food chains, and reduce the total
18 nutrient inventory (Odum, 1993). This transformation, however, sets the stage for recovery,
19 which permits the perturbed ecosystem to adapt to changing environments (Holling, 1986).
20 Therefore, these perturbations are seldom more than a temporary setback, and recovery can be
21 rapid (Odum, 1969).
22 In contrast, anthropogenic stresses usually are severe, debilitating stresses. Severely
23 stressed ecosystems do not recover readily, but may be further degraded (Odum, 1969; Rapport
24 and Whitford, 1999). Anthropogenic stresses can be classified into four main groups:
25 (1) physical restructuring (e.g., changes resulting from land use); (2) introduction of exotic
26 species; (3) over harvesting; and (4) discharge of toxic substances into the atmosphere, onto land,
27 and into water. Ecosystems lack the capacity to adapt to the above stresses and maintain their
28 normal structure and functions unless the stress is removed (Rapport and Whitford, 1999). These
29 stresses result in a process of degradation marked by a decrease in biodiversity, reduced primary
30 and secondary production, and a lower capacity to recover and return to its original state.
31 In addition, there is an increased prevalence of disease, reduced nutrient cycling, increased
March 2001 4-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 dominance of exotic species, and increased dominance by smaller, short-lived opportunistic
2 species (Odum, 1985; Rapport and Whitford, 1999). Discharge of toxic substances into the
3 atmosphere, onto land, and into water can cause acute and chronic stresses and, once the stress is
4 removed, a process of succession begins which can ultimately return the ecosystem to a
5 semblance of its former structure. Air pollution stresses, if acute, are usually short term and the
6 effects soon visible. Chronic stresses, on the other hand, are long-term stresses whose effects
7 occur at different levels of ecosystem organization and appear only after long-term exposures, as
8 in the case of acidic deposition in the northeast or ozone in California (Shortle and Bondietti,
9 1992; U.S. Environmental Protection Agency, 1996b).
10 The possible effects of air pollutants on ecosystems have been categorized by Guderian
11 (1977) as follows:
12 (1) accumulation of pollutants in the plant and other ecosystem components (such as soil
13 and surface-and groundwater),
14 (2) damage to consumers as a result of pollutant accumulation,
15 (3) changes in species diversity because of shifts in competition,
16 (4) disruption of biogeochemical cycles,
17 (5) disruption of stability and reduction in the ability of self-regulation,
18 (6) breakdown of stands and associations, and
19 (7) expanses of denuded zones.
20 How changes in these functions can result from PM deposition and influence ecosystems is
21 discussed in the following text. It should be remembered that, although the effects of PM are
22 being emphasized, the vegetational components of ecosystems also are responding to multiple
23 stresses from other sources.
24
25 4.2.2.3 Ecosystem Response to Direct Plant Effects
26 The presence of PM in the atmosphere may affect vegetation directly, following physical
27 contact with the foliar surface (Section 4.2), but in most cases, the more significant impacts are
28 indirect. These impacts may be mediated by suspended PM (i.e., through effects on radiation and
29 climate) and by particles that pass through the vegetative canopies to the soil. Particulate matter,
30 as considered in this chapter is a heterogeneous mixture of particles differing in size, origin, and
31 chemical constituents and their impacts vary depending on the chemical nature of PM being
March 2001 4-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 deposited on vegetation or soil. Particulate inputs and ecosystem cycling of key elements are
2 considered below.
3 The majority of studies dealing with direct effects of particulate dust and trace metals on
4 vegetation have focused on responses of individual plant species and were conducted in the
5 laboratory or in controlled environments (Saunders and Godzik, 1986). A few have considered
6 the effects of particles on populations, communities, and ecosystems. Most of these focused on
7 ecosystems in industrialized areas heavily polluted by deposits of both chemically inert and
8 active dusts. Effects can result from direct deposition or indirectly by deposition onto the soil.
9 Reductions in growth, yield, flowering, and reproduction of plants from parti culate deposition
10 have been reported (Saunders and Godzik, 1986). Sensitivities of individual species have been
11 associated with changes in composition and structure of natural ecosystems.
12 Evidence from studies of effects of PM deposition, specifically chemically inert and active
13 dusts indicates that, within a population, plants exhibit a wide range of sensitivity, which is the
14 basis for the natural selection of tolerant individuals (Saunders and Godzik, 1986). Rapid
15 evolution of certain populations of tolerant species at sites with heavy trace element and nitrate
16 deposition has been observed. Tolerant individuals present in low frequencies in populations
17 when growing in unpolluted areas have been selected for tolerance at both the seedling and adult
18 stages when exposed to trace metal or nitrate deposition (Ormrod, 1984; U.S. Environmental
19 Protection Agency, 1993). Chronic pollutant injury to a forest community may result in the loss
20 of sensitive species, loss of tree canopy, and maintenance of a residual cover of pollutant-tolerant
21 herbs or shrubs that are recognized as successional species (Table 4-3; Smith, 1974). Frequently,
22 trace metals that penetrate the above-ground plant parts are less injurious than when taken up
23 through the roots (Guderian, 1986).
24 Responses of ecosystems to stresses (unless severe or catastrophic) are difficult to
25 determine because the changes are subtle (Garner, 1991). This is particularly true of responses to
26 particles. Changes in the soil may not be observed until accumulation of the pollutant has
27 occurred for 10 or more years except in the severely polluted areas around heavily industrialized
28 point sources (Saunders and Godzik, 1986). In addition, the presence of other co-occurring
29 pollutants makes it difficult to attribute the effects to PM alone. In other words, the potential for
30 alteration of ecosystem function and structure exists, but it is difficult to quantify, especially
March 2001 4-26 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-3. ECOSYSTEM FUNCTIONS IMPACTED BY AIR POLLUTION
EFFECTS ON TEMPERATE FOREST ECOSYSTEMS
Forest Soil and Vegetation: Activity and Response
Ecosystem Consequence and Impact
1. Forest tree reproduction, alteration, or inhibition
2. Forest nutrient cycling, alteration
a. Reduced litter decomposition
b. Increased plant and soil leaching and soil
weathering
c. Disturbance of microbial symbioses
3. Forest metabolism
a. Decreased photosynthesis
b. Increased respiration
c. Altered carbon allocation
4. Forest stress, alteration
a. Phytophagous insects, increased or decreased
activity
b. Microbial pathogens, increased or decreased
activity
c. Foliar damage increased by direct air pollution
influence
1. Altered species composition
2. Reduced growth, less biomass
3. Reduced growth, less biomass
4. Altered ecosystem stress:
increased or decreased insect infestations;
increased or decreased disease epidemics;
and reduced growth, less biomass, and
altered species composition
Source: Smith (1974).
1
2
3
4
5
9
10
11
12
13
when there are other pollutants present in the ambient air, which 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. 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, 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. Q. prinus dominated the
area, and the larger trees were 60 to 80 years old. The dusty site was dominated by white oak
March 2001
4-27
DRAFT-DO NOT QUOTE OR CITE
-------
1 (Q. alba), whereas Q. rubra and Tulip poplar (Liriodendron tulipifera) were subcodominants.
2 The largest trees were 100 years old and had necrotic leaves, peeling bark, and appeared to be in
3 generally poor condition except for L. tulipifera (which thrived in localized areas). The site
4 contained a tangled growth of seedlings and shrubs, a few saplings, and a prevalence of green
5 briar (Smilax spp.) and grape (Vitis spp). The sapling strata in the area was represented by Acer
6 rubrum, hickory (Carya spp.), dogwood (Cornus florida), and hop-hornbeam (Ostrya
1 virginiana). Saplings of none of the leading dominant trees were of importance in this stratum.
8 The most obvious form of vegetation in the seedling-shrub stratum, because of their tangled
9 appearance, were C. florida, Ostrya virginiana, redbud (Cercis canadensis), and sugar maple
10 (Acer saccarum).
11 Crust formation reduced photosynthesis, induced premature leaf fall, destruction of leaf
12 tissues, inhibited growth of new tissue and reduced the formation of carbohydrate needed for
13 normal growth and storage (Brandt and Rhoades, 1973). The authors (Brandt and Rhoades,
14 1972), citing Odum (1969), also stated that a result of the accumulation of toxic pollutants in the
15 biosphere as the result of human activities, is the simplification of both plant and animal
16 communities. In plant communities, structure is determined by sampling various strata within
17 the community. Each stratum comprises a particular life form (e.g., herbs, seedlings, saplings,
18 trees). Dust accumulation favored growth of some species and limited others. For example,
19 Acer saccharum was more abundant in all strata of the dusty site when compared with the control
20 site where it was present only as a seedling. The growth of L. tulipifera, C. florida,
21 O. virginiana, black haw (Viburnum prunifolium), and C. canadensis appeared to be favored by
22 the dust. Growth of conifers and acidophiles such as rhododendron (Rhododendron maximum),
23 however, was limited. Although dust accumulation began in 1945, the heaviest accumulation
24 occurred between 1967 and 1972 during the time of the study.
25 Changes in community composition were associated closely with changes in the growth of
26 the dominant trees. Decrease in density of seedlings and saplings and in mean basal area, as well
27 as lateral growth of A. rubrum, Q. prinus, and Q. rubra, occurred in all strata. On the other hand,
28 all of these characteristics increased in L. tulipifera, which was a subordinate species before dust
29 accumulation began but had assumed dominance at the time of the study. Reduction in growth of
30 the dominant trees had apparently given L. tulipifera competitive advantage because of its ability
31 to tolerate dust. Changes in soil alkalinity occurred because of the heavy deposition of limestone
March 2001 4-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 dust; however, the facilities necessary for critical analysis of the soils were not available. From
2 the foregoing, it is obvious that PM physical effects in the vicinity of limestone quarries and
3 processing plants can impact ecosystems.
4
5 Acidic Deposition
6 The effects of acidic deposition have been discussed in several previous reports. The 1982
7 EPA document, Air Quality Criteria for Particulate Matter and Sulfur Oxides, devoted a chapter
8 to the effects of acidic deposition (U.S. Environmental Protection Agency, 1982). In 1984, EPA
9 published The Acidic Deposition Phenomenon and Its Effects (Altshuller and Linthurst, 1984),
10 and, in 1991, NAPAP published the result of its extensive study, Acidic Deposition: State of
11 Science and Technology (Irving, 1991). The major effects of acidic deposition occur through the
12 soil and are discussed under indirect effects. However, included among the direct responses of
13 forest trees to acidic deposition are increased leaching of nutrients from foliage; accelerated
14 weathering of leaf cuticular surfaces; increased permeability of leaf surfaces to toxic materials,
15 water, and disease agents; and altered reproductive processes (Altshuller and Linthurst, 1984).
16
17 Trace Elements
18 Possible direct responses of trace elements on vegetation result from their deposition and
19 residence on the phyllosphere (i.e., leaf surfaces). Fungi and other microorganisms living on the
20 leaves of trees and other vegetation play an important role in leaf decomposition after litterfall
21 (Miller and McBride, 1999; Jensen, 1974; Millar, 1974). Early needle senescence and abscission
22 in the San Bernardino Forest changed fungal microflora successional and decomposition patterns
23 by altering the taxonomic diversity and population density of microflora that normally develop
24 on needles while they are on the tree. Changing the fungal community on the needles weakened
25 the decomposer community, decreasing the rate of decomposition, and altered nutrient cycling
26 (Bruhn, 1980). Nutrient availability was influenced by accumulation of carbohydrates and
27 mineral nutrients in the heavy litter under those stands with the most severe needle injury and
28 defoliation (U.S. Environmental Protection Agency, 1996b). Possible impacts of heavy metals
29 on nutrient cycling and their effects on leaf microflora appear not to have been studied.
30 A trace metal must be brought into solution before it can enter into the leaves or bark of
31 vascular plants. Low solubility limits entry. In those instances when trace metals are absorbed,
March 2001 4-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 they frequently are bound in the leaf tissue and then are lost when the leaf drops off (Hughes,
2 1981) and can affect litter decomposition, an important source of soil nutrients. Changes in litter
3 decomposition processes influence nutrient cycling in the soil and limit the supply of essential
4 nutrients. Both Cotrufo et al. (1995) and Niklinska et al. (1998) point out that heavy metals have
5 impacts on forest litter decomposition. Cotrufo et al. (1995) observed that decomposition of oak
6 leaves containing Fe, Zn, Cu, Cr, Ni, and Pb was influenced strongly during the early stages by
7 metal contamination. Fungal mycelium was significantly less abundant in litter and soil in
8 contaminated sites, when compared with control sites. Niklinska et al. (1998) stated that toxic
9 effects of heavy metals on soil respiration rate have been reported by many scientists, and that, in
10 polluted environments, this results in accumulation of undecomposed organic matter. However,
11 they state that results of experiments should identify the most important "natural" factors
12 affecting soil/litter sensitivity because the effects of heavy metals on respiration rates depend on
13 the dose of heavy metals, the type of litter, types of metals deposited, and the storage time before
14 respiration tests are made.
15 Trace metals, particularly heavy metals (e.g., cadmium, copper, lead, chromium, mercury,
16 nickel, zinc) have the greatest potential for influencing forest growth (Smith, 1991).
17 Experimental data indicate that the broadest spectrum of growth suppression of foliar microflora
18 resulted from iron, aluminum, and zinc. These three metals also inhibited spore formation, as did
19 cadmium, chromium, manganese, and nickel (see Smith, 1990e). In the field, the greatest injury
20 occurs from pollution near mining, smelting, and other industrial sources (Ormrod, 1984). Direct
21 metal phytotoxicity can occur only if the metal can move from the surface into the leaf or directly
22 from the soil into the root.
23
24 Organic Compounds
25 Secondary organic compounds formed in the atmosphere, the effects of some of which are
26 discussed below, have been referred to under the following terms: toxic substances, pesticides,
27 hazardous air pollutants (HAPS), air toxics, semivolatile organic compounds (SOCs), and
28 persistent organic pollutants (POPS). Again, it should be noted that the chemical substances
29 denoted by such headings are not criteria air pollutants controlled by the NAAQS under
30 Section 109 of the Clean Air Act (CAA) (U.S. Code, 1991), but rather are controlled under
31 Sect. 112, Hazardous Air Pollutants. Their possible effects on humans and ecosystems are
March 2001 4-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 discussed in a number of government documents and in many other publications. They are
2 mentioned here because, in the atmosphere, many of the chemical compounds are partitioned
3 between gas and particle phases. As particles, they can become airborne, be distributed over a
4 wide area, and impact remote ecosystems. Some of the chemical compounds are of concern
5 because they may reach toxic levels in food chains of both animals and humans, whereas others
6 tend to decrease or maintain the same toxicity as they move through the food chain. Some
7 examples of movement through food chains are provided below.
8 Many chemical compounds from a variety of anthropogenic sources are released into the
9 ambient air (See Section 4.2.1). In the atmosphere, the emitted compounds initially go through a
10 mixing process, and the airborne particles then are distributed over a wide area and ultimately
11 deposited on ecosystem components. Atmospheric deposition of polychlorinated dibenzo-p-
12 dioxins and dibenzofurans (PCDD/Fs), as an example, can be divided into three different forms:
13 (1) dry gaseous, (2) dry particle-bound, and (3) wet deposition. Dry particle-bound deposition
14 occurs when the PM containing the pollutant is deposited on the plant surface, whereas wet
15 deposition ranges from hail through rain to fog and dew fall (McLachlan, 1996b).
16 Human exposure to PCDD/Fs has been demonstrated to be caused almost exclusively by
17 the ingestion of animal fat from fish, meat, and dairy products. Almost half of human exposure
18 to PCDD/Fs is caused by consumption of beef and dairy products (McLachlan, 1996b). Cattle
19 obtain most of their PCCD/Fs though grass. Therefore, the grass-cattle-milk/beef pathway is
20 critical for human exposure. It has been shown that root uptake/translocation is an insignificant
21 pathway of PCDD/Fs to aerial plant parts. Wet and dry particle deposition are the most
22 important for the accumulation of the higher chlorinated cogeners in vegetation. The persistence
23 of PCDD/Fs in plants has not been investigated extensively; however, biodegradation probably
24 does not occur in that these compounds are found primarily in the lipophilic cuticle and are very
25 resistant to microbial degradation (McLachlan, 1996b). Feed contaminated with soil containing
26 the pollutant also can be another source of exposure of beef and dairy cattle as well as chickens.
27 The PCDD/Fs are near a steady state in milk cows and laying hens; however, animals raised for
28 meat production (such as beef cattle and pigs) may accumulate them. The beef cattle and pigs
29 cannot excrete the contaminants in a lipid-rich matrix such as milk or eggs. All of the PCDD/Fs,
30 ingested are stored in the body. In agricultural food chains, there is a biodilution of PCDD/Fs,
31 with the fugacity decreasing by up to three orders of magnitude between the air and cows milk
March 2001 4-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 (McLachlan, 1996b). Fiirst et al. (1993), based on surveys to determine the factors that influence
2 the presence of PCDD/PCDF in cows milk, earlier concluded that regardless of which pathway,
3 soil - grass - cow or air - grass - cow, it was the congener of the chemical that was most
4 important.
5 Persistent polychlorinated pollutants (POPS), such as PCBs, PCDFs, and PCDDs, can be
6 transported as particles through the atmosphere from industrial and agricultural sources; be
7 brought down via wet and dry deposition in remote regions, such as the Arctic; and have been
8 detected in all levels of the Arctic food chain (Oehme et al., 1995). High concentrations of PCB
9 (1 to 10 ppm) were found in seals; but the concentrations increased to 10 to 100 ppm in polar
10 bears. The polar bear is the top predator in the Arctic and feeds preferentially on ringed seals and
11 also, to a lesser extent, on other seal species. Bioconcentration factors of organochlorines in the
12 Arctic food web, reaching 107 for fish and seals, are biomagnified in polar bears (Oehme et al.,
13 1995). Polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans
14 (PDCF/s) also have also been found in seals (Oehme et al., 1995). Milk taken from
15 anaesthetized polar bears was also found to contain PCDD/PCDF. Very little is known regarding
16 the intake of milk by polar bear cubs. However, estimates of the intake of milk containing
17 detectable levels of PCDD/PCDF and PCB and the additional consumption of seal blubber
18 confirm that these pollutants are passed on to the next generation (Oehme et al., 1995).
19 Section 112 of the CAA, provides the legislative basis for U.S. hazardous air pollutant
20 (HAP) programs. In response to mounting evidence that air pollution contributes to water
21 pollution, Congress included Section 112m (Atmospheric Deposition to Great Lakes and Coastal
22 Waters) in the 1990 CAA Amendments, which directs the Environmental Protection Agency
23 (EPA) to establish a research program on atmospheric deposition of HAPS to the "Great
24 Waters".
25 Actions taken by EPA and others to evaluate and control sources of Great Waters pollutants
26 of concern appear to have positively affected trends in pollutant concentrations measured in air,
27 sediment, and biota. Details concerning these effects may be found in "Deposition of Air
28 Pollutants to the Great Waters", Third Report to Congress (U. S. Environmental Protection
29 Agency, 2000a). The Third Report (EPA-453/R-00-005, June 2000), like the First and Second
30 Reports to Congress, focuses on 15 pollutants of concern, including pesticides, metal
31 compounds, chlorinated organic compounds, and nitrogen compounds. The new scientific
March 2001 4-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 information in the Third Report supports and builds on three broad conclusions presented in the
2 previous two EPA Reports to Congress and discussed below.
3 (1) Atmospheric deposition from human activities can be a significant contributor of toxic
4 chemicals and nitrogen compounds to the Great Waters. The relative importance of
5 atmospheric loading for a particular chemical in a water body depends on many factors (e.g.,
6 characteristics of the water body, properties of the chemical, and the kind and amount of
7 atmospheric deposition versus or water discharges).
8 (2) A plausible link exists between emissions into the air of Great Waters toxic pollutants of
9 concern; the atmospheric deposition of these pollutants (and their transformation products);
10 and the concentrations of these pollutants found in the water, sediments, and biota, especially
11 fish and shellfish. For mercury, fate and transport modeling and exposure assessments
12 predict that the anthropogenic contribution to the total amount of methylmercury in fish is, in
13 part, the result of anthropogenic mercury releases from industrial and combustion sources
14 increasing mercury body burdens (i.e., concentrations) in fish. Also, the consumption offish
15 is the dominant pathway of exposure to methylmercury for fish-consuming humans and
16 wildlife. However, what is known about each stage of this process varies with each pollutant
17 (for instance, the chemical species of the emissions and its transformation in the
18 atmosphere).
19 (3) Airborne emissions from local as well as distant sources, from both within and outside the
20 United States, contribute pollutant loadings to waters through atmospheric deposition.
21 Determining the relative roles of particular sources—local, regional, national, and possibly
22 global, as well as anthropogenic, natural, and reemission of pollutants—contributing to
23 specific water bodies is complex, requiring careful monitoring, atmospheric modeling, and
24 other analytical techniques.
25
26 4.2.2.4 Indirect Effects of Particulate Matter In Ecosystems
27 The presence of PM in the atmosphere directly affects vegetation following physical
28 contact with foliar surfaces (as discussed above in Section 4.2.2.2), but in many cases the more
29 significant impacts are indirect. These impacts may be mediated by suspended PM (i.e., through
30 effects on radiation and climate) and by particles that pass through vegetative canopies to reach
31 the soil. Effects mediated in the atmosphere are considered briefly below and in greater detail
March 2001 4-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 later, under Section 4.5. Indirect plant responses are chiefly soil mediated and depend primarily
2 on the chemical composition of the individual elements deposited in PM. The individual
3 elements must be bioavailable to have an effect. The soil environment, composed of mineral and
4 organic matter, water, air, and a vast array of bacteria, fungi, algae, actinomycetes, protozoa,
5 nematodes, and arthropods, is one of the most dynamic sites of biological interactions in nature
6 (Wall and Moore, 1999; Alexander, 1977). The quantity of organisms in soils varies by locality.
7 Bacteria and fungi are usually most abundant in the rhizosphere, the soil around plant roots that
8 all mineral nutrients must pass through. Bacteria and fungi benefit from the nutrients in the root
9 exudates (chiefly sugars) in the soil and, in turn, they play an essential role by making mineral
10 nutrients available for plant uptake (Wall and Moore, 1999; Rovira and Davey, 1974). Their
11 activities create chemical and biological changes in the rhizosphere by decomposing organic
12 matter and making inorganic minerals available for plant uptake. Bacteria are essential in the
13 nitrogen and sulfur cycles and make these elements available for plant uptake and growth (see
14 Section 4.3.3). Fungi are directly essential to plant growth. Attracted to the roots by the
15 exudates, they develop mycorrhizae, a mutualistic, symbiotic relationship, that is integral in the
16 uptake of the mineral nutrients (Allen, 1991). The impact in ecosystems of PM, particularly
17 nitrates, sulfates, and metals, is determined by their affect on the growth of the bacteria involved
18 in nutrient cycling and the fungi involved in plant nutrient uptake.
19
20 Particulate Matter-Related Atmospheric Turbidity: Effects on Vegetative Processes
21 Photosynthetic processes underlie the contribution of vegetative surfaces to nutrient and
22 energy cycling. Photosynthesis and the heat-driven processes of water cycling depend on net
23 receipts and characteristics of the radiation environment. These characteristics may be altered
24 substantially when the atmosphere becomes turbid because of particulate loading.
25 Specific wavelengths of interest depend on the vegetation process under consideration.
26 Canopy temperature and water relations are particularly sensitive to long-wave, infrared
27 radiation, whereas primary photosynthetic charge separations depend on short-wave radiation in
28 the visible and photosynthetically active range (0.4 to 0.7 //m).
29 Effects of anthropogenic aerosols on the radiation environment at the Earth's surface are
30 difficult to assess. The residence time of suspended particles varies with size and environmental
31 conditions (seconds to months or years), and concentrations are spatially and temporally variable.
March 2001 4-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 In particularly polluted urban and near-urban areas, unambiguous particulate impacts on radiation
2 and local climate may be observed. Visibility was degraded by 50% in a large plume originating
3 in the St. Louis urban area during the midweek, midday period (Pueschel, 1993). In contrast,
4 visibility was reduced by only 20% on weekends, when traffic and industrial emissions were
5 reduced. The area affected by this plume includes highly productive agricultural land.
6 Empirical relationships between mass of specific components of the aerosol and radiation
7 scattering have been developed (e.g., Pueschel, 1993), from which regional visibility (or
8 radiation attenuation) isopleths can be constructed if appropriate mass data are available. These
9 estimates support trends observed by direct measurement of turbidity (e.g., Flowers et al., 1969;
10 U.S. Environmental Protection Agency, 1982).
11 Sulfates, nitrates, and elemental carbon dominate effects on visibility, in part, because they
12 frequently dominate the mass profiles and, in part, because they exhibit particularly large
13 absorption coefficients (see Section 4.3). Absorption by particles containing carbon may range
14 from 5 to 10% in rural areas to up to 50% in urban areas (U.S. Environmental Protection Agency,
15 1982). In west-coast cities with contrasting particulate sources and loadings, the common
16 component that related PM to visibility degradation was sulfate between 0.65 and 3.6 //m
17 (Barone et al., 1978). For example, in Los Angeles, sulfate and nitrate had similar effects on
18 visibility (White, 1976), despite the dominance of nitrate from transportation sources in the
19 aerosol, although this is changing with controls on point sources of sulfate (Farber et al., 1994).
20 No long-term global trend of increasing atmospheric optical depth has been documented
21 (Bolle et al., 1986; Pueschel, 1993), although seasonal and regional impacts are substantial. The
22 classic study by Flowers et al. (1969) demonstrated large regional distinctions in turbidity across
23 the United States. Typically, the western deserts, plains, and Rocky Mountains exhibited low
24 mean annual turbidity, whereas the more humid and densely vegetated eastern half of the country
25 exhibited much greater turbidities. In the mid-1970s, visible range in the mountainous southwest
26 exceeded 110 km and radiation attenuation was ca. 2.6%; whereas, in the east, visible range was
27 below 24 km and radiation attenuation was ca. 10%. Visibility in the eastern United States has
28 decreased generally since the 1940s (Flowers et al., 1969; Trijonis and Shapland, 1979; U.S.
29 Environmental Protection Agency, 1982). Correlative trends in visibility degradation and
30 emissions of sulfur oxides suggest that particulate sulfate may account for much of the turbidity.
March 2001 4-3 5 DRAFT-DO NOT QUOTE OR CITE
-------
1 These trends are typical of urban industrial areas around the world. Turbidity has increased
2 above Mexico City (Binenko and Harshvardhan, 1993) since the 1911 to 1928 period. During
3 this early period, a single annual peak of turbidity coincided with the end of the dry period, and
4 natural sources dominated. By 1957 to 1962, the number of annual peaks had increased, as
5 anthropogenic sources came to dominate. During this period, atmospheric transmission of direct-
6 beam solar radiation decreased by about 10% (Binenko and Harshvardhan, 1993). Visibility in
7 the Los Angeles basin has improved very slightly in the past decades (Farber et al., 1994), as
8 sulfate emissions have been controlled by regulation. The composition of the aerosol has
9 changed, particularly in inland areas, as the former dominance of sulfate shifts to a
10 preponderance of secondary organics.
11 Particles interact with solar radiation through scattering and absorption. Absorption of
12 short-wavelength solar radiation reduces the amount of radiation reaching the Earth's surface and
13 leads to atmospheric heating. If the absorbing particles reradiate in the infrared range, then some
14 of this energy is lost as long-wave reradiation to space. This loss mechanism is minimized
15 because most of the anthropogenic aerosol in the troposphere resides in the planetary boundary
16 layer (Bolle et al., 1986), even within the lower 500 m (Binenko and Harshvardhan, 1993), where
17 the temperature is similar to that of the surface. Some of this energy is captured at the surface as
18 down-welling infrared radiation.
19 These wavelengths directly impact canopy temperatures and influence transpirational water
20 use by vegetation. The presence of absorbing aerosols reduces the ratio of photosynthetically
21 active radiation to total radiation received at the surface, potentially reducing photosynthetic
22 water use efficiency. The net effect of aerosol absorption on the surface depends on the relative
23 magnitudes of the particulate absorption coefficients in the visible and infrared area and on the
24 albedo of the Earth's surface. In general, absorption is not a dominant parti culate effect.
25 Scattering of radiation dominates the effects of parti culate loading on visibility and
26 turbidity. Nonabsorbing, scattering aerosols raise the overall albedo of the atmosphere and
27 reduce the amount of radiation reaching the surface by the amount reflected or backscattered to
28 space. As atmospheric turbidity increases, so does the scattering of light, including forward
29 scattering of photosynthetically active radiation that intercepts the Earth's surface (Hoyt, 1978).
30 The largest effect is described by Mie-scattering theory. Forward scattering reduces the
31 intensity of direct radiation by disrupting the solar beam, thereby increasing the path length and
March 2001 4-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 probability of absorption and also increases the intensity of diffuse (sky) radiation. In a clear
2 atmosphere, diffuse radiation may be on the order of 10% of total solar radiation (Choudhury,
3 1987). However, in highly turbid, humid conditions, this fraction may increase, even up to 100%
4 of solar radiation in extreme cases. The direct-to-diffuse-radiation ratio is highest at solar noon
5 and lowest near dawn or dusk, when the path length through the atmosphere is longest.
6 Particle scattering is wavelength dependent, causing objects to appear blue- or red- tinged,
7 depending on viewing and illumination angles and on the light quality, the alteration of which is
8 a minor contributor to photosynthetic light-use efficiency. The wavelength dependence of
9 scattering decreases rapidly from extreme sensitivity for very fine particles to little dependence at
10 10 (j,m. Equations relating scattering at a reference wavelength to scattering at wavelengths of
11 interest are rigorously applicable only to spherical particles but may be extended to nonspherical
12 particles of equal volume (Janzen, 1980).
13 World Meteorological Organization (WHO) data summarized in U.S. Environmental
14 Protection Agency (1982) indicated that turbidity in the eastern United States commonly resulted
15 in radiation losses of ca. 3.5% because of backscattered radiation and ca. 3.5% because of
16 absorption, with a resulting total reduction of incident radiation to ca. 93% of total solar
17 radiation. However, 28% of the radiation reaching the surface was converted from direct
18 radiation to diffuse, or sky, radiation. Under more polluted conditions, losses were ca. 9%
19 backscattered and 9% absorbed, reducing total radiation to 82% of total solar radiation and
20 converting 72% from direct beam to diffuse radiation. Photosynthetically active radiation (0.4 to
21 0.7 //m) typically is enriched in diffuse radiation relative to total or direct beam radiation.
22
23 Altered Radiative Flux: Effects on Vegetative Processes
24 Canopy photosynthesis is typically a nearly linear function of incident radiation,
25 overcoming saturation exhibited by individual leaves by distributing the light throughout the
26 multilayer canopy. Light penetration into canopies limits photosynthetic productivity (Rosenberg
27 et al., 1983). The uppermost leaves of many canopies are at or above light saturation for
28 photosynthetic processes. The simplest radiative transfer functions describing plant canopies
29 relate total down-welling radiation (direct plus diffuse radiation measured above the canopy) to
30 radiation interception at each leaf level through a Beer's Law analogy. The expected exponential
31 decline in radiation through the canopy depends only on total radiation and a bulk canopy
March 2001 4-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 extinction coefficient that depends on leaf size, orientation, and distribution, as well as on
2 reflectance and absorption in wavelengths of interest. These simplified models predict radiation
3 distribution adequately for homogeneous canopies. Turbidity affects canopy processes only by
4 attenuating the total radiation impinging on the canopy surface.
5 In more complex, and more realistic, canopy-response models (e.g., Choudhury, 1987),
6 radiation is considered in its direct and diffuse components. Foliar interception by canopy
7 elements is considered for both up- and down-welling radiation (a two-stream approximation).
8 In this case, the effect of atmospheric PM on turbidity affects canopy processes both by radiation
9 attenuation and by influencing the efficiency of radiation interception throughout the canopy
10 through conversion of direct to diffuse radiation (Hoyt, 1978). Diffuse radiation is more
11 uniformly distributed throughout the canopy and increases canopy photosynthetic productivity by
12 distributing radiation to lower leaves. The treatment of down-welling direct-beam radiation in
13 the two-stream approach remains an elaboration of the simplified Beer's Law analogy, with solar
14 angle, leaf area distribution, and orientation individually parameterized (Choudhury, 1987).
15 Diffuse down-welling radiation is a function of diffuse and direct radiation at the top of the
16 canopy and penetration within the canopy, according to cumulative leaf area density and foliage
17 orientation. Up-welling (diffuse) radiation results from scattering and reflectance within the
18 canopy, and by the soil, of both direct and diffuse down-welling radiation.
19 The effect of the altered distribution between diffuse and direct radiation impacts
20 photosynthesis in upper, exposed leaves as a function of leaf angle and in total canopy
21 photosynthesis as a function of penetration of radiation within the canopy. This depends on
22 canopy structure, leaf optical properties, and leaf area density, as well as on solar angle and
23 atmospheric turbidity. Absorption of radiation by particles heats the upper atmosphere and
24 results in reduced vertical temperature gradients. This could reduce the intensity of atmospheric
25 turbulent mixing. The magnitude of such potential effects on turbulent transport within canopies
26 remains unknown, although damping of eddy transport could inhibit canopy gas exchange.
27 Suppressed tropospheric mixing also could intensify local temperature inversions and increase
28 the severity of pollution episodes (Pueschel, 1993), with direct inhibitory effects on
29 photosynthetic processes.
30 The most significant effect of aerosols on vegetation is probably through their role as cloud
31 condensation nuclei because clouds have substantial impact on radiation receipts at the surface.
March 2001 4-3 8 DRAFT-DO NOT QUOTE OR CITE
-------
1 An important characteristic of fine particles is their ability to affect the flux of solar radiation
2 passing through the atmosphere directly, by scattering and absorbing solar radiation, and
3 indirectly, by acting as cloud condensation nuclei which in turn influence the optical properties
4 of clouds (Chameides et al., 1999). Regional haze has been estimated to diminish surface solar
5 visible radiation by approximately 8%. Crop yields have been reported as being sensitive to the
6 amount of sunlight received. The potentially significant impact of regional haze on the yield of
7 crops because of reduction in solar radiation has been examined by Chameides et al. (1999).
8 Using a case study approach, Chameides et al. (1999), studied the affects of regional haze on
9 crop production in China, where regional haze is especially severe. A rudimentary assessment of
10 the direct effect of atmospheric aerosols on agriculture suggests that optimal crop yields of
11 approximately 70% of the crops are being depressed by at least 5 to 3% by regional scale air
12 pollution and its associated haze (Chameides et al., 1999).
13
14 Effects of Solar Ultraviolet Radiation on Terrestrial Ecosystems
15 The transmission of solar UV-B radiation through the earth's atmosphere is controlled by
16 ozone, clouds and particles. The depletion of stratospheric ozone caused by the release of
17 chlorofluorcarbons (CFCs) and other substances, such as halides, has resulted in heightened
18 concern about potentially deleterious increases in the amount of solar UV-B (SUVB) radiation
19 reaching the Earth's surface (see Section 4.5). One salient consideration is that, although CFC
20 production is at a peak level now, the problem likely will continue well into the future because of
21 the length of time it takes for molecules to reach the stratosphere (Greenberg et al, 1997).
22 The vulnerability of terrestrial plants to UV-B results from their requirement for sunlight
23 for photosynthesis. Each 1% decline in ozone has been predicted to decrease crops yield by 1%
24 (Greenberg et al., 1997). In addition to inhibiting photosynthesis, UV-B radiation triggers
25 numerous responses in plants (e.g., membrane, protein, and DNA damage; delayed maturation;
26 diminished growth; activation of chemical stress; flavonoid synthesis; leaf thickening)
27 (Table 4-4). It is not known which of the injury and damage effects are most detrimental to plant
28 growth (Table 4-4). Effects of increased UV-B on plant growth are likely to be incremental.
29 Because plants evolved under the selective pressure of ambient UV-B radiation in sunlight, they
30 have developed adaptive mechanisms (Greenberg et al., 1997). Although inhibition of
31 photosynthesis is a detrimental growth effect, flavonoid synthesis represents acclimation.
March 2001 4-39 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-4. TYPES OF PLANT RESPONSES TO ULTRAVIOLET-B RADIATION3
Acclimation and Morphological Responses Damage and Injury Responses
Altered biomass distribution
Altered leaf cell division
Cotyledon curling
Increased DNA repair
Increased flavonoid biosynthesis
Increased leaf thickness
Increased leaf number
Increased number of tillars
Leaf wrinkling
Reduced leaf area
Reduced hypocotyl growth
Reduced shoot height
Reduced stomatal density
Altered gene expression
Degradation of auxin
Degradation of chlorophyll and carotenoids
Degradation of proteins
Diminished biomass
Epidermal collapse
Inhibition of growth
Inhibition of photosynthesis
Increased stomatal conductance
Lower seed yield
Oxidation of DNA
Peroxidation of lipids
Prymidine dimer formation
"Entries in alphabetical order.
1 Plants growing under full light have been shown to be protected against UV-B effects but not
2 when growing under weak visible light (Bjorn, 1996). A common adaptation is alteration in leaf
3 transmission properties, which results in attenuation of UV-B in the epidermis before it can reach
4 the leaf interior.
5 Plant species vary enormously in their response to UV-B exposures, and large differences
6 in response occur among different genotypes within a species. In general, dicotyledonous plants
7 are more sensitive than monocotyledons from similar environments. In addition, plant responses
8 may differ depending on stage of development. Therefore, extrapolation of experimental
9 responses from seedlings to mature plants must be taken with caution (Bjorn, 1996). The above
10 facts are especially important when considering the effects of UV-B on agricultural plants.
11 For example, among soybeans and rice, there are varieties for which growth and crop yield are
March 2001
4-40
DRAFT-DO NOT QUOTE OR CITE
-------
1 severely decreased by increased UV-B radiation and other varieties that are not affected or may
2 even be stimulated. On the other hand, the growth of the same sensitive soybeans when grown
3 under water stress was not inhibited. Many crop plants grown in temperate regions originated in
4 more tropical areas, hence, a gene pool for more resistant varieties is likely to exist (Bjorn, 1996).
5 Crop plants, unlike forest trees and vegetation in natural ecosystems, are only exposed for one
6 generation, and thus, it may be possible to readily change the genotype if a variety proves to be
7 sensitive.
8 Trees, forests, and perennial evergreen plants are long-lived when compared to agricultural
9 systems, making it possible for UV-B exposure impacts to accumulate with time. Saplings and
10 young and small trees react differently when compared to mature trees; also, on evergreen trees,
11 needles of different ages respond differently (Bjorn, 1996). Breeding and testing trees is a slow
12 process, and, for this reason, much care needs to be taken when planting large areas with trees of
13 a single species and one provenance (e.g., Stika Spruce [Picea sitchensis] in Britain). The
14 response of only a few broad-leaved trees have been studied. The most investigated genus has
15 been loblolly pine (Pinus taeda) (Bjorn, 1996).
16 A few studies indicate that the photomorphogenesis (changes in leaf thickness under UV-B
17 that results in a transition from shade to sun leaves, Table 4-4) and the variable responses of
18 native plants in ecosystems to UV-B exposures results in changes in interactions between various
19 plants species, changes between plants and other organisms, and between plants and their abiotic
20 environment. These preliminary studies suggest that in natural ecosystems, composed of many
21 different plant species, with complex interactions between plants and between plants and other
22 organisms, there may develop effects of UV-B that cannot be determined from experiments on
23 single plant species. The effects of UV-B on natural plant systems, therefore, should be of
24 greater concern than on agricultural crops (Bjorn, 1996).
25
26 Nitrogen Deposition Effects
27 Nitrogen has long been recognized as the nutrient most important for plant growth. Plants
28 usually absorb nitrogen through their roots by absorbing NH4 + or NO3" or informed by symbiotic
29 organisms in the roots. Plants, however, vary in their ability to absorb ammonium and nitrate
30 (Chapin et al., 1987). Nitrogen is of overriding importance in plant metabolism and, to a large
31 extent, governs the utilization of phosphorus, potassium, and other nutrients. Most of the
March 2001 4-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 nitrogen in soils is associated with organic matter. Typically, the availability of nitrogen via the
2 nitrogen cycle controls net primary productivity and possibly the decomposition rate of plant
3 litter. Photosynthesis is influenced by nitrogen uptake in that ca. 75% of the nitrogen in a plant
4 leaf is used during the process of photosynthesis. The nitrogen-photosynthesis relationship is,
5 therefore, critical to the growth of trees and other plants (Chapin et al., 1987).
6 Because nitrogen is not readily available and is usually in shortest supply, it is the chief
7 element in agricultural fertilizers. Atmospherically deposited nitrogen also can act as a fertilizer
8 in soil low in nitrogen. Not all plants, however, are capable of utilizing extra nitrogen. Inputs of
9 nitrogen to natural ecosystems that alleviate deficiencies and increase growth of some plants can
10 impact competitive relationships and alter species composition and diversity (Ellenberg, 1987;
11 Kenk and Fischer, 1988; U.S. Environmental Protection Agency, 1993).
12 The impact of increasing nitrogen inputs (e.g., NOX, nitrates, nitric acid) on the nitrogen
13 cycle and forests, wetlands, and aquatic ecosystems is discussed in detail elsewhere (U.S.
14 Environmental Protection Agency, 1993, 1997a; Garner, 1994; World Health Organization,
15 1997). The most important effects of nitrogen deposition are accumulation of nitrogen
16 compounds resulting in the enhanced availability of nitrate or ammonium, soil-mediated effects
17 of acidification, and increased susceptibility to stress factors (Bobbink et al., 1998). A major
18 concern is "nitrogen saturation", the result of the deposition of large amounts of particulate
19 nitrates. Nitrogen saturation results when additions to soil background nitrogen (nitrogen
20 loading) exceed the capacity of plants and soil microorganisms to utilize and retain nitrogen
21 (Aber et al., 1989, 1998; Garner, 1994; U.S. Environmental Protection Agency, 1993). Under
22 these circumstances, ecosystems become unable to utilize excessive nitrogen inputs and
23 disruptions of ecosystem functioning may result (Hornung and Langan, 1999).
24 Growth of most forests in North America is limited by the nitrogen supply. Severe
25 symptoms of nitrogen saturation, however, have been observed in high-elevation, nonaggrading
26 spruce-fir ecosystems in the Appalachian Mountains, as well as in the eastern hardwood
27 watersheds at Fernow Experimental Forest near Parsons, WV. Mixed conifer forests and
28 chaparral watersheds with high smog exposure in the Los Angeles Air Basin also are nitrogen
29 saturated and exhibit the highest stream water NO3" concentrations for wildlands in North
30 America (Bytnerowicz and Fenn, 1996; Fenn et al., 1998). Not all forest ecosystems react in the
31 same manner to nitrogen deposition. High-elevation alpine watersheds in the Colorado Front
March 2001 4-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 Range and a deciduous forest in Ontario, Canada, also are naturally saturated even though
2 nitrogen deposition has been moderate («8 kg.ha ha'Vyear"1). The Harvard Forest hardwood
3 stand in Massachusetts, however, has absorbed >900 kg N/ha without significant NO3" leaching
4 during a nitrogen amendment study of 8 years (Table 4-5; Fenn et al., 1998). Johnson et al.
5 (199 la) reported that measurements showing the leaching of nitrates and aluminum (A13+) from
6 high elevation forests in the Great Smoky Mountains indicates that these forests have reached
7 saturation.
8 Possible ecosystem responses to nitrate saturation, as postulated by Aber and coworkers
9 (Aber et al., 1989), include a permanent increase in foliar nitrogen and reduced foliar phosphorus
10 and lignin caused by the lower availability of carbon, phosphorus, and water; reduced
11 productivity in conifer stands because of disruptions of physiological function; decreased root
12 biomass and increased nitrification and nitrate leaching; and (4) reduced soil fertility, resulting
13 from increased cation leaching, increased nitrate and aluminum concentrations in streams, and
14 decreased water quality. Saturation implies that some resource other than nitrogen is limiting
15 biotic function.
16 Water and phosphorus for plants and carbon for microorganisms are the resources most
17 likely to be the secondary limiting factors. The appearance of nitrogen in soil solution is an early
18 symptom of excess nitrogen. In the final stage, disruption of forest structure becomes visible
19 (Garner, 1994).
20 Changes in nitrogen supply can have a considerable impact on an ecosystem's nutrient
21 balance (Waring, 1987). Large chronic additions of nitrogen influence normal nutrient cycling
22 and alter many plant and soil processes involved in nitrogen cycling (Aber et al., 1989).
23 Among the processes affected are (1) plant uptake and allocation, (2) litter production,
24 (3) immobilization (includes ammonification [the release of ammonia] and nitrificatrion
25 [conversion of ammonia to nitrate during decay of litter and soil organic matter]), and (4) nitrate
26 leaching and trace gas emissions (Figure 4-2; Aber et al., 1989).
27 Subsequent studies have shown that, although initially, there was an increase in nitrogen
28 mineralization (i.e., the conversion of soil organic matter to nitrogen in available form [see item
29 3 above]), nitrogen mineralization rates were reduced under nitrogen-enriched conditions. Also,
30 studies suggest that, during saturation, soil microbial communities change from predominantly
31 fungal (mycorrhizal) communities to those dominated by bacteria (Aber et al., 1998).
March 2001 4-43 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-5. NITROGEN-SATURATED FORESTS IN NORTH AMERICA, INCLUDING
ESTIMATED N INPUTS AND OUTPUTS
Location
Forest Type
Elevation N Input N Output
(m) (kg-ha^-year1) (kg-ha'-year1)
Reference
Adirondack Mts.
northeastern New York
Catskill Mts.,
southeastern New York
Turkey Lakes Watershed,
Ontario, Canada
Northern hardwoods or
hardwood/conifer mix
Mainly hardwood;
some eastern hemlock
Sugar maple and yellow
birch
396-661
335-675
350-400
9.3a
10.2a
7.0-7.7 (as
throughfall)
Stage 1 N
loss"
Stage 1 and 2
N lossb
17.9-23.6
Driscoll and
Van Dreason (1993)
Stoddard(1994)
Foster etal. (1989)
Johnson and
Whitetop Mt,
southwestern Virginia
Femow, West Virginia
Great Smoky Mts.
National Park, Tennessee
Great Smoky Mts.
National Park, Becking
Site, North Carolina
Great Smokey Mts.
National Park, Tower
Site, North Carolina
Red spruce
Mixed hardwood
American beech
Red spruce
Red spruce
1650
32C
735-870 15-20
1600
1800
1740
3.1d
10.3d
26.6
47C
6.1
2.9
19.2
20.3
Lindberg(1992a)
Joslin and Wolfe
(1992)
Joslin etal. (1992)
Gilliametal. (1996)
Peterjohnetal. (1996)
Johnson and Lindberg
(1992b)
Johnson etal. (1991a)
Johnson etal. (1991a)
Front Range, Colorado
San Dimas, San Gabriel
Mts. southern California
Camp Paivika,
San Bernadino Mts.,
southern California
Klamath Mts,
northern California
Thompson Forest, Cascade
Mts., Washington
Alpine tundra,
subalpine conifer
Chapparral and
grasslands
Mixed conifer
Western coniferous
Red alder
3000-4000
580-1080
1600
NA
220
7.5-8.0
23. 3e
30
Mainly
geologic8
4. 7 plus > 100
as N2 fixation
7.5
0.04-19.4
7-26f
NA8
38.9
Williams etal. (1996)
Riggan etal. (1985)
Fenn etal. (1996)
Dahlgren(1994)
Johnson and Lindberg
(1992b)
"Estimated total N deposition from wet deposition data is from Driscoll et al. (1991) for the Adirondacks, and from Stoddard and
Murdoch (1991) for the Catskills. Total deposition was estimated based on the wet deposition/total N deposition ratio (0.56) at
Huntington Forest in the Adirondacks (Johnson and Lindberg, 1992b). Nitrogen deposition can be higher in some areas, especially
at high-elevation sites such as Whiteface Mountain (15.9 kg-ha^-year"1; Johnson and Lindberg, 1992b).
bStage 1 and 2 of N loss according to the watershed conceptual model of Stoddard (1994). Nitrogen discharge (kg-ha^-year"1) data
are not available; only stream water NO3" concentration trend data were collected.
'Values appear high compared to other sites, especially N leaching losses. Joslin and Wolfe (1992) concede that "there is
considerable uncertainty associated with the estimates of atmospheric deposition and leaching fluxes." However, elevated NO3"
concentrations in soil solution, and lack of a growth response to N fertilization (Joslin and Wolfe, 1994) support the hypothesis that
the forest at Whitetop Mountain is N saturated.
dEstimated total N deposition from throughfall data. Total deposition was estimated based on the throughfall/total N deposition
ration (0.56) from the nearby Smokies Tower site (Johnson and Lindberg, 1992b).
"Annual throughfall deposition to the chaparral ecosystem.
'Nitrogen output is from unpublished streamwater data (M.E. Fenn and M. A. Poth). The low value represents a year of average
precipitation, and the high value is for 1995, when precipitation was nearly double the long-term average. Nitrogen output
includes N export in streamwater and to groundwater.
8Annual input and output data are not known, although N deposition in this forest is probably typical for much of the rural western
United States (2-3 kg N-ha^-year"1 (Young et al., 1988). Excess N is from weathering of ammonium in mica schist bedrock. The
ammonium was rapidly nitrified, leading to high NO3" concentrations in soil solution (Dahlgren, 1994).
March 2001
4-44
DRAFT-DO NOT QUOTE OR CITE
-------
Microbial
Decomposition
— Process altered by
nitrogen saturation
Figure 4-2. Nitrogen cycle (dotted lines indicate processes altered by nitrogen satuation).
Source: Garner (1994).
1 The availability of nutrients is an important factor in determining species composition, and
2 nitrogen is usually the growth-limiting nutrient (Bobbink, 1998). Most of the plants growing in
3 nutrient-poor habitats have become adapted to them over time and can only compete successfully
March 2001
4-45
DRAFT-DO NOT QUOTE OR CITE
-------
1 on soils low in nitrogen (Bobbink, 1998; Chapin, 1991). All plants growing in low resource
2 environments (e.g., infertile soil, shaded understory, deserts, tundra) have been observed to have
3 certain similar characteristics: a slow growth rate, low photosynthetic rate, and low capacity for
4 nutrient uptake. An important feature to plants adapted to low-resource environments is that they
5 grow slowly and tend to respond less, even when provided with an optimal supply and balance of
6 resources (Pearcy et al., 1987; Chapin, 1991). Plants adapted to cold, moist environments grow
7 more leaves than roots as the relative availability to nitrogen increases; however, other nutrients
8 may soon be limiting. The capacity of gymnosperms in general, and in subalpine and boreal
9 species in particular, to reduce nitrates in either roots or leaves appears to be limited. In addition,
10 the ability of trees to use nitrogen varies with the age of the tree and the density of the stand
11 (Waring, 1987).
12 Because the competitive equilibrium of plants in any community is finely balanced, the
13 alteration of one of a number of environmental parameters, (e.g., continued nitrogen additions),
14 can change the vegetation structure of an ecosystem (Bobbink, 1998; Skeffmgton and Wilson,
15 1988). Increases in soil nitrogen play a selective role. When nitrogen becomes more readily
16 available, plants adapted to living in an environment of low nitrogen availability will be replaced
17 by plants capable of using increased nitrogen because they have a competitive advantage.
18 The long-term impacts of increased nitrogen deposition have been studied in several
19 western and central European plant communities: lowland heaths, species-rich grasslands,
20 mesotrophic fens, ombrotrophic bogs, upland moors, forest-floor vegetation, and freshwater
21 lakes (Bobbink, 1998). Large changes in species composition have been observed in regions
22 with high nitrogen loadings or infield experiments after years of nitrogen addition (Bobbink
23 etal., 1998). The increased input of nitrogen gradually increased availability of nitrogen in the
24 soil, and its retention because of low rates of leaching and denitrification resulted in faster litter
25 decomposition and rate of mineralization. Faster growth and greater height of nitrophilic species
26 enables these plants to shade out the slower growing species, particularly those in oligotrophic or
27 mesotrophic conditions (Bobbink, 1998; Bobbink et al., 1998). Excess nitrogen inputs to
28 unmanaged heathlands in the Netherlands has resulted in nitrophilous grass species replacing
29 slower growing heath species (Roelofs et al., 1987; Garner, 1994). Van Breemen and Van Dijk
30 (1988) noted that over the past several decades the composition of plants in the forest herb layers
March 2001 4-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 has been shifting toward species commonly found on nitrogen-rich areas. It also was observed
2 that the fruiting bodies of mycorrhizal fungi had decreased in number.
3 Other studies in Europe point out the effects of excessive nitrogen deposition on mixed-oak
4 forest vegetation along a deposition gradient largely controlled by soil acidity, nitrogen supply,
5 canopy composition, and location of sample plots (Brunet et al., 1998; Falkengren-Grerup,
6 1998). Results of the study, using multivariate methods, suggest that nitrogen deposition has
7 affected the field-layer vegetation directly by increased nitrogen availability and, indirectly, by
8 accelerating soil acidity. Time series studies indicate that 20 of the 30 field-layer species
9 (nonwoody plants) that were associated most closely with high nitrogen deposition increased in
10 frequency in areas with high nitrogen deposition during the past decades. Included in the field-
11 layer species were many generally considered nitrophilous; however, there were several acid
12 tolerant species (Brunet et al, 1998). Falkengren-Grerup (1998), in an experimental study
13 involving 15 herbs and 13 grasses, observed that species with a high nitrogen demand and a
14 lesser demand for other nutrients were particularly competitive in areas with acidic soils and high
15 nitrogen deposition. The grasses grew better than herbs with the addition of nitrogen. It was
16 concluded that, at the highest nitrogen deposition, growth was limited for most species by the
17 supply of other nutrients; and, at the intermediate nitrogen concentration, the grasses were more
18 efficient than the herbs in utilizing nitrogen. Nihlgard (1985) suggested that excessive nitrogen
19 deposition may contribute to forest decline in other specific regions of Europe. Also, Schulze
20 (1989), Heinsdorf (1993), and Lamersdorf and Meyer (1993) attribute magnesium deficiencies in
21 German forests, in part, to excessive nitrogen deposition.
22 Plant succession patterns and biodiversity are affected significantly by chronic nitrogen
23 additions in some North American ecosystems (Figure 4-3). Fenn et al. (1998) report that
24 long-term nitrogen fertilization studies in both New England and Europe, as well, suggest that
25 some forests receiving chronic inputs of nitrogen may decline in productivity and experience
26 greater mortality. Long-term fertilization experiments at Mount Ascutney, Vermont, suggest that
27 declining coniferous forest stands with slow nitrogen cycling may be replaced by deciduous
28 fast-growing forests that cycle nitrogen rapidly (Fenn et al., 1998).
29 In experimental studies of nitrogen deposition conducted by Wedin and Tilman (1996) over
30 a 12-year period on Minnesota grasslands, plots dominated by native warm-season grasses
31 shifted to low-diversity mixtures dominated by cool-season grasses at all but the lowest rates of
March 2001 4-47 DRAFT-DO NOT QUOTE OR CITE
-------
N-Saturated Ecosystems
in North America
Review of Ecosystem Effects
and Responses to Excess N
1. Nitrogen Inputs:
> Atmospheric deposition, N2 fixation, fertilization
Nitrogen Retention:
> In plant biomass and soil organic matter
> The role of soil microbes and woody residues
> Abiotic retention
Nitrogen Outputs:
>-Hydrologic transport, gaseous emissions from soil
> Removal in harvest, fire emissions, and soil erosion
2. Characteristics Predisposing Forests 3. Ecosystem Responses to Excess Nitrogen:
to N Saturation:
> Stand vigor and succession, forest type
> Previous land use-stand history
> Soil N accumulation
> Topography and climate
> Nitrogen deposition
> Nitrate leaching and export
>• Eutrophicationof estuaries
> Toxicity of surface waters
> Foliar nutrient responses
> Nitrogen mineralization and nitrification
> Effects on soil organic matter
> Soil acidification, cation depletion, Al toxicity
> Foliar nutrient responses
> Greenhouse gas fluxes
4. Regional N Saturation Conceptual Models:
> New England forests
*• California forests
*• Colorado alpine ecosystems
Figure 4-3. Diagrammatic overview of excess nitrogen (N) in North America.
Source: Fennetal. (1998).
March 2001
4-48
DRAFT-DO NOT QUOTE OR CITE
-------
1 nitrogen addition. Grasslands with high nitrogen retention and carbon storage rates were the
2 most vulnerable to loss of species and major shifts in nitrogen cycling. The shift to low-diversity
3 mixtures was associated with the decrease in biomass carbon to nitrogen (C:N) ratios, increased
4 nitrogen mineralization, increased soil nitrate, high nitrogen losses, and low carbon storage
5 (Wedin and Tilman, 1996). Naeem et al. (1994) experimentally demonstrated (under controlled
6 environmental conditions) that loss of biodiversity, in addition to loss of genetic resources, loss
7 of productivity, loss of ecosystem buffering against ecological perturbation, and loss of aesthetic
8 and commercially valuable resources, also may alter or impair ecosystems services.
9 The C:N ratio of the forest floor can be changed by nitrogen deposition over time. This
10 change appears to occur when the ecosystem becomes nitration saturated (Gundersen et al.,
11 1998a). Long-term changes in C:N status have been documented in Central Europe and indicate
12 that nitrogen deposition has changed the forest floor. In Europe, low C:N ratios coincide with
13 high deposition regions (Gundersen et al., 1998a). A strong decrease in forest floor root biomass
14 has been observed with increased nitrogen availability. Roots and the associated mycorrhizae
15 appear to be an important factor in the accumulation of organic matter in the forest floor at
16 nitrogen limited sites. If root growth and mycorrhizal formation are impaired by nitrogen
17 deposition, the stability of the forest floor may be affected by stimulating turnover and decreasing
18 the root litter input to the forest floor and thus decrease the nitrogen that can be stored in the
19 forest floor pool (Gundersen et al., 1998b). Nitrogen-limited forests have a high capacity for
20 deposited nitrogen to be retained by the plants and microorganisms competing for available
21 nitrogen (Gundersen et al., 1998b). Nitrate leaching has been correlated significantly with nitrate
22 status but not with nitrate depositions. Forest floor C:N ratio has been used as a rough indicator
23 of ecosystem nitrogen status in mature coniferous forests and the risk of nitrate leaching;
24 analyses of European databases indicated an empirical relationship between forest floor C:N ratio
25 and nitrate leaching (Gundersen et al., 1998a). Nitrate leaching was observed when the
26 deposition received was more than 10kg N/ha. All of the data sets supported a threshold at
27 which nitrate leaching seems to increase at a C:N ratio of 25. Therefore, to predict the rate of
28 changes in nitrate leaching it is necessary to be able to predict the rate of changes in the forest
29 floor C:N ratio. Understanding the variability in forest ecosystem response to nitrogen input is
30 essential in assessing pollution risks (Gundersen et al., 1998a).
March 2001 4-49 DRAFT-DO NOT QUOTE OR CITE
-------
1 The plant root is an important region of nutrient dynamics. The rhizosphere includes the
2 soil that surrounds and is influenced by plant roots (Wall and Moore, 1999). The mutualistic
3 relationship between plant roots, fungi, and microbes is critical for the growth of the organisms
4 involved. The plant provides shelter and carbon, whereas the symbiont provides access to a
5 limiting nutrient such as nitrogen and phosphorus. As indicated above, changes in soil nitrogen
6 influence the mycorrhizal-plant relationship. Mycorrhizal fungal diversity is associated with
7 above-ground plant biodiversity, ecosystem variability, and productivity (Wall and Moore, 1999).
8 Aber et al. (1998) showed a close relationship between mycorrhizal fungi and the conversion of
9 dissolved inorganic nitrogen to soil nitrogen. During nitrogen saturation, soil microbial
10 communities change from being fungal, and probably being dominated by mycorrhizae, to being
11 dominated by bacteria. The loss of mycorrhizal function has been hypothesized as the key
12 process leading to increased nitrification and nitrate mobility. Increased nitrate mobility leads to
13 increased cation leaching and soil acidification (Aber et al., 1998).
14 The interrelationship of above- and below-ground flora is illustrated by the natural invasion
15 of heath lands by oaks (Quercus robur). Soils are dynamic entities, the features of which can
16 change like the rest of the ecosystem with age and management. The soil-forming factors under
17 the heath have been vegetation type during the last 2000 years, whereas the invasion by oaks has
18 been taking place for only a few decades. Clear changes in the ground floor and soil morphology
19 takes place when trees colonize heath (Nielsen et al., 1999). The distribution of roots also
20 changed under the three different vegetation types. Under both heather and the Sitka spruce
21 plantation, the majority of roots are confined to the uppermost horizons, whereas under oak, the
22 roots are distributed more homogeneously. There was also a change in the C:N ratio when
23 heather was replaced by oaks. Also, the spontaneous succession of the heath by oaks changed
24 the biological nutrient cycle into a deeper vertical cycle, when compared to the heath where the
25 cycle is confined to the upper soil horizons. Soils similar to those described in this study
26 (Jutland, Denmark), with mainly an organic buffer system, seem to respond quickly to changes in
27 vegetation (Nielsen et al., 1999).
28 In addition to excess nitrogen deposition effects on terrestrial ecosystems of the types noted
29 above (e.g., dominant species shifts and other biodiversity impacts), direct atmospheric nitrogen
30 deposition and increased nitrogen inputs via runoff into streams, rivers, lakes, and oceans can
31 have notable impacts on aquatic ecosystems as well. One illustrative example is recently
March 2001 4-50 DRAFT-DO NOT QUOTE OR CITE
-------
1 reported research (summarized by Paerl et al., in press) characterizing impacts of nitrogen
2 deposition on the Pamlico Sound, NC, estuarine complex, which serves as a key fisheries nursery
3 supporting an estimated 80% of commercial and recreational fmfish and shellfish catches in the
4 southeastern U.S. Atlantic coastal region. Such direct atmospheric nitrogen deposition onto
5 waterways feeding into the Pamlico Sound or onto the sound itself and indirect nitrogen inputs
6 via runoff from upstream watersheds contribute to conditions of severe water oxygen depletion,
7 formation of algae blooms in portions of the Pamlico Sound estuarine complex, and altered fish
8 distributions, catches, and physiological states and incidence of disease. Under extreme
9 conditions of especially high rainfall rate events (e.g., hurricanes) affecting watershed areas
10 feeding into the sound, the effects of nitrogen runoff (in combination with excess loadings of
11 metals or other nutrients) can be massive (e.g., creation of the widespread "dead-zone" affecting
12 large areas of the Pamlico Sound for many months after hurricane Fran in 1996 and hurricanes
13 Dennis, Floyd, and Irene in 1999 impacted eastern North Carolina).
14
15 Sulfur Deposition Effects
16 Sulfur is an essential plant nutrient and, as such, is a major component of plant proteins.
17 The most important source of sulfur is sulfate taken up from the soil by plant roots even though
18 plants can utilize atmospheric SO2 (Marschner, 1995). The availability of organically bound
19 sulfur in soils depends largely on microbial decomposition, a relatively slow process. The major
20 factor controlling the movement of sulfur from the soil into vegetation is the rate of release from
21 the organic to the inorganic compartment (May et al., 1972; U. S. Environmental Protection
22 Agency, 1982; Marschner, 1995). Sulfur plays a critical role in agriculture as an essential
23 component of the balanced fertilizers needed to grow and increase worldwide food production
24 (Ceccotti and Messick, 1997). Atmospheric deposit!on is an important component of the sulfur
25 cycle. This is true not only in polluted areas where atmospheric deposition is very high, but also
26 in areas of low sulfur input. Additions of sulfur into the soil in the form of SO4 2" could alter the
27 important organic-sulfur/organic-nitrogen relationship involved in protein formation in plants.
28 The biochemical relationship between sulfur and nitrogen in plant proteins indicates that neither
29 element can be assessed adequately without reference to the other. There is a regulatory coupling
30 of sulfur and nitrogen metabolism. Sulfur deficiency reduces nitrate reductase and, to a similar
31 extent, also glutamine synthetase activity. Nitrogen uptake in forests, therefore, may be loosely
March 2001 4-51 DRAFT-DO NOT QUOTE OR CITE
-------
1 regulated by sulfur availability, but sulfate additions in excess of needs do not necessarily lead to
2 injury (Turner and Lambert, 1980; Hogan et al., 1998).
3 Only two decades ago, there was little information comparing sulfur cycling in forests with
4 nutrients, especially nitrogen. With the discovery of deficiencies in some unpolluted regions
5 (Kelly and Lambert, 1972; Humphreys et al., 1975; Turner et al., 1977; Schnug, 1997) and
6 excesses associated with acidic deposition in other regions of the world (Meiwes and Khanna,
7 1981; Shriner and Henderson, 1978; Johnson et al., 1982a,b), interest in sulfur nutrition and
8 cycling in forests has heightened. General reviews of sulfur cycling in forests have been written
9 by Turner and Lambert (1980), Johnson (1984), Mitchell et al. (1992a,b), and Hogan et al.
10 (1998). The salient elements of the sulfur cycle as it may be affected by changing atmospheric
11 deposition are summarized by Johnson and Mitchell (1988). Sulfur has become the most
12 important limiting factor in European agriculture because of the desulfurization of industrial
13 emissions (Schnug, 1997).
14 Most studies dealing with the impacts of sulfur deposition on plant communities have been
15 conducted in the vicinity of point sources and have investigated above-ground effects of SO2 or
16 acidifying effects of sulfate on soils (Krupa and Legge, 1998; Dreisinger and McGovern, 1970;
17 Legge, 1980; Winner and Bewley, 1998a,b; Laurenroth and Michunas, 1985; U.S. Environmental
18 Protection Agency, 1982). Krupa and Legge (1986), however, observed a pronounced increase
19 in foliar sulfur concentrations in all age classes of needles of the hybrid pine lodgepole x jack
20 pine (Pinus contorta x P. banksiana). This vegetation had been exposed to chronic low
21 concentrations of sulfur gas pollution (SO2), hydrogen sulfide (H2S), and fugitive sulfur aerosol
22 for more than 20 years. Observations under the microscope showed no sulfur deposits on the
23 needle surfaces and led to the conclusion that the sulfur was derived from the soil. The oxidation
24 of elemental sulfur and the generation of protons is well known for the soils of Alberta, CN.
25 This process is mediated by bacteria of the Thiobacillus sp. As elemental sulfur gradually is
26 converted to protonated SO4, it can be leached downward and readily taken up by plant roots.
27 The activity of Thiobacillus sp. is stimulated by elemental sulfur additions (Krupa and Legge,
28 1986).
29
30
31
March 2001 4-52 DRAFT-DO NOT QUOTE OR CITE
-------
1 Effects of Acidic Deposition on Forest Soils
2 Substantial and previously unsuspected changes in soils are occurring both in polluted areas
3 of eastern North America, the United Kingdom, Sweden, and Central Europe and in less polluted
4 regions of Australia and western North America (reviewed by Johnson et al., 1991b; see review
5 by Huntingdon, 2000). In some cases, trends are toward more acidic soils (e.g., Markewitz et al.,
6 1998), and, in others, there are no consistent trends, with some soils showing increases and some
7 showing decreases at different sampling times, and some showing no change (e.g., Johnson and
8 Todd, 1998; Trettin et al., 1999; Yanai et al., 1999).
9 Significant changes have occurred at many sites in the eastern United States during recent
10 decades. Temporal trends in tree ring chemistry were examined as indicators of historical
11 changes in the chemical environmental of red spruce. Chemical changes in tree-ring chemistry
12 reflect changing inputs of regional pollutants to forests. If significant base cation mobilization
13 and depletion of base cations from eastern forest soils has occurred, a temporal sequence of
14 changes in uptake patterns and possibly in tree growth would be expected. Patterns of tree-ring
15 chemistry principally at high-elevation sites in the eastern United States, leads to the conclusion
16 that significant changes in soil chemistry have occurred in many of these sites during recent
17 decades leading to changes in growth (Bondietti and McLaughlin, 1992). These changes are
18 spatially and temporally consistent with emissions of SO2 and NO2 across the region, suggesting
19 that increased acidification of forest soils has occurred.
20 Increases in levels of Al and Fe typically occur as base cations are removed from soils by
21 tree uptake. A region-wide Ca increase above expected levels followed by a decrease suggests
22 that increased mobilization began perhaps 30 to 40 years ago (Bondietti and McLaughlin, 1992).
23 The period of Ca mobilization coincides with a region-wide increase in growth rate of red spruce,
24 whereas the period of decreasing levels of Ca in wood corresponds temporally with patterns of
25 decreasing radial growth at high elevation sites throughout the region during the past 20 to 30
26 years. The decline in wood Ca suggests that Ca loss may have been increased to the point at
27 which base saturation of soils has been reduced (Bondietti and McLaughlin, 1992).
28 Studies by Shortle and Bondietti, (1992) support the view that changes in soil chemistry in
29 eastern North America forest sites occurred many decades ago "before anybody was looking".
30 Sulfur and nitrogen emissions began increasing in eastern North America in the 1920s and
31 continued to increase into the 1980s, when sulfur began to decrease but nitrogen emissions have
March 2001 4-53 DRAFT-DO NOT QUOTE OR CITE
-------
1 not (Garner et al., 1989). Shortle and Bondietti (1992) present evidence that, from the late 1940s
2 into the 1960s, the mor humus layer of acid-sensitive forest sites in eastern North America
3 underwent a significant change that resulted in the loss of exchangeable essential base cations
4 and interrupted the critical base nutrient cycles between mature trees and the root-humus
5 complex. The timing of the impact appears to have coincided with the period when the SOX and
6 NOX emissions in eastern North America subject to long-range transport were increasing the most
7 rapidly (See above; Shortle and Bondietti, 1992). Although forest ecosystems other than the
8 high-elevation spruce-fir forests are not currently manifesting symptoms of injury directly
9 attributable to acid deposition, less sensitive forests throughout the United States are
10 experiencing gradual losses of base cation nutrient, which in many cases will reduce the quality
11 of forest nutrition over the long term (National Science and Technology Council, 1998). In some
12 cases, it may not even take decades, because these forests already have been receiving sulfur and
13 nitrogen deposition for many years. The current status of forest ecosystems in different U.S.
14 geographic regions varies, as does their sensitivity to nitrogen and sulfur deposition. Variation in
15 potential future forest responses or sensitivity are caused, in part, by differences in deposition of
16 sulfur and nitrogen, ecosystem sensitivities to sulfur and nitrogen additions, and responses of
17 soils to sulfur and nitrogen inputs (National Science and Technology Council, 1998).
18 Acidic deposition has played a major role in recent soil acidification in some areas of
19 Europe and, to a more limited extent, eastern North America. Examples include the study by
20 Hauhs (1989) at Lange Bramke, Germany, which indicated that leaching was of major
21 importance in causing substantial reduction in soil-exchangeable base cations over a 10-year
22 period (1974-1984). Soil acidification and its effects result from the deposition of nitrate (NO3")
23 and sulfate (SO4 2") and the associated hydrogen (H +) ion. The effects of excessive nitrogen
24 deposition on soil acidification and nutrient imbalances have been well established in Dutch
25 forests (Van Breemen et al., 1982; Roelofs et al., 1985; Van Dijk and Roelofs, 1988).
26 For example, Roelofs et al. (1987) proposed that NH3 /NH4+ deposition leads to heathland
27 changes via two modes: (1) acidification of the soil and the loss of cations K+, Ca2+, and Mg2+;
28 and (2) nitrogen enrichment that results in "abnormal" plant growth rates and altered competitive
29 relationships. Nihlgard (1985) suggested that excessive nitrogen deposition may contribute to
30 forest decline in other specific regions of Europe. Falkengren-Grerup (1987) noted that, during
March 2001 4-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 about 50 years, unexpectedly large increases in growth of beech (Fagus sylvatica L.) were
2 associated with decreases in pH and exchangeable cations in some sites in southernmost Sweden.
3 Likens et al. (1996) suggested that soils are changing at the Hubbard Brook Watershed,
4 NH, because of a combination of acidic deposition and reduced base cation deposition. They
5 surmised, based on long-term trends in stream-water data, that large amounts of Ca and Mg have
6 been lost from the soil-exchange complex over a 30-year period from approximately 1960 to
7 1990. The authors speculate that the declines in base cations in soils may be the cause of recent
8 slowdowns in forest growth at Hubbard Brook. In a follow-up study, however, Yanai et al.
9 (1999) found no significant decline in Ca and Mg concentrations in forest floors at Hubbard
10 Brook over the period 1976 to 1997. They also found both gains and losses in forest floor Ca
11 and Mg between 1980 and 1990 in a regional survey. Thus, they concluded that "Forest floors in
12 the region are not currently experiencing rapid losses of base cations, although losses may have
13 preceded the onset of these three studies."
14 Hydrogen ions entering a forest ecosystem first encounter the forest canopy, where they are
15 often exchanged for base cations that then appear in throughfall (Figure 4-4 depicts a model of
16 H+ sources and sinks). Base cations leached from the foliage must be replaced through uptake
17 from the soil, or foliage cations will be reduced by the amounts leached. In the former case, the
18 acidification effect is transferred to the soil, where H + is exchanged for a base cation at the
19 root-soil interface. Uptake of base cations or NH4 + by vegetation or soil microorganisms causes
20 the release of H + in order to maintain charge balance. Uptake of nutrients in anionic form (NO3",
21 SO4 2", PO4 3") causes the release of OH" in order to maintain charge balance. Thus, the net
22 acidifying effect of uptake is the difference between cation and anion uptake. The form of ions
23 taken up is known for all nutrients but nitrogen, where either NH4+ or NO3" can be taken up.
24 In that, nitrogen is a nutrient taken up in great quantities, the uncertainty in the ionic form of
25 nitrogen taken up creates great uncertainty in the overall H+ budget for soils (Johnson 1992).
26 The cycles of base cations differ from those of N, P, and S in several respects. The fact that
27 Ca, K, and Mg exist primarily as cations in solution whereas N, P, and S exist primarily as anions
28 has major implications for the cycling of the nutrients and the effects of acid deposition on these
29 cycles. The most commonly accepted model of base cation cycling in soils is one in which base
30 cations are released by weathering of primary minerals to cation exchange sites, where they are
31 then available for either plant uptake or leaching (Figure 4-4). The introduction of H + by
March 2001 4-55 DRAFT-DO NOT QUOTE OR CITE
-------
Deposition
hT
NO; sol
Soil
Organism
Uptake
2+
Ca
2OH
2H+ + NO3
Nitrification
CO2 + H2O
Carbonic Acid Formation
R-COOH
Organic Acid Formation
Tree
Uptake
2OH
SoiS
Organism
Uptake
2OH
Leaching
Figure 4-4. Schematic of sources and sinks of hydrogen ions in a forest (from Taylor et al.,
1994).
1 atmospheric deposition or by internal processes will impact directly the fluxes of Ca, K, and
2 Mg via cation exchange or weathering processes. Therefore, soil leaching is often of major
3 importance in cation cycles, and many forest ecosystems show a net loss of base cations
4 (Johnson, 1992a).
5 Two basic types of soil change are involved: (1) a short-term intensity type change
6 resulting from the concentrations of chemicals in soil water, and (2) a long-term capacity change
March 2001
4-56
DRAFT-DO NOT QUOTE OR CITE
-------
1 based on the total content of bases, aluminum and iron stored in the soil (Reuss and Johnson,
2 1986; Van Breemen, 1983). Changes in intensity factors can have a rapid impact on the
3 chemistry of soil solutions. Increases in the amounts of sulfur and nitrogen in acidic deposition
4 can cause immediate increases in acidity and mobilization of aluminum in soil solutions.
5 Increased aluminum concentrations and an increase in the Ca/Al ratio in soil solution have been
6 linked to a significant reduction in the availability of essential base cations to plants, an increase
7 in plant respiration, and increased biochemical stress (National Science and Technology Council,
8 1998).
9 Rapid changes in intensity, resulting from the addition of increased amounts of nitrogen or
10 sulfur in acidic deposition, can have a rapid impact on the chemistry of soil solutions by
11 increasing the acidity and mobilizing aluminum. Increased concentrations of aluminum and an
12 increase in the ratio of calcium-to-aluminum in soil solution have been linked to significantly
13 reduced availability of essential cations to plants.
14 Capacity changes are the result of many factors acting over long time periods. The content
15 of base cations (calcium, magnesium, sodium, and potassium) in soils result from additions from
16 the atmospheric deposition, decomposition of vegetation and geologic weathering. Loss of base
17 cations may occur through plant uptake and leaching. Increased leaching of base cations may
18 result in nutrient deficiencies in soils as has been happening in some sensitive forest ecosystems
19 (National Science and Technology Council, 1998).
20 A major concern has been that soil acidity would lead to nutrient deficiency. Calcium is
21 essential in the formation of wood and the maintenance of cells, the primary plant tissues
22 necessary for tree growth. Trees obtain Ca from the soil, but to be taken up by roots, the Ca
23 (a positively charged ion) must be dissolved in soil water (Lawrence and Huntington, 1999).
24 Tree species may be adversely affected if high Al to nutrient ratios limit uptake of Ca and Mg
25 and create a nutrient deficiency (Shortle and Smith, 1988; Garner, 1994). Acid deposition by
26 lowering the pH of aluminum-rich soil can increase aluminum concentrations in soil water
27 through dissolution and ion-exchange processes. When in solution, aluminum can be taken up
28 by roots, transported through the tree and, eventually, deposited on the forest floor in leaves and
29 branches. Aluminum is more readily taken up than is Ca because it has a higher affinity for
30 negatively charged surfaces than does Ca. When present in the forest floor, Al tends to displace
31 adsorbed Ca and causes it to be more readily leached. The continued buildup of Al in the forest
March 2001 4-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 floor layer, where nutrient uptake is greatest, can decrease the availability of Ca to the roots
2 (Lawrence et al., 1995) and lower the efficiency of Ca uptake because Al is more readily taken up
3 than is Ca 2+when the ratio of Ca to Al in soil water is less than one (Lawrence and Huntington,
4 1999). A 1968 Swedish report to the United Nations postulated a decrease in forest growth of ca.
5 1.5%/year as result of Ca2+ loss by leaching (Johnson and Taylor, 1989). The concern that soil
6 acidification and nutrient deficiency may result in forest decline remains extant today.
7 Aluminum toxicity is a possibility in acidified soils. Atmospheric deposition (or any other
8 source of mineral anions) can increase the concentration of Al, especially A13+, in soil solution
9 without causing significant soil acidification (Johnson and Taylor, 1989). Aluminum can be
10 brought into soil solution in two ways: (1) by acidification of the soil and (2) by an increase in
11 the total anion and cation concentration of the soil solution. The introduction of mobile, mineral
12 acid anions to an acid soil will cause increases in the concentration of aluminum in the soil
13 solution, but extremely acid soils in the absence of mineral acid anions will not produce a
14 solution high in aluminum. An excellent review of the relationships among the most widely used
15 cation-exchange equations and their implications for the mobilization of aluminum into soil
16 solution is provided by Reuss (1983).
17 Aluminum toxicity may influence forest tree growth, where acid deposition and natural
18 acidifying processes increase soil acidity. Aluminum concentrations have been observed to
19 exhibit a strongly descending gradient from bulk soil through the rhizosphere to the root (Smith,
20 1990a). Once it enters the forest tree roots, Al accumulates in root tissue (Thornton et al., 1987;
21 Vogt et al., 1987a,b). There is abundant evidence that Al is toxic to plants. Reductions of Ca
22 uptake by roots has been associated with increases in Al uptake (Clarkson and Sanderson, 1971).
23 Calcium plays a major role in cell membrane integrity and cell wall structure. A number of
24 studies have suggested that the toxic effect of aluminum on forest trees could be caused by Ca2+
25 deficiency (Shortle and Smith, 1988; Smith, 1990a). Mature trees have a high Ca2+ requirement
26 relative to agricultural crops (Rennie, 1955). Shortle and Smith (1988) attributed the decline of
27 red spruce in eight stands across northern New England from Vermont to Maine to an imbalance
28 of A13+ and Ca2+ in the fine root environment. Aluminum in the soil solution reduces Ca uptake
29 by competing for binding sites in the cortex of fine roots. Reduction in Ca uptake suppresses
30 cambial growth and reduces the rate of wood formation (annual ring formation), decreases the
31 amount of functional sapwood and live crown, and predisposes trees to disease and injury from
March 2001 4-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 stress agents when the functional sapwood becomes less than 25% of cross-sectional stem area
2 (Smith, 1990a).
3 Air pollution is not the sole cause of soil change. High rates of acidification are occurring
4 in less polluted regions of the western United States and Australia because of internal soil
5 processes, such as tree uptake of nitrate and nitrification associated with excessive nitrogen
6 fixation (Johnson et al., 1991b). Many studies have shown that acidic deposition is not a
7 necessary condition for the presence of extremely acid soils, as evidenced by their presence in
8 unpolluted, even pristine forests of the northwestern United States and Alaska (Johnson et al.,
9 1991b). The soil becomes acidic when H+ ions attached to NH4+ or HNO3 remain in the soil after
10 nitrogen is taken up by plants. For example, Johnson et al. (1982b) found significant reductions
11 in exchangeable K + over a period of only 14 years in a relatively unpolluted Douglas fir
12 Integrated Forest Study (IFS) site in the Washington Cascades. The effects of acid deposition at
13 this site were negligible relative to the effects of natural leaching (primarily carbonic acid) and
14 nitrogen tree uptake (Cole and Johnson, 1977). Even in polluted regions, numerous studies have
15 shown the importance of tree uptake of NH4+ and NO3" in soil acidification. Binkley et al. (1989)
16 attributed the marked acidification (pH decline of 0.3 to 0.8 units and base saturation declines of
17 30 to 80%) of abandoned agricultural soil in South Carolina over a 20-year period to NH4+ and
18 NO3" uptake by a loblolly pine plantation.
19 An interesting example of uptake effects on soil acidification is that of Al uptake and
20 cycling (Johnson et al., 1991b). Aluminum accumulation in the leaves of coachwood
21 (Ceratopetalum apetaluni) in Australia has been found to have a major impact on the distribution
22 and cycling of base cations (Turner and Kelly, 1981). The presence of C. apetalum as a
23 secondary tree layer beneath brush cox (Lophostemon confertus) was found to lead to increased
24 soil exchangeable Al3+ and decreased soil exchangeable Ca2+ (Turner and Kelly, 1981). The
25 constant addition of aluminum-rich litter fall obviously has had a substantial effect on soil
26 acidification, even if base cation uptake is not involved directly.
27 Given the potential importance of particulate deposition for base cation status of forest
28 ecosystems, the findings of Driscoll et al. (1989) and Hedin et al. (1994) are especially relevant.
29 Driscoll et al. (1989) noted a decline in both SO42" and base cations in both atmospheric
30 deposition and stream water over the past two decades at Hubbard Brook Watershed, NH. The
March 2001 4-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 decline in SO42" deposition was attributed to a decline in emissions, and the decline in stream
2 water SO42" was attributed to the decline in sulfur deposition.
3 Hedin et al. (1994) reported a steep decline in atmospheric base cation concentrations in
4 both Europe and North America over the past 10 to 20 years. The reductions in SO 2 emissions
5 in Europe and North America in recent years have not been accompanied by equivalent declines
6 in net acidity related to sulfate in precipitation. These current declines in sulfur deposition have,
7 in varying degrees, been offset by declines in base cations and may be contributing "to the
8 increased sensitivity of poorly buffered systems." Analysis of the data from the Integrated Forest
9 Studies (IPS) supports the authors' contention that atmospheric base cation inputs may seriously
10 affect ecosystem processes. Johnson et al. (1994a) analyzed base cation cycles at the Whiteface
11 Mountain IPS site in detail and concluded that Ca losses from the forest floor were much greater
12 than historical losses, based on historical changes in forest floor Ca observed in an earlier study
13 (Johnson et al., 1994b). Further, the authors suggest that the difference between historical and
14 current net loss rates of forest floor Ca may be caused by sharply reduced atmospheric inputs of
15 calcium after about 1970 and exacerbated by sulfate leaching (U.S. Environmental Protection
16 Agency, 1999).
17 The calcium/aluminum molar ratio has been suggested as a valuable ecological indicator of
18 an approximate threshold beyond which the risk of forest injury from Al stress and nutrient
19 imbalances increases (Cronan and Grigal, 1995). The Ca/Al ratio also can be used as an
20 indicator to assess forest ecosystem changes over time in response to acidic deposition, forest
21 harvesting, or other process that contribute to acid soil infertility. This ratio, however, may not
22 be a reliable indicator of stress in areas with both high atmospheric deposition of ammonium and
23 magnesium deficiency via antagonism involving ammonium rather than aluminum, and in areas
24 with soil solutions with calcium concentrations greater than 500 micromoles per liter (National
25 Science and Technology Council, 1998). Cronan and Grigal (1995) based on a review of the
26 literature have made the following estimates for determining the adverse impact of acidic
27 deposition on tree growth or nutrition:
28 • forests have a 50% risk of adverse impacts if the Ca/Al ration is 1.0,
29 • the risk is 75% if the ratio is 0.5, and
30 • the risk approaches 100% if the ratio is 0.2.
March 2001 4-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 The Ca/Al ratio of soil solution provides only an index of the potential for Al stress. Cronan and
2 Grigal (1995), state that the overall uncertainty of the Ca/Al ratio associated with a given
3 probability ratio is considered to be approximately ±50%. Determination of thresholds for
4 potential forest impacts requires the use of the four successive measurement endpoints in the soil,
5 soil solution, and plant tissue listed below.
6 (1) Soil base saturation less than 15% of effective cation exchange capacity
7 (2) Soil solution Ca/Al molar ratio less than 1.0 for 50% risk
8 (3) Fine roots tissue Ca/Al molar ratio less than 0.2 for 50% risk
9 (4) Foliar tissue Ca/Al molar ratio less than 12.5 for 50% risk
10 The application of the Ca/Al ratio indicator for assessment and monitoring of forest health risks
11 has been recommended for sites or in geographic regions where the soil base saturation <15%.
12
13 Critical Loads
14 In Europe, the critical load concept generally has been accepted as the basis for abatement
15 strategies to reduce or prevent injury to the functioning and vitality of forest ecosystems caused
16 by long-range transboundary acidic deposition (Lokke, et al., 1996). The critical load has been
17 defined as a "quantitative estimate of an exposure to one or more pollutants below which
18 significant harmful effects on specified sensitive elements of the environment do not occur
19 according to present knowledge" (Lokke et al., 1996). A biological indicator, a chemical
20 criterion, and a critical value are the elements used in the critical load concept. The biological
21 indicator is the organism used to indicate the status of the receptor ecosystem, the chemical
22 criterion is the parameter that results in harm to the biological indicator, and the critical value is
23 the value of the chemical criterion below which no significant harmful response occurs to the
24 biological indicator (Lokke et al., 1996). Trees, and sometimes other plants, are used as the
25 biological indicators in the case of critical loads for forests. The critical load calculation using
26 the current methodology, is essentially an acidity/alkalinity mass balance calculation. The
27 chemical criterion must be expressible in terms of alkalinity. Initially, the Ca/Al ratio was used,
28 but, recently, the (Ca+Mg+K)/Al ratio has been used (Lokke et al., 1996).
29 Ideally, changes in acidic deposition should result in changes in the status of the biological
30 indicator used in the critical load calculation. However, the biological indicator is the integrated
31 response to a number of different stresses. Furthermore, there are other organisms more sensitive
March 2001 4-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 to acid deposition than trees. At high concentrations, Al3+ is known to be toxic to plants,
2 inhibiting root growth and, ultimately, plant growth and performance (Lokke et al., 1996;
3 National Science and Technology Council, 1998). Sensitivity to Al varies considerably between
4 species and within species because of changes in nutritional demands and physiological status,
5 which are related to age and climate. Experiments have shown that there are large variations in
6 Al sensitivity, even among ecotypes.
7 Mycorrhizal fungi as possible biological indicators have been suggested by Lokke et al.
8 (1996) because they are intimately associated with tree roots, depend on plant assimilates, and
9 play an essential role in plant nutrient uptake, influencing the ability of their host plants to
10 tolerate different anthropogenically generated stresses. Mycorrhizas and fine roots are an
11 extremely dynamic component of below-ground ecosystems and can respond rapidly to stress.
12 They have a relatively short life span, and their turnover appears to be strongly controlled by
13 environmental factors. Changes in mycorrhizal species composition or the loss of dominant
14 mycorrhizal species in areas where diversity is already low may lead to increased susceptibility of
15 plant to stress (Lokke et al., 1996). Stress affects the total amount of carbon fixed by plants and
16 modifies carbon allocation to biomass, symbionts and secondary metabolites. Because
17 mycorrhizal fungi are dependent for their growth on the supply of assimilates from the host
18 plants, stresses that shift the allocation of carbon reserves to the production of new leaves at the
19 expense of supporting tissues will be reflected rapidly in decreased fine root and mycorrhizzal
20 biomass (Winner and Atkinson, 1986). The physiology of carbon allocation has also been
21 suggested as an indicator of anthropogenic stress (Andersen and Rygiewicz, 1991). Soil
22 dwelling animals are important for decomposition, soil aeration, and nutrient redistribution in the
23 soil. They contribute to decomposition and nutrient availability mainly by increasing the
24 accessibility of dead plant material to microorganisms. Earthworms decrease in abundance and
25 in species number in acidified soils Lokke et al., 1996).
26
27 Biogeochemical Cycling—The Integrated Forest Study
28 The Integrated Forest Study (IPS) (Johnson and Lindberg, 1992a) has provided the most
29 extensive data set available on wet and dry deposition and the effects of deposition on the cycling
30 of elements in forest ecosystems. The overall patterns of deposition and cycling have been
31 summarized by Johnson and Lindberg (1992a), and the reader is referred to that reference for
March 2001 4-62 DRAFT-DO NOT QUOTE OR CITE
-------
1 details. The following is a summary of particulate deposition, total deposition, and leaching in
2 the IPS sites.
3 Particulate deposition in the IPS was separated at the 2-//m level; a decision was made to
4 include total particulate deposition in this analysis and may include the deposition of particles
5 larger than 10 //m.
6 Particulate deposition contributes considerably to the total impact of base cations to most of
7 the IPS sites. On average, particulate deposition contributes 47% to total calcium deposition
8 (range: 4 to 88%), 49% of total potassium deposition (range: 7 to 77%), 41% to total magnesium
9 deposition (range: 20 to 88%), 36% to total sodium deposition (range: 11 to 63%), and 43% to
10 total base cation deposition (range: 16 to 62%). Of the total particulate deposition, the vast
11 majority (>90%) is >2 //m.
12 Figures 4-5 through 4-8 summarize the deposition and leaching of calcium, magnesium,
13 potassium, and total base cations for the IPS sites. As noted in the original synthesis (Johnson
14 and Lindberg, 1992a), some sites show net annual gains of base cations (i.e., total deposition
15 > leaching), some show losses (total deposition < leaching), and some are approximately in
16 balance. Not all cations follow the same pattern at each site. For example, calcium shows net
17 accumulation at the Coweeta, TN; Durham (Duke), NC; and Florida sites (Figure 4-5), potassium
18 shows accumulation at the Duke; Florida; Douglas-fir; red alder; Thompson, WA; Huntingdon
19 Forest, NY; and Whiteface Mountain, NY, sites (Figure 4-7), and magnesium accumulated only
20 at the Florida sites (Figure 4-6). Only at the Florida site is there a clear net accumulation of total
21 base cations (Figure 4-8).
22 The factors affecting net calcium accumulation or loss include the soil-exchangeable cation
23 composition, as noted previously; base cation deposition rate; the total leaching pressure because
24 of atmospheric sulfur and nitrogen inputs, as well as natural (carbonic and organic) acids; and
25 biological demand (especially for potassium). At the Florida site, which has a very cation-poor,
26 sandy soil (an Ultic Haploquod derived from marine sand), the combination of all these factors
27 leads to net base cation accumulation from atmospheric deposition (Johnson and Lindberg,
28 1992a). The site showing the greatest net base cation losses, the red alder stand in Washington
29 state, is one that is under extreme leaching pressure by nitrate produced because of excessive
30 fixation by that species (Van Miegroet and Cole, 1984). In the red spruce site in the Smokies,
31 the combined effects of SO42" and NO3" leaching are even greater than in the red alder site
March 2001 4-63 DRAFT-DO NOT QUOTE OR CITE
-------
1,000
CD
CD
'CD
.c
-i-«
"CD
'5
o-
LU
-1,000
-1,500-
-2,000
63%
41%
^
w
• <2^
^ Wet
^ Leaching
Pe
55%
rcent c
65%
I
1
)f total
40%
•
depos
26%
=
I
ition a;
25%
I
I
I
i
s partic
49%
1
;les:
58%
I
88%
0
4%
I
^
D y
|
i
M
CP DL GS LP FS DF RA NS FF MS WF ST
I Warmer Sites >• * Colder Sites >•
Figure 4-5. Calcium deposition in >2-jum particles, <2-(j,m particles, and wet forms (upper
bars) and leaching (lower bars) in the Integrated Forest Study sites.
CP = Pinus strobus, Coweeta, TN; DL = Pinus taeda, Durham (Duke), NC;
GS = Pinus taeda, B. F. Grant Forest, GA; LP = Pinus taeda, Oak Ridge, TN;
FS = Pinus eliottii, Bradford Forest, FL; DF = Psuedotsuga menziesii,
Thompson, WA; RA = Alnus rubra; Thompson WA; NS = Picea abies,
Nordmoen, Norway; HF = northern hardwood, Huntington Forest, NY;
MS = Picea rubens, Howland, ME; WF = Picea rubens, Whiteface Mountain,
NY; and ST = Picea rubens, Clingman's Dome, NC.
1 (Figure 4-9), but a considerable proportion of the cations leached from this extremely acid soil
2 consist of FT and A13+ rather than of base cations (Johnson and Lindberg, 1992a). Thus, the red
3 spruce site in the Smokies is approximately in balance with respect to calcium and total base
4 cations, despite the very high leaching pressure at this site (Figures 4-5 and 4-8).
5 The relative importance of particulate base cation deposition varies widely with site and
6 cation and is not always related to the total deposition rate. The proportion of calcium deposition
7 in parti culate form ranges from a low of 4% at the Whiteface Mountain site to a high of 88% at
8 the Maine site (Figure 4-5). The proportion of potassium deposition as particles ranges from
9 7% at the Smokies site to 77% at the Coweeta site (Figure 4-7), and the proportion of total base
March 2001
4-64
DRAFT-DO NOT QUOTE OR CITE
-------
400 -
03
CD
>.
co 0 -
.c
-(—i
0 ?oo -
03 ~Z(J(J
'5
LU -400 -
-600 -
40%
I
w
m
49%
^
I
> 2 |jm
< 2 pin
Wet
Leaching
p<
44%
9rcent
48%
i
i
i
oftota
36%
y
depo-
27%
^
i
sition a
27%
^
I
I
i
s parti
39%
i
cles:
46%
m
88%
i
20%
I
1
26%
—
i
i
CP DL GS LP FS DF RA NS FF MS WF ST
Warmer Sites ^^ Colder Sites
Figure 4-6. Magnesium deposition in >2-(j,m particles, <2-(j,m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-5 for legend.
1 cation deposition ranges from 16% at the Whiteface site to 62% at the Maine site (Figure 4-8).
2 Overall, paniculate deposition at the site in Maine accounted for the greatest proportion of
3 calcium, magnesium, potassium, and base cation deposition (88, 88,57, and 62%, respectively),
4 even though total deposition was relatively low. At some sites, the relative importance of
5 particulate deposition varies considerably by cation. At the Whiteface Mountain site, particulate
6 deposition accounts for 4, 20, and 40% of calcium, magnesium, and potassium deposition,
7 respectively. At the red spruce site in the Smokies, particulate deposition accounts for 46, 26%,
8 7% of calcium, magnesium, and potassium deposition, respectively.
9 As observed in the IPS synthesis, SO42" and NOj leaching often are dominated by
10 atmospheric sulfur and nitrogen (Johnson and Lindberg, 1992a). The exceptions to this are in
11 cases where natural nitrogen inputs are high (i.e.,the nitrogen-fixing red alder stand), as are
12 leaching rates, even though nitrogen deposition is low, and where soils adsorb much of the
March 2001
4-65
DRAFT-DO NOT QUOTE OR CITE
-------
co
*,
'co
c
_CD
CO
'5
CT
UJ
400'
300-
200-
100'
0'
-100-
-200
-300
-400
-500
77%
s
m
76%
^
1
Q > 2 pm
^ Wet
^ Leaching
p<
61%
n
m
Brcent
48%
.— — -
I
of tota
54%
Pi
depos
40%
H
ition a
40%
1
I
\
s partic
68%
1
L/J
elates
23%
i
' 57%
w
m
40%
W
7%
1
i
%
I
!
%
CP DL GS LP FS
I Warmer Sites
DF
RA NS
FF MS WF ST
- Colder Sites ^
Figure 4-7. Potassium deposition in >2-(j,m particles, <2-(j,m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-5 for legend.
1
2
3
4
5
6
1
8
9
10
11
12
13
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 4-9 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 (wet and gaseous deposition, internal production).
As noted in the IPS synthesis, total SO42" and NO3" inputs account for a large proportion
(28 to 88%) 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.
March 2001
4-66
DRAFT-DO NOT QUOTE OR CITE
-------
co
Q)
cb
3,000
2,000
1,000
-1,000
-2,000 -
-3,000 -
-4,000
53%
1
47%
i
1
FJ > 2 Mm
^J < 2 Mm
^ Wet
^ Leaching
P
48%
555
1
ercent
62%
I
i
of tota
49%
^_^^
i
1
1 depo
28%
H
1
sition £
28%
S
I
I
I
3S part
47%
1
culate
44%
I
s:
62%
1
16%
1
31%
I
I
1
CP DL GS LP FS DF RA NS FF MS WF ST
Warmer Sites >-4 Colder Sites
Figure 4-8. Base cation deposition in >2-^m particles, <2-/j,m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-5 for legend.
1 It was noted previously in this chapter that the contribution of particles to total deposition
2 of nitrogen and sulfur at the IPS sites is lower than is the case for base cations. On average,
3 particulate deposition contributes 18% to total nitrogen deposition (range: 1 to 33%) and 17%
4 to total sulfur deposition (range: 1 to 30%). Particulate deposition contributes only a small
5 amount to total H+ deposition (average = 1%; range: 0 to 2%). (It should be noted, however,
6 that particulate H+ deposition in the >2-//m fraction was neglected.)
7 Based on the IPS data, it appears that the particulate deposition has a greater effect on base
8 cation inputs to soils than on base cation losses associated with inputs of sulfur, nitrogen, and H+.
9 It cannot be determined what fraction of the mass of these particles are <10 //m, but only a very
10 small fraction is <2 //m. These inputs of base cations have considerable significance, not only to
11 the base cation status of these ecosystems, but also to the potential of incoming precipitation to
12 acidify or alkalize the soils in these ecosystems. As noted above, the potential of precipitation to
March 2001
4-67
DRAFT-DO NOT QUOTE OR CITE
-------
7,000
6,000
",_ 5,000
CD
4,000 -H
% of total cation leaching balanced by SO^" and NOj from Particles (P) and other (O) sources
P: 4 %
0: 28%
7%
55%
1%
6%
8%
78%
9%
30%
10%
42%
Other Anions
Particulate Sulphur and Nitrogen
Other Sulphur and Nitrogen Sources
10%
88%
8%
55%
6%
73%
18%
69%
1%
77%
DF RA NS HF MS WF
9%
69%
I
ST
Figure 4-9. 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. See Figure 4-5 for legend.
1
2
3
4
5
9
10
11
acidify or alkalize soils depends on the ratio of base cations to H+ in deposition, rather than
simply on the inputs 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 magnesium and potassium.
Sodium is a rather special case, in that it is a poorly absorbing 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 simply compare soil pool sizes with the net balances. This comparison is
made for exchangeable pools and net balances for a 25-year period in Figures 4-10 to 4-12.
It readily is seen that net leaching losses of cations pose no threat in terms of depleting
soil-exchangeable Ca2+, K+, or magnesium ion within 25 years at the Coweeta, Duke, Georgia,
March 2001
4-68
DRAFT-DO NOT QUOTE OR CITE
-------
Soil Exchangeable
(Dep - Leaching)*25
-50,000
CP DL GS LP FS DF RA NS HF MS WF ST
Warmer Sites >-*< Colder Sites
Figure 4-10. Soil exchangeable Ca2+ pools and net annual export of Ca2+ (deposition minus
leaching times 25 years) in the Integrated Forest Study sites. See Figure 4-5
for legend.
1 Oak Ridge, or Douglas-fir sites. There, however, is a potential for significant depletion at the red
2 alder, Whiteface Mountain (magnesium), and Smokies red spruce sites.
3 The range of values for soil-exchangeable turnover is very large, reflecting variations in
4 both the size of the exchangeable pool and the net balance of the system. Soils with the highest
5 turnover rates are those most likely to experience changes in the shortest time interval, other
6 things being equal. Thus, the Whiteface Mountain, Smokies, and Maine red spruce sites; the
7 Thompson red alder site; and the Huntington Forest northern hardwood site appear to be most
8 sensitive to change. The actual rates, directions, and magnitudes of changes that may occur in
9 these soils (if any) will depend on weathering inputs and vegetation outputs, in addition to
10 deposition and leaching. It is noteworthy that each of the sites listed above as sensitive has a
11 large store of weatherable minerals, whereas many of the other soils, with larger exchangeable
12 cation reserves, have a small store of weatherable minerals (e.g., Coweeta white pine, Duke
March 2001
4-69
DRAFT-DO NOT QUOTE OR CITE
-------
100,000
Soil Exchangeable
(Dep - Leaching)*25
GS LP FS
Warmer Sites
HF MS WF
Colder Sites
Figure 4-11. Soil exchangeable Mg2+ pools and net annual export of Mg2+ (deposition
minus leaching times 25 years) in the Integrated Forest Study sites.
See Figure 4-5 for legend.
1 loblolly pine, Georgia loblolly pine, and Oak Ridge loblolly pine) (Johnson and Lindberg, 1992a;
2 April and Newton, 1992).
3 Base cation inputs are especially important to the Smokies red spruce site because of
4 potential aluminum toxicity and calcium and magnesium deficiencies. Johnson et al. (199la)
5 found that soil solution aluminum concentrations occasionally reached levels found to inhibit
6 calcium uptake and cause changes in root morphology in solution culture studies of red spruce
7 (Raynal et al., 1990). In a follow-up study, Van Miegroet et al. (1993) found a slight but
8 significant growth response to calcium and magnesium fertilizer in red spruce saplings near the
9 Smokies red spruce site. Joslin et al. (1992) reviewed soil and solution characteristics of red
10 spruce in the southern Appalachians, and it would appear that the IPS site is rather typical.
11 Wesselink et al. (1995) reported on the complicated interactions among changing
12 deposition and soils at this site (including repeated sampling of soil exchangeable base cation
13 pools) from 1969 to 1991 and compared these results with those of a simulation model. They
March 2001
4-70
DRAFT-DO NOT QUOTE OR CITE
-------
iuu,uuu •
140 000 -
t_
(0
9J_ 1 20 000 -
ra
-c 1 00 000 -
ff
fe fin ono .
vw
•c «n ooo -
y
*"^ 40 non .
il
(0 on 000 -
'5
HI 0 •
-20 000 -
1L
_i — i
_D
CP DL
-------
1 variety of forest ecosystems and to determine if these effects are in any way related to current or
2 potential forest decline. Acidic deposition is having a significant effect on nutrient cycling in
3 most of the forest ecosystems studied in the IPS project. The exceptions were the relatively
4 unpolluted Douglas fir, red alder, and Findley Lakes in Washington state. The nature of the
5 effects, however, varies from one location to another (Johnson, 1992). In all but the relatively
6 unpolluted Washington sites, atmospheric deposition was having a significant, often
7 overwhelming effect on cation leaching from the soils. In general, nutrient budget data from IPS
8 and literature suggest that the susceptibility of southeastern sites to base cation depletion from
9 soils and the development of cation deficiencies by that mechanism appears to be greater than in
10 northern sites (Johnson, 1992).
11 Atmospheric deposition may have affected significantly the nutrient status of some IPS
12 sites through the mobilization of Al. Soil solution Al levels in the Smokies sites approach and
13 sometimes exceed levels noted to impede cation uptake in solution culture studies. It is therefore
14 possible that the rates of base cation uptake and cycling in these sites have been reduced because
15 of soil solution Al. To the extent that atmospheric deposition has contributed to these elevated
16 soil solution Al levels, it has likely caused a reduction in base cation uptake and cycling rates at
17 these sites. Nitrate and sulfate are the dominant anions in the Smokies sites, and nitrate pulses
18 are the major cause of Al pulses in soil solution (Johnson, 1992). The connection between Al
19 mobilization and forest decline is not clear. The decline in red spruce certainly has been more
20 severe in the Northeast than in the Southeast, yet all evidence indicates that Al mobilization is
21 most pronounced in the southern Appalachians. However, at the Whiteface Mountain site
22 selected for study because it was in a state of decline, soil solution levels there are lower than in
23 the Smokies, which are in a visibly obvious state of decline (e.g., no dieback other than the fir
24 killed by the balsam wooly adelgid, no needle yellowing, etc). Thus, Al mobilization constitutes
25 a situation worthy of further study (Johnson, 1992).
26 The simple calculations shown above give some idea of the importance of particulate
27 deposition in these forest ecosystems, but they cannot account for the numerous potential
28 feedbacks between vegetation and soils nor for the dynamics through time that can influence the
29 ultimate response. One way to examine some of these interactions and dynamics is to use
30 simulation modeling. The nutrient cycling model (NuCM) has been developed specifically for
31 this purpose and has been used to explore the effects of atmospheric deposition, fertilization, and
March 2001 4-72 DRAFT-DO NOT QUOTE OR CITE
-------
1 harvesting on some of the IPS sites (Johnson et al., 1993). The NuCM model is a stand-level
2 model that incorporates all major nutrient cycling processes (uptake, translocation, leaching,
3 weathering, organic matter decay, and accumulation).
4 Johnson et al. (1999) used the NuCM model to simulate the effects of reduced S, N, and
5 base cation (CB) deposition on nutrient pools, fluxes, soil, and soil solution chemistry in two
6 contrasting southern Appalachian forest ecosystems: (1) the red spruce and (2) Coweeta
7 Hardwood sites from the IPS project. The scenarios chosen for these simulations included
8 "no change", 50% N and S deposition, 50% CB deposition, and 50% N, S, and CB deposition
9 (50% N, S, CB). The NuCM simulations suggested that, for the extremely acid red spruce site,
10 S and N deposition is the major factor affecting soil solution Al concentrations and CB deposition
11 is the major factor affecting soil solution CB concentrations. The effects of S and N deposition
12 were largely through changes in soil solution SO42" and NO3" and, consequently, mineral acid
13 anion (MAA) concentrations rather than through changes in soils. This is illustrated in
14 Figures 4-13 and 4-14, which shows simulated soil solution mineral acid anions, base cations,
15 Al, and soil base saturation in B horizon from in the red spruce site. The 50% S and N scenario
16 caused reductions in soil solution SO42", NO3" and, therefore, MAA concentrations, as expected,
17 and this, in turn, caused short-term reductions in base cation concentrations. However, by the
18 end of the 24-year simulation, base cations in the 50% S, N scenario were nearly as high as in the
19 no change scenario because base saturation had increased and the proportion of cations as Al
20 decreased. The 50% CB scenario had virtually no effect on soil solution SO42", NO3" and,
21 therefore, MAA concentrations, as expected, but did cause a long-term reduction in base cation
22 concentrations. This was caused by a long-term reduction in base saturation (Figure 14). Thus,
23 the effects of CB deposition were solely through changes in soils rather than through changes in
24 soil solution MAA, as postulated by Driscoll et al. (1989). In the less acid Coweeta soil, base
25 saturation was high and little affected by scenario (not shown), Al was unimportant, and S and
26 N deposition had a much greater effect than CB deposition in all respects (Figure 15).
27 In summary, Johnson et al. (1999) found that the results of the red spruce simulations
28 support the hypothesis of Driscoll et al. (1989) in part: CB deposition can have a major effect on
29 CB leaching through time in an extremely acid system. This effect occurred through changes in
30 the soil exchanger and not through changes in soil solution MAA concentration. On the other
31 hand, S and N deposition had a major effect on Al leaching at the Noland Divide site. This
March 2001 4-73 DRAFT-DO NOT QUOTE OR CITE
-------
Red Spruce
500 —
400
Mineral Acid Anions
100 —
No Change
« 50%N,S
-A: 50% BC
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Base Cations
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure 4-13. Simulated soil solution mineral acid anions and base cations in the red spruce
site with no change, 50% N and S deposition, and 50% base cation
deposition. Redrawn from Johnson et al. (1999).
March 2001
4-74
DRAFT-DO NOT QUOTE OR CITE
-------
Red Spruce
500
1234567
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Year
Base Saturation
4 —
1234567
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Year
Figure 4-14. Simulated soil solution Al and soil base saturation in the red spruce site with
no change, 50% N and S deposition, and 50% base cation deposition.
Redrawn from Johnson et al. (1999).
March 2001
4-75
DRAFT-DO NOT QUOTE OR CITE
-------
Coweeta
100
80 —
Mineral Acid Anions
No Change
50% S,N
50% BC
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Base Cations
1234567
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Year
Figure 4-15. Simulated soil solution mineral acid anions and base cations in the Coweeta
site with no change, 50% N and S deposition, and 50% base cation
deposition. Redrawn from Johnson et al. (1999).
March 2001
4-76
DRAFT-DO NOT QUOTE OR CITE
-------
1 occurred primarily because of changes in soil solution MAA concentration. At the less acidic
2 Coweeta site, CB deposition had a minor effect on soils and soil solutions, whereas S and N
3 deposition had delayed but major effects on CB leaching because of changes in SO42" and MAA
4 concentrations.
5
6 Trace Element Effects
1 Trace metals are natural elements that are ubiquitous in small (trace) amounts in soils,
8 ground water and vegetation. Many are essential elements required for growth by plants and
9 animals as micronutrients. Naturally occurring surface mineralizations can produce metal
10 concentrations in soils and vegetation that are as high, or higher, than those in the air and
11 deposited near man-made sources (Freedman and Hutchinson, 1981). The occurrence and
12 concentration of trace metals in any ecosystem component depend on the sources of the metal via
13 the soil or as particulate. Even when air pollution is the primary source, continued deposition
14 can result in the accumulation of trace metals in the soil (Martin and Coughtrey, 1981). Many
15 metals are deposited into soils by chemical processes and are not available to plants (Saunders
16 and Godzik, 1986).
17 When aerial deposition is the primary source of metal particles, both the chemical form and
18 particle size deposited determine the heavy metal concentration in the various ecosystem
19 components (Martin and Coughtrey, 1981). Human activities introduce heavy metals into the
20 atmosphere and have resulted in the deposition of antimony, cadmium, chromium, copper, lead,
21 molybdenum, nickel, silver, tin, vanadium, and zinc (Smith, 1990c). Extensive evidence
22 indicates that heavy metals deposited from the atmosphere to forests accumulate either in the
23 richly organic forest floor or in the soil layers immediately below, areas where the activity in
24 roots and soil is greatest. The greater the depth of soil, the lower the metal concentration. The
25 accumulation of metal in the soil layers where the biological activity is greatest, therefore, has the
26 potential for being toxic to roots and soil organisms and interfering with nutrient cycling (Smith,
27 1990e). Though all metals can be directly toxic at high levels, only toxicity from copper, nickel,
28 and zinc have been documented frequently. Toxicity of cadmium, cobalt, and lead has been seen
29 only under unusual conditions (Smith, 1990e). Exposures at lower concentrations have the
30 potential, over the long-term, for interfering with the nutrient-cycling processes when they affect
31 mycorrhizal function.
March 2001 4-77 DRAFT-DO NOT QUOTE OR CITE
-------
1 Accumulation of heavy metals in litter presents the greatest potential for interference with
2 nutrient cycling. Accumulation of metals in the litter occurs chiefly around brass works and lead
3 and zinc smelters. There is some evidence that invertebrates inhabiting soil litter do accumulate
4 metals. Earthworms from roadsides were shown to contain elevated concentrations of cadmium,
5 nickel, lead, and zinc; however, interference with earthworm activity was not cited (Martin and
6 Coughtrey, 1981). It has been shown, however, that when soils are acidic, earthworm abundance
7 decreases and bioaccumulation of metals from soil may increase exponentially with decreasing
8 pH (Lokke et al, 1996). Organisms that feed on earthworms living in soils with elevated levels
9 of Cd, Ni, Pb, and Z for extended periods could accumulate lead and zinc to toxic levels (Martin
10 and Coughtrey, 1981). Increased concentrations of heavy metals have been found in a variety of
11 small mammals living in areas with elevated heavy metal concentrations in the soils.
12 Studies by Babich and Stotsky (1978) support the concept that increased accumulation of
13 litter in metal-contaminated areas is the result of effects on the microorganismal populations.
14 Cadmium toxicity to microbial populations was observed to decrease and prolong logarithmic
15 rates of microbial increase, to reduce microbial respiration and fungal spore formation and
16 germination, to inhibit bacterial transformation, and to induce abnormal morphologies. Also, the
17 effects of cadmium, copper, nickel, and zinc on the symbiotic activity of fungi, bacteria, and
18 actinomycetes were reported by Smith (1991). The formation of mycorrhizae by Glomus
19 mosseae with onions was reduced when zinc, copper, nickel, or cadmium was added to the soil.
20 The relationship of the fungus with white clover, however, was not changed. It was suggested
21 that the effect of heavy metals on vesicular-arbuscular mycorrhizal fungi will vary from host to
22 host (Gildon and Tinker, 1983). Studies with ericoid plants indicated that, in addition to Calluna
23 vulgar is, mycorrhizae also protect Vaccinium macrocarpa and Rhodendron ponticum from heavy
24 metals (Bradley et al., 1981). Heavy metals tend to accumulate in the roots, and shoot toxicity is
25 prevented.
26 The effects of sulfur deposition on litter decomposition in the vicinity of smelters also must
27 be considered. Metal smelters emit SO2 as well as heavy metals. Altered litter decomposition
28 rates have been well documented near SO2 sources (Prescott and Parkinson, 1985). The presence
29 of sulfur in litter has been associated with reduced microbial activity (Bewley and Parkinson,
30 1984). Additionally, the effects on symbiotic activity of fungi, bacteria and actinomycetes were
31 reported by Smith (1990b).
March 2001 4-78 DRAFT-DO NOT QUOTE OR CITE
-------
1 The potential pathways of accumulation of trace metals in terrestrial ecosystems, as well as
2 the possible consequences of trace metal deposition on ecosystem functions is summarized in
3 Figure 4-16. The generalized trophic levels found in an ecosystem and the various physiological
4 and biological processes that could be affected by trace metals are shown in the figure.
5 Reduction in physiological processes can affect productivity, fecundity, and mortality (Martin
6 and Coughtrey, 1981). Therefore, any effects on structure and function of an ecosystem are
7 likely to occur through the soil and litter (Tyler, 1972).
8 Trace metals deposited from the atmosphere to forests accumulate either in the richly
9 organic forest floor or in the soil layers immediately below, layers where greatest biological
10 activity occurs. The shallow-rooted species plant species are those most likely to take up metals
11 from the soil (Martin and Coughtrey, 1981).
12 Certain species of plants are tolerant of metal contaminated soils (e.g., soils from mining
13 activities) (Antonovics et al., 1971). Certain species of plants also have been used as
14 bioindicators of metals (e.g., Astragalius is an accumulator of selenium). The sources of both
15 macroelements and trace metals in the soil of the Botanical Garden of the town of Wroclow,
16 Poland, were determined by measuring the concentrations of the metals in Rhododendron
17 catawbiense, Ilex aquifolium, and Mahonia aquifolium growing in the garden and comparing the
18 results with the same plant species growing in two other botanical gardens in nonpolluted areas.
19 Air pollution deposition was determined as the source of metals in plants rather than the soil
20 (Samecka-Cymerman and Kempers, 1999).
21 Biological accumulation of metals through the plant-herbivore and litter-detrivore chains
22 can occur. A study of the accumulation of cadmium, lead, and zinc concentrations in
23 earthworms suggested that cadmium and zinc were concentrated, but not lead. Studies indicate
24 that heavy metal deposition onto the soil, via food chain accumulation, can cause excessive
25 levels and toxic effects in certain animals. Cadmium appears to be relatively mobile within
26 terrestrial food chains; however, the subsequent mobility of any metal after it is ingested by a
27 herbivorous animal depends on the site of accumulation within body tissues. Although food
28 chain accumulation may not in itself cause death, it can reduce the breeding potential in a
29 population (Martin and Coughtrey, 1981).
30 In actual case studies, it was observed that the deposition of copper and zinc particles
31 around a brassworks resulted in an accumulation of incompletely decomposed litter. In one
March 2001 4-79 DRAFT-DO NOT QUOTE OR CITE
-------
g.
to
o
o
-------
1 study, litter accumulation was reported up to 7.4 km from the stack of a primary smelter in
2 southeastern Missouri. Similar results were reported around a metal smelter at Avonmouth,
3 England. In the latter case, litter accumulation was associated closely with concentrations
4 specifically of cadmium, as well as with those of lead, copper, and zinc (Martin and Coughtrey,
5 1981). Experimental data (using mesh bags containing litter) supports the hypothesis that
6 reduced decomposition occurs close to heavy metal sources.
7 Accumulations of metals emitted in particulate matter also were reported in soil litter close
8 to a metal smelter at Palmerton, PA, in 1975 and 1978. The continued presence of cadmium,
9 lead, zinc, and copper in the upper soil horizons (layers) were observed 6 years after the smelter
10 terminated operation in 1980. Metal levels were highest near the smelter. The relationship of
11 decreasing amounts of metal in body tissues also held true for amphibians and mammals. Levels
12 of cadmium in kidneys and liver of white-tailed deer (Odocoileus virginaus) were five times
13 higher at Palmerton than in those collected 180 km southwest downwind. The abnormal
14 amounts of metal in the tissues of terrestrial vertebrates and the absence or low abundance of
15 wildlife at Palmerton indicated that ecological processes within 5 km of the zinc smelter
16 continued to be markedly influenced even 6 years after its closing (Storm et al., 1994).
17 The effects of lead in ecosystems are discussed in the Air Quality Criteria for Lead
18 (U.S. Environmental Protection Agency, 1986b). Studies have shown that there is cause for
19 concern in three areas where ecosystems may be extremely sensitive to lead: (1) delay of
20 decomposition because the activity of some decomposer microorganisms and invertebrates is
21 inhibited by lead, (2) subtle shifts toward plant populations tolerant of lead, and (3) lead in the
22 soil and on the surfaces of vegetation circumvent the processes of biopurification. The problems
23 cited above arise because lead is deposited on the surface of vegetation, accumulates in the soil,
24 and is not removed by the surface and ground water of the ecosystem (U.S. Environmental
25 Protection Agency, 1986b).
26
27 4.2.3 Ecosystem Goods and Services and Their Economic Valuation
28 Human existence on this planet depends on ecosystems and the services and products they
29 provide. The essential services and products provided by the planet's collective biodiversity (the
30 earth's flora, fauna, and microorganisms) are clean air, clean water, clean soil, and clean energy
31 (Table 4-6). Today, governments around the world pursue a "bottom line" that driven is by an
March 2001 4-81 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-6. PRIMARY GOODS AND SERVICES PROVIDED BY ECOSYSTEMS
Ecosystem
Goods
Services
Agroecosvstems
Coastal ecosystems
Forest ecosystems
Freshwater
Grassland
ecosystems
1 Food crops
1 Fiber corps
1 Crop genetic resources
1 Fish and shellfish
• Fishmeal (animal feed)
1 Seaweeds (for food and industrial
use)
•Salt
1 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
1 Fish
1 Hydroelectricity
1 Genetic resources
1 Livestock (food, game, hides, and
fiber)
• Drinking and irrigation water
1 Genetic resources
• Maintain limited watershed functions (infiltration,
flow control, and partial soil protection)
• Provide habitat for birds, pollinators, and soil
organisms important to agriculture
• Sequester atmospheric carbon
• Provide employment
• Moderate storm impacts (mangroves, barrier
islands)
• Provide wildlife (marine and terrestrial) habitat
and breeding areas/hatcheries/nurseries
• Maintain biodiversity
• Dilute and treat wastes
• Provide harbors and transportation routes
• Provide human and wildlife habitat
• Provide employment
• Contribute aesthetic beauty and provide recreation
• Remove air pollutants, emit oxygen
• Cycle nutrients
• Maintain array of watershed functions (infiltration,
purification, flow control, soil stabilization)
• Maintain biodiversity
• Sequester atmospheric carbon
• Moderate weather extremes and impacts
• Generate soil
• Provide employment
• Provide human and wildlife habitat
• Contribute aesthetic beauty and provide recreation
• Buffer water flow (control timing and volume)
• Dilute and carry away wastes
• Cycle nutrients
• Maintain biodiversity
• Provide aquatic habitat
• Provide transportation corridor
• Provide employment
• Contribute aethetic beauty and provide recreation
• Maintain array of watershed functions (infiltration,
purification, flow control, and soil stabilization)
• Cycle nutrients
• Remove air pollutants and emit oxygen
• Maintain biodiversity
• Generate soil
• Sequester atmospheric carbon
• Provide human and wildlife habitat
• Provide employment
• Contribute aesthetic beauty and provide recreation
Source: World Resources (2000-2001).
March 2001
4-82
DRAFT-DO NOT QUOTE OR CITE
-------
1 economy that is disconnected from the natural world and is fundamentally destructive of local
2 ecosystems (Suzuki, 1997). For this reason, human society needs to be reconnected to the
3 biologically diverse ecosystems and the natural world of which they are a part (Suzuki, 1997).
4 There is a need to understand the biodiversity that encompasses all levels of biological
5 organization, including populations, individuals, species and ecosystems (Wilson, 1997).
6 Populations, geographical entities within a species of organisms, usually distinguished
7 ecologically or genetically, are essential to the conservation of species diversity. Their number
8 and size influence the probability of the future existence of the entire species (Hughes et al.,
9 1997). The number, biodiversity, structure, and functions of ecosystem populations, provide
10 ecosystem products (goods) and services. For any given population, the number of individuals,
11 the genetic variation between individuals, and the area occupied affects ecosystem functioning
12 and the delivery of ecosystem services and other benefits provided by that population (Hughes,
13 et al., 1997). Loss of population diversity means loss of the benefits described in Table 4-6 and,
14 in particular, with time, the loss of the life-support systems on which humanity relies (Hughes
15 etal., 1997).
16 Attempts have been made to value biodiversity and the world's ecosystem services and
17 natural capital (Pimentel et al., 1997; Costanza et al., 1997). Pimentel et al. (1997) estimated
18 economic and environmental benefits for services contributed from all biota (biodiversity) in the
19 United States, including their genes, at $319 billion per year. Costanza et al. (1997) have
20 estimated the total value of ecosystem services by biome for the entire bioshere. Ecosystems
21 provide at least $33 trillion worth of services annually. Approximately, 63% of the estimated
22 value is contributed by marine ecosystems ($20.9 trillion per year), most of which comes from
23 coastal ecosystems ($10.6 trillion per year). About 38% of the estimated value comes from
24 terrestrial ecosystems, mainly from forests ($4.7 trillion per year) and wetlands ( $4.9 trillion per
25 year). Costanza et al. (1997) state that it may never be possible to make a precise estimate of the
26 services provided by ecosystems. Their estimates, however, indicate the relative importance of
27 ecosystem services.
28 Heal (2000), however, feels that attempts to value ecosystems and their services are
29 probably misplaced. "Economics cannot estimate the importance of natural environments to
30 society: only biology can do that" (Heal, 2000). The role of economics is to help design
31 institutions that will provide incentives to the public and policy makers for the conservation of
March 2001 4-83 DRAFT-DO NOT QUOTE OR CITE
-------
1 important natural systems and for mediating human impacts on the biologically diverse
2 ecosystems and the biosphere so that they are sustainable. The approach of Harwell et al. (1999)
3 also deals with the need to understand human impacts on ecosystems so that ecosystem
4 management can define what ecological conditions are desired. Further, they state that the
5 establishment of ecological goals involves a close linkage between scientists and decision
6 makers, in which science informs decision makers and the public by characterizing the ecological
7 conditions that are achievable under particular management regimes. Decision makers then can
8 make choices that reflect societal values, including issues of economics, politics, and culture.
9 For management to achieve their goals, the general public, scientific community, resource
10 managers, and decision makers need to be routinely apprised of the condition or integrity of
11 ecosystems, so that ecological goals may be established (Harwell et al., 1999).
12 The above assessment of new information leads to the clear conclusion that atmospheric
13 PM at levels currently found in the United States has the potential to alter ecosystem structure
14 and function in ways that may reduce their ability to meet societal needs. The possible direct
15 effects of airborne PM on individual plants were discussed in Section 4.2.1 above. The major
16 impacts of airborne PM on ecosystems, however, are the indirect effects on plant populations that
17 occur through the soil and affect the cycling of nutrients necessary for plant growth and vigor, as
18 discussed in Section 4.2.2. By altering the cycling of nitrogen, nitrogen deposition changes the
19 biodiversity of ecosystems and their functioning and, by altering the vigor of forest tree stands,
20 alters forest succession. Also, nitrogen deposition in combination with the deposition of sulfur in
21 the form of acid rain alters the biogeochemical cycling of soil mineral nutrients and changes the
22 biodiversity and functioning of forest ecosystems. The changes in the ability of forest vegetation
23 and soil microorganisms to utilize nutrients results in the leaching of nitrates and other minerals
24 from the soils. The nitrate and mineral runoff impacts coastal and aquatic ecosystems and, thus,
25 influences the services important to human life provided by these ecosystems as well (Table 4-6).
26
27
28
29
March 2001 4-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.3 EFFECTS ON VISIBILITY
2 4.3.1 Introduction
3 Visibility is defined as the degree to which the atmosphere is transparent to visible light and
4 the clarity (transparency) and color fidelity of the atmosphere (National Research Council, 1993).
5 Visibility impairment is defined as any humanly perceptible change in visibility (light extinction,
6 visual range, contrast, or coloration). Visual range is described as the farthest distance at which a
7 large black object can be distinquished against the horizontal sky (U.S. Environmental Protection
8 Agency, 1979). For regulatory purposes, visibility impairment is classified into two principal
9 forms: (1) "reasonably attributable" impairment, attributable to a single source or small group of
10 sources and (2) regional haze, described as any perceivable change in visibility (light extinction,
11 visual range, contrast, or coloration) from which would have existed under natural conditions
12 that is caused predominantly by a combination of many sources over a wide geographical area
13 (U.S. Environmental Protection Agency, 1999).
14 The objective of the visibility discussion in this section is to summarize the linkage
15 between air pollution, in particular particulate matter, and visibility. This section summarizes the
16 information discussed in the previous 1996 PM air quality criteria document (PM AQCD) and
17 includes additional relevant information available since publication of that document. For a
18 more detailed discussion on visibility, the reader is referred to the earlier PM AQCD entitled, Air
19 Quality Criteria for Parti culate Matter (U.S. Environmental Protection Agency, 1996a), the
20 Recommendations of the Grand Canyon Visibility Transport Commission (Grand Canyon
21 Visibility Transport Commission, 1996), the National Research Council (National Research
22 Council, 1993), the National Acid Precipitation Assessment Program (Trijonis et al., 1991), and
23 the U.S. Environmental Protection Agency (1995a).
24
25 4.3.2 Factors Affecting Atmospheric Visibility
26 4.3.2.1 Anthropogenic Pollutants
27 Visibility impairment may be connected to air pollutant properties, including size
28 distribution, aerosol chemical composition, and relative humidity. In the United States, visibility
29 impairment is caused by sulfate and nitrate particles in the 0.1- to 1.0-micron (//m) range, and
30 organic aerosols, carbon soot, and crustal dust. Generally, sulfates are responsible for most of
March 2001 4-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 the visibility impairment in the United States, as measured by light extinction, accounting for
2 approximately two-thirds of the light extinction in the eastern United States. Sulfate
3 concentrations are higher in summer months than in the wintertime (Malm et al., 1994).
4 Exceptions to the sulfate-related effects on visibility include California, where the primary cause
5 of visibility effects is ambient nitrate, and Alaska, where visibility impairment is caused by fine
6 soil plus coarse mass (classified as coarse extinction) or organics, thought to be from natural
7 sources (Sisler and Cahill, 1993).
8
9 4.3.2.2 Human Vision
10 Human vision is one of the factors that affects the way an object is viewed. Vision is the
11 response to the electromagnetic radiation that enters the eye between wavelengths of 400 and
12 700 nm. The cones, a receptor cell in the retina, govern visibility interpretations.
13 The eye perceives the lightest and brightest object in a scene as white, and determines the
14 color of other objects by comparison. The ability of the eye to perceive contrasts, the degree of
15 color difference between the lightest and darkest object in a scene, changes in response to the
16 illumination and setting. The effects of illumination on visibility are discussed in the following
17 subsection. At increasing distances the brightness of a target or object will approach the
18 brightness of the horizon making the target indistinquishable from the horizon, hence, visual
19 range.
20
21 4.3.2.3 Characteristics of the Atmosphere
22 The appearance of a distant object is determined by illumination of the sight path by the
23 direct rays of the sun, diffused skylight, light that has been reflected from the surface of the Earth
24 (path radiance or air light), and the light reflected from the object itself. Some of the light in the
25 sight path is absorbed or scattered towards the observer. The remaining light is absorbed or
26 scattered in other directions. The portion of scattered light from the object being viewed that
27 reaches the observer is the transmitted radiance. The radiance seen by the observer looking at a
28 distant object is the sum of the transmitted radiance and the path radiance. Figure 4-17
29 demonstrates light being absorbed and scattered by the atmosphere and a target object.
30 On a clear day when the sun is high in the sky, 80 to 90% of the visible solar radiation
31 reaches the surface of the Earth without being scattered or absorbed. Rayleigh scattering by
March 2001 4-86 DRAFT-DO NOT QUOTE OR CITE
-------
OBSERVER
Figure 4-17. Light reflected from a target toward an observer. The intervening
atmosphere scatters a portion of this light out of the sight path and scatters
light from the sun into the sight path. Some particles and gases also absorb a
portion of the light from the target. The light scattered into the sight path
increases with distance from the target, whereas the light transmitted from
the target decreases with distance from the target. The visual range is the
closest distance between the target and the observer at which the transmitted
light no longer can be distinguished from the light scattered into the sight
path.
Source: Watson and Chow (1994).
1 gases is the major component of light extinction in relatively unpolluted areas. Mie scattering is
2 the scattering of all visible wavelengths equally (Shodor Education Foundation, Inc., 1996). It is
3 the attenuation of light in the atmosphere by scattering because of particles of a size comparable
4 to the wavelength of the incident light (National Acid Precipitation Assessment Program, 1991).
5 The term, multiple scattering, is used when light is scattered more than once in a turbid medium.
6 The great majority of light absorption by particles is caused by black carbonaceous particles,
7 assumed to be elemental carbon, that are products of incomplete combustion (Rosen et al., 1978;
8 Japar et al., 1986; Watson and Chow, 1994). Malm et al. (1996) suggested that organic carbon
March 2001
4-87
DRAFT-DO NOT QUOTE OR CITE
-------
1 also acts to scatter and absorb light. The estimated natural visibility for the east and west is 60 to
2 90 mi and up to 140 mi, respectively. Current visibility conditions range from 18 to 40 mi in the
3 rural east to 35 to 90 mi in the rural west (U.S. Environmental Protection Agency, 2000b).
4 At the surface, a variable fraction of the solar radiation is reflected back upwards, referred
5 to as surface reflectance or the albedo, illuminating the atmosphere from above and below. The
6 amount of solar radiation reflected depends on the color of the terrain. Dark-colored terrain
7 reflects less radiation than light-colored terrain.
8 Visibility within a sight path longer than approximately 100 km (60 mi) is affected by
9 changes in the properties of the atmosphere over the length of the sight path. The atmosphere
10 generally will not have uniform optical properties over distances greater than a few tens of
11 kilometers. Air quality within a sight path can affect the illumination of the sight path by
12 scattering or absorbing solar radiation before it reaches the Earth's surface. The light-extinction
13 coefficient, oext, is a measure of the fraction of light that is lost as it travels through the
14 atmosphere. The light-extinction coefficient is the sum of the light-scattering coefficient, oscat,
15 and the light-absorption coefficient, oabs, expressed in units of inverse lengths of the atmosphere
16 (megameters ; Mm"1). Typical extinction coefficients range from 0.01 km"1 (10 Mm"1) in
17 relatively clean air to ~ 1000 Mm"1 in highly polluted areas (Watson and Chow, 1994).
18 The light-extinction coefficient can be divided into coefficients for the following
19 components:
20 oag, light absorption by gases,
21 osg, light scattering by gases (Rayleigh scattering),
22 oap, light absorption by particles, and
23 osp, light scattering by particles.
24 Light scattering by particles, osp, can be divided to indicate scattering by coarse and fine particles:
25 osfp, light scattering by fine particles and oscp, light scattering by coarse particles.
26
27 4.3.3 Optical Properties of Particles
28 Visibility impairment is typically caused by fine particles. Fine particles are small enough
29 in comparison with the wavelength of visible light that their optical properties are nearly the
30 same as those of homogeneous spheres of the same volume and average index of refraction.
31 Accordingly, Mie equations (Mie, 1908; Kerker, 1969), for calculating the optical properties of
March 2001 4-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 homogeneous spheres also may be used to calculate the optical properties of fine particles with
2 the only uncertainties being in the fine particle size distribution and index of refraction (Richards,
3 1973). However, within the range of indices of refraction that most commonly occur in
4 atmospheric fine particles, the results of Mie calculations can be scaled to account for the effect
5 of the index of refraction. Coarse particles have less of an impact on visibility than do fine
6 particles. However, in most actual cases, the dominant uncertainty in using the optical properties
7 for coarse particles calculated with Mie equations is the uncertainty in the particle size
8 distribution. Uncertainties exist in the use of Mie calculations for calculating light absorption for
9 course particles because the refractive index of the particle is generally not known, and the
10 light-absorbing particles are not spherical in shape, making the calculated light absorption
11 efficiency factor less reliable. Also, light absorption by elemental carbon particles can be
12 reduced when the particle is covered by some chemical species (Dobbins et al., 1994).
13 Conversely, light absorption by carbon particles can be enhanced when coated with a
14 nonabsorbing refractive material such as ammonium sulfate (Fuller et al., 1999).
15 The output of the Mie calculations includes efficiency factors for extinction, Qext,
16 scattering, Qscat, and absorption, Qabs. The Qext, Qscat, and Qabs give the fraction of the incident
17 radiation falling on a circle with the same diameter as the particle that is either scattered or
18 absorbed. The light scattering or absorption efficiency factor (in units of m2/g) is the change in
19 the light scattering or absorption efficiencies per unit change in mass of the fine particle
20 constituent. The scattering and absorption efficiencies are determined by estimating the size
21 distribution of each particle. The results of the calculations for the light absorption efficiencies
22 contains significant uncertainties because the components of the index of refraction is generally
23 unknown and the light-absorbing particles are frequently chained agglomerates that do not have a
24 spherical shape. Multiplying the values of the light-scattering efficiency factor by the aerosol
25 volume concentration (in units of //m3/cm3) gives the value of the light-scattering coefficient, osp,
26 (in units of Mm"1) for these particles.
27 Richards et al. (1991) reported a scattering efficiency for fine particles of ammonium
28 sulfate of 1.2 m2/g based on Mie calculations. The value was in agreement with the value
29 determined using the integrating nephelometer readings and the sulfate concentrations. Sulfate
30 scattering efficiencies have been reported to increase by a factor of two when the size distribution
31 went from 0.15 to 0.5 //m (McMurry et al., 1996). The calculated scattering efficiencies for
March 2001 4-89 DRAFT-DO NOT QUOTE OR CITE
-------
1 sulfates were 4.1 m2/g for 100% mass removal and 3.4 and 5.6 m2/g for 25% mass removal.
2 Calculated scattering efficiencies for carbon particles ranged from 0.9 to 8.1 m2/g (Zhang et al.,
3 1994; Sisler and Malm, 2000; Sloane et al., 1991). A scattering efficiency of 1.0 and 0.6 m2/g
4 was reported for soil and coarse mass, respectively (Trijonis et al., 1987).
5 Scattering efficiencies of 2.4 and 3.1 m2/g for fine particles were reported by White et al.
6 (1994) and Waggoner et al. (1981), using an integrating nephelometer. Coarse particle scatter
7 less light, resulting in lower scattering efficiencies. Scattering efficiencies for coarse particles
8 ranged from 0.4 to 0.6 m2/g, based on integrating nephelometer readings (White et al., 1994;
9 Trijonis et al., 1987; White and Macias, 1990; Watson et al., 1991).
10 Absorption efficiencies for elemental carbon particles have been reported to range from
11 9 to 10 m2/g (Japar et al., 1984; Adams et al., 1989; Sloane et al., 1991). Based on a review of
12 the available data, Horvath (1993) reported that measured light absorption efficiencies for light
13 absorbing carbon ranges from 3.8 to 17 m2/g. According to Horvath (1993), calculated
14 absorption efficiencies are too high, ranging from 8 to 12 m2/g for monodispersed carbon
15 particles. Fuller et al. (1999) suggested that isolated spheres of light absorbing carbon have a
16 specific absorption of less than 10 m2/g. Light absorption by carbon particles only will be greater
17 than 10 m2/g if the particles are internally mixed and the occluding particles are sufficiently large.
18 Absorption values for graphitic and amorphous carbon spheres for primary sizes typical of diesel
19 soot are around 5 m2/g. Light absorption by aggravated carbon at visible wavelengths is
20 enhanced by no more than 30% and diminishes if encapsulated by a nonabsorbing aerosol.
21 Malm et al. (1996) suggested a combined scattering and absorption efficiency of 10 m2/g for
22 organic carbon.
23 Light-extinction budgets may be estimated using the light extinction efficiency and the
24 measured species concentrations. Light-extinction budgets estimate the fraction of the total light
25 extinction contributed by each chemical species in the sight path; however, the values obtained
26 will depend on the assumptions used (Malm et al., 1996; Lowenthal et al., 1995; Sisler and
27 Malm, 1994).
28
29
30
March 2001 4-90 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.3.4 Effect of Relative Humidity on Particle Size and Light-Scattering
2 Properties
3 Ambient particles contain water, even on relatively dry days. As the relative humidity
4 increases, the particle absorbs more water and increases in size and volume. It is the increase in
5 particle size and volume that acts to increase the light scattering properties of most particles
6 (Malm etal., 1996).
7 Ambient particles are a mixture of chemical compounds. The amount of increase in
8 particle size with increasing relative humidity is dependent on the particle composition (Zhang
9 et al., 1993). Available data indicate that particles containing ammonium salts are in a liquid
10 solution at relative humidities above 80%. Particles containing inorganic salts and acids are
11 more hygroscopic than particles composed primarily of organic species (Day et al., 1996;
12 McMurry and Stolzenburg, 1989; Saxena et al., 1995; Zhang et al., 1993, 1994; Sloane et al.,
13 1991). Particles containing the more hygroscopic salts and acid species deliquesce and undergo
14 changes in particle size in response to changes in relative humidity. For sulfate and nitrate
15 aerosols, light-scattering properties are similar for all mixture types and compositions, as long as
16 there is the same particle size distribution (Tang, 1997). Saxena et al. (1995) found that the
17 hygroscopic properties of inorganic particles can be altered positively or negatively in the
18 presence of organics. Based on limited data, nonurban organics were found to add to water
19 absorption by inorganics, whereas the urban organics diminished the absorption of water by
20 inorganic particles at relative humidities of 80 to 93%. Figure 4-18 demonstrates the humidity
21 effect on the scattering coefficients for several internally mixed (individual particles containing
22 one or more species) and externally mixed (species that co-exist as separate particles) aerosols.
23 The total scattering computed for an aerosol is relatively insensitive to whether the sample is
24 internally or externally mixed (Malm et al., 1997). Figure 4-19 demonstrates changes in the
25 scattering coefficient ratio, ospw/aspd, where aspw is the scattering coefficient under humid
26 conditions, and ospd is the scattering coefficient under dry conditions. The figure demonstrates
27 that light scattering is a function of relative humidity and chemical composition. The monitoring
28 data were generated as part of the Southeastern Aerosol and Visibility Study (Day et al., 2000).
29 There is also a relative humidity-related effect on the scattering efficiency. Ammonium
30 sulfate fine-particle-scattering efficiency varied from 1.5 to 4.5 m2/g, with low relative humidity
31 and median particle sizes ranging from 0.07 to 0.66 //m (McMurry et al., 1996). Sloane et al.
March 2001 4-91 DRAFT-DO NOT QUOTE OR CITE
-------
0.1
CL
-------
T3
Q.
Q.
'o
it
75%. The calculated
scattering efficiencies for organic carbon ranged from 3.81 m2/g when the relative humidity was
March 2001
4-93
DRAFT-DO NOT QUOTE OR CITE
-------
1 <63% to 6.9 m2/g at relative humidities above 75% (Omar et al., 1999). A more detailed
2 discussion of the effects of relative humidity on the size distribution of ambient particles appears
3 in Chapter 2 of this document.
4
5 4.3.5 Measures of Visibility
6 4.3.5.1 Human Observations
7 The National Weather Service has in recent decades recorded hourly visibility readings at
8 all major airports in the United States based on human observations of the most distant targeted
9 object's perceivability. Human observation of visibility, although providing a historical record of
10 visibility readings in the United States, are dependent on the individual and the availability of a
11 target and generally are related poorly to air quality.
12
13 4.3.5.2 Light-Extinction Coefficient and Parameters Related to the Light-Extinction
14 Coefficient
15 The most frequently used indicator for visibility characterization for air quality is the
16 light-extinction coefficient because it is closely linked to air quality (U.S. Environmental
17 Protection Agency, 1996a). Various meteorological conditions (moisture and cloud cover) can
18 affect the light-extinction coefficient; however, these effects can be minimized (Husar et al.,
19 1994; Blandford, 1994; Mercer, 1994). The light-extinction coefficient can be measured directly
20 using a transmissometer (Molenar et al., 1990, 1992) or can be estimated by measuring the
21 components of light extinction (scattering and absorption) and calculating the sum (Malm et al.,
22 1994).
23 The light-extinction coefficient is the quantitative measure of haziness, defined as
24 oext = K/visual range, where K is the Koschmieder constant. The value of K is determined both
25 by the threshold sensitivity of the human eye and the initial contrast of the visible object against
26 the horizon sky.
27 The visual range may be calculated from the light-extinction coefficient using the
28 Koschmieder equation by assuming the atmosphere and the illumination over a sight path in the
29 daytime is uniform, and that the threshold contrast is 2% (Katsev and Zege, 1994; Koschmieder,
30 1924). These assumptions are, however, invalid for visual ranges greater than 100 km (U.S.
31 Environmental Protection Agency, 1996a). Visual range is an understandable, and for most
March 2001 4-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 purposes, an appropriate measure of the optical environment. It has the disadvantage of being
2 related inversely to aerosol concentration.
3
4 Visual Range = 3.91/oext
5
6 The deciview index is an atmospheric haze index that expresses uniform changes in
7 haziness in common increments from pristine conditions to extremely visibility impaired
8 environments. The deciview scale is linear with perceived visual changes, starting near zero for
9 a pristine atmosphere (particle-free) at a 1.8-km elevation, and increases with increasing
10 haziness. The deciview index may be calculated from the light-extinction coefficient for green
11 light. For consistency, a Rayleigh scattering value of 10 Mm"1 is used.
12
13 dv = 10 logw (oext /JO Mm -1)
14
15 Under ideal conditions, a just noticeable change in the light-extinction coefficient should
16 represent a one or two deciview change in the deciview scale, about a 10 to 20% change in the
17 extinction coefficient. Any change in the deciview scale should have a change of similar
18 magnitude in the visual appearance of the scene in cases where the assumptions used to develop
19 the deciview scale are met (Pitchford and Malm, 1994; Sisler and Malm, 2000). Figure 4-20
20 illustrates the relationship of light extinction in Mn"1, deciview index, and visual range in
21 kilometers. Although the deciview is related to extinction, it is scaled in such a way that is
22 perceptually correct (Fox et al., 1999).
23
24
Extinction (Mm"1) 10 20 30 40 50 70 10° 20° 30° 40° 50° 70° 100°
Deciviews (dv)
I
0
I
I
7
I
I
11
I
I
14
I
i i mi
16 19 23
I Mill
I
30
I
I
34
I
I
37
I
I
39
I
I I III
42 46
I Mil
Visual Range (km) 400 2oo 130 100 so eo 40 20 13 10 a 6 4
Figure 4-20. Comparison of extinction (Mn"1) and visual range (km).
Source: Fox et al. (1999).
March 2001 4-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 Figures 4-21a,b illustrate a change in deciview scale based on reconstructed extinction
2 coefficients for the Great Plains Region (Badlands) using data from the Interagency Monitoring
3 of Protected Visual Environments Network (IMPROVE). Details about the IMPROVE network
4 appears in Section 4.5.6. The data are sorted by year into three groups based on the cumulative
5 frequency of occurrence of PM25: best visibility days (10th percentile), median (50th percentile),
6 and worst visibility days (90th percentile) (Sisler and Malm, 2000).
7 Richards (1999) suggests that the deciview index may not be a good tool for measuring
8 visibility impairment in areas restricted by boundaries. The deciview index is, however, suitable
9 for measuring visibility conditions over a broad geographic region, which is consistent with the
10 definition of regional haze, uniform haze caused by pollutant sources over broad areas (U.S.
11 Environmental Protection Agency, 1999).
12
13 4.3.5.3 Light-Scattering Coefficient
14 Light-scattering by particles has been reported to account for 68 to 86% of the total
15 extinction coefficient in several cities in California (Eldering et al., 1994). The light-scattering
16 coefficient is closely linked to fine particle concentrations, making it a good tool for determining
17 small particle-related effects on visibility. When the light-scattering coefficient is increased,
18 visibility is impaired because the transmitted radiance is decreased and the path radiance is
19 increased. (See discussion in the previous sections on transmitted radiance and path radiance.)
20 The light-scattering coefficient can be measured directly with an open and enclosed integrating
21 nephelometer and a forward scatter visibility monitor (Molenar et al., 1992; National Oceanic
22 and Atmospheric Administration, 1992). The light-scattering coefficient also may be calculated
23 using analytical approximations of the particle size distributions, log normal size distributions, or
24 sectional particle size distributions. In the sectional approach, the size composition distribution
25 is represented by a set of particle size sections. The chemical composition of each size section is
26 assumed to be the same (Wu et al., 1996).
27
28 4.3.5.4 Fine Particulate Matter Concentrations
29 The influence of particles on visibility degradation is dependent on the particle
30 composition, solubility, and size (Pryor and Steyn, 1994). Fine particle species have been
31 classified into five major types: (1) sulfates, (2) nitrates, (3) organics, (4) light absorbing carbon,
March 2001 4-96 DRAFT-DO NOT QUOTE OR CITE
-------
Fine Mass PM
88
90
91 92 93 94
Sample Year
95 96
Visibility Impairment
88 89 90 91 92 93 94
Sample Year
95 96
• 10th Percentile
•50th Percentile
•90th Percentile
Figure 4-21a,b. Plots of the 10th, 50th, and 90th percentile groups for PM2 5 and deciview
at the Badlands National Park. The sample year began in March of each
year.
Source: Sisler and Malm (2000).
March 2001
4-97
DRAFT-DO NOT QUOTE OR CITE
-------
1 and (5) soil (Malm et al., 1994). The coefficient of light-scattering by fine particles is primarily
2 responsible for visibility impairment making fine particle concentration a suitable indicator of
3 particle related effects on visibility. Several studies have demonstrated a relationship between
4 the coefficient for light-scattering by particles, measured using an integrating nephelometer, and
5 fine particle concentrations (Dattner, 1995; Waggoner and Weiss, 1980; Waggoner et al., 1981;
6 White et al., 1994). Figure 4-22 demonstrates visual range based on particle concentrations and
7 extinction efficiencies for road dust and sulfate.
400
300 -
CD
ro
o:
"ro
200 -
100 -
Sulfate A Road Dust
change in visibility based on 1 ug/m
difference in concentration
10 20 30 40 50 60 70
Concentration (ug/m3)
80
90
100
Figure 4-22. Reduction in visual range as a function of increasing fine (sulfate) and coarse
(dust) particle concentrations.
Source: Watson and Chow (1994).
4.3.5.5 Discoloration
Discoloration may be used as a quantitative measurement of atmospheric color changes in
urban hazes. Atmospheric color changes is a component of plume visibility models. The color
March 2001
4-98
DRAFT-DO NOT QUOTE OR CITE
-------
1 of haze will primarily depend on the scene used and human vision. For plume visibility, the
2 threshold for perception of color differences depend on the apparent width of the plume and is
3 greater for color patches separated by sharp edges. Methods for specifying the colors of hazes
4 include the CIE XYZ system of color matching, the Hunt94 color-appearance model, and the
5 visual colorimeter, VISUAL colorimeter for Atmospheric Research (Trijonis et al., 1991;
6 Mahadev and Henry, 1999).
7
8 4.3.6 Visibility Monitoring Methods and Networks
9 Visibility monitoring studies measure the properties of the atmosphere either at the sampler
10 inlets (point measurements), as is the case with air quality measurements, or by determining the
11 optical properties of a sight path through the atmosphere (path measurements). Instrumental
12 methods for measuring visibility are generally of three types: (1) direct measurement of light
13 extinction of a sight path using a transmissometer, (2) measurement of light scattering at one
14 location using an integrating nephelometer, and (3) measurement of ambient aerosol mass
15 concentration and composition (Mathai, 1995).
16 The largest instrumental visibility monitoring network in the United States is designed to
17 provide real-time data for runway visibility to aid in controlling airport operations.
18 An automated observing system, Automated Surface Observing System (ASOS), is being placed
19 at airports around the country. This monitoring network is sponsored by the National Weather
20 Service, the Federal Aviation Administration, and the Department of Defense. More than
21 500 airports are currently commissioned and an additional 500 are expected to come online in the
22 next few years.
23 The visibility sensor, instead of measuring how far one can see, measures the clarity of the
24 air using a forward scatter visibility meter. The forward scatter meter was found to correlate
25 fairly well with extinction coefficient measurements from the Optec Transmissometer. The
26 clarity is then converted to what would be perceived by the human eye using a value called
27 Sensor Equivalent Visibility (SEV). Values derived from the sensor are not affected by terrain,
28 location, buildings, trees, lights, or cloud layers near the surface. The sensor transmits an
29 average 1-min value for a 10-min period. The sensor only samples 0.75 ft of the atmosphere.
30 An algorithm processes the air passing through the sensor over the 10-min measurement period
31 to provide a generally accurate visibility measurement for within 2 to 3 mi of the site. Moisture,
March 2001 4-99 DRAFT-DO NOT QUOTE OR CITE
-------
1 dust, snow, rain, or particles in the light beam affect the amount of light scattered (National
2 Weather Service, 1998).
3 Visibility data from the ASOS network is typically reported in small increments, up to
4 10 mi, for the purposes of airport operations. However, beginning in 1998, the raw visibility
5 data, including light extinction measurements equaling to visual ranges exceeding 10 mi, have
6 been archived in databases available from the National Climatic Data Center. Data for visibility
7 at larger distances from ASOS sites are available at the sensors for only a short period of time.
8 The data can be directly downloaded from the site. The ASOS data may be useful for
9 characterizing visibility in urban and suburban areas across the country. It also may be used in
10 future analyses to better understand the effects of fine PM on visibility in non-class I areas.
11 The largest monitoring network that includes both visibility and air quality measurements is
12 the Interagency Monitoring of Protected Visual Environments (IMPROVE) network. The
13 IMPROVE network was formed as a collaborative effort between the EPA and federal land
14 management agencies (National Park Service, U.S. Forest Service, Bureau of Land Management,
15 and Fish and Wildlife Service) responsible for Class I areas and the land around them (National
16 Park Service, 1998; Malm et al., 1994; Sisler et al., 1993; U.S. Environmental Protection
17 Agency, 1995a; Eldred et al., 1997; Perry et al., 1997). The primary monitoring objectives of the
18 IMPROVE program are to establish visibility levels, identify anthropogenic sources of
19 impairment, document progress towards elimination of visibility impairment in protected areas
20 from anthropogenic sources, and promote the development of visibility monitoring equipment
21 and the collection of comparable visibility data (National Park Service, 1998; Evans and
22 Pitchford, 1991). Presently over 70 sites employ the IMPROVE program monitoring methods.
23 It is anticipated that an additional 80 sites will be added in 2000.
24 Table 4-7 contains PM25 monitoring data from 30 IMPROVE sites for the years 1988 to
25 1996. The data includes averaged PM2 5 mass and specific species contributions. The data are
26 divided into eastern and western regions. The eastern regions, in addition to Washington, DC,
27 include Acadia National Park and Appalachia and consist of data from Shenandoah and the
28 Great Smoky Mountains National Parks. The western regions include the Northern Great Plains,
29 West Texas, Sonora, the Colorado Plateau, Central Rockies, Cascade, Sierra Humbolt, West
30 Coast, Sierra Nevada, Southern California, and Alaska (Sisler and Malm, 2000).
31
March 2001 4-100 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-7. AVERAGED REGIONAL PM2 5 MASS AND EXTINCTION
SUMMARIES FOR THE YEARS 1988 TO 1996a
Region
Alaska
Appalachia
Cascades
Colorado Plateau
Central Rockies
Coastal
Northeast
Northern Great Plains
Northern Rockies
Southern California
Sonora
Sierra Nevada
Sierra Humbolt
Washington, DC
West Texas
PM25
1.71
(11.9)
10.81
(97.6)
4.67
(50.6)
3.15
(17.3)
2.87
(15.8)
4.40
(43.5)
6.13
(59.3)
4.26
(30.3)
5.15
(39.5)
8.64
(51.7)
4.09
(21.3)
4.40
(25.2)
2.67
(16.7)
16.90
(132.8)
5.11
(27.0)
Sulfate
0.55
(5.1)
6.53
(71.7)
1.30
(29.1)
1.06
(6.7)
0.80
(5.5)
1.35
(18.4)
3.32
(40.6)
1.61
(14.6)
0.98
(15.0)
1.45
(9.3)
1.52
(8.3)
0.96
(7.0)
0.52
(5.2)
7.91
(73.2)
2.13
(12.9)
Nitrate
Organics
0.06
(0.06)
0.60
(6.9)
0.23
(5.0)
0.21
(1.3)
0.18
(1.2)
0.90
(10.9)
0.40
(4.8)
0.51
(4.7)
0.31
(4.7)
O CO
3.53
(22.6)
0.24
(1.3)
0.47
(3.5)
0.16
(1.5)
2.16
(19.9)
0.25
(1.5)
Organics
0.77
(3.1)
2.73
(10.9)
2.51
(10.0)
1.08
(4.3)
1.11
(4.4)
1.65
(6.6)
1.84
(7.3)
1.35
(5.4)
0 ££
Z.oo
(11.5)
2.29
(9.2)
1.28
(5.1)
2.16
(8.6)
1.36
(5.5)
4.44
(17.8)
1.29
(5.2)
Fine Soil
0.22
(1.0)
0.52
(4.3)
0.22
(4.1)
0.64
(1.7)
0.64
(1.4)
0.25
(2.5)
0.23
(3.4)
0.63
(1.6)
0.57
(4.1)
0.94
(4.2)
0.84
(2.0)
0.55
(2.6)
0.42
(2.0)
0.82
(15.6)
1.27
(1.7)
Elemental
Carbon
0.10
(2.2)
0.43
(3.8)
0.41
(2.3)
0.17
(3.3)
0.14
(3.2)
0.25
(5.1)
0.34
(3.0)
0.16
(4.0)
0.41
(4.1)
0.42
(6.3)
0.20
(4.6)
0.26
(3.5)
0.20
(2.5)
1.56
(6.3)
0.17
(5.7)
aMass is in ^g/m3. Extinction summaries in parenthesis are in Mm.
Adapted: Sisler and Malm (2000).
March 2001
4-101
DRAFT-DO NOT QUOTE OR CITE
-------
1 The U.S. Environmental Protection Agency is currently in the process of establishing a
2 national PM25 monitoring network of approximately 1,700 monitors at over 1,100 sites. The
3 PM2 5 monitoring effort will be coordinated with visibility monitoring efforts currently in place,
4 such as IMPROVE, to maximize benefits of both programs. The monitoring network is expected
5 to be fully implemented by the end of 2000 or shortly thereafter (U.S. Environmental Protection
6 Agency, 1997b; U.S. Environmental Protection Agency, 2000b).
7
8 4.3.7 Visibility Modeling
9 There are several types of models available for the evaluation of pollution-related effects on
10 visibility. Plume visibility models and regional haze models are source models that simulate the
11 transport, dispersion, and transformation of chemical species in the atmosphere. Plume models
12 use the resulting air quality data to calculate the values of parameters related to human
13 perception, such as contrast and color differences. Regional haze models calculate aerosol
14 species concentrations and the light-extinction coefficient. Models for the photographic
15 representation of haze use air quality data as an input and perform the optical calculations
16 required to create images that represent the visual effects of the air quality.
17
18 4.3.7.1 Regional Haze
19 Regional haze models may be used to assess the impact of pollutant sources on an
20 identified area or region, in most cases identified class I wilderness areas, or to evaluate the
21 impact of new or existing air quality regulations. Light extinction by fine particles is used to
22 determine the effect of anthropogenic pollutants on regional visibility degradation (regional
23 haze). In the United States, these anthropogenic particles are composed primarily of sulfate
24 compounds, organic compounds, and, to a much lesser extent, nitrate compounds, with the
25 exception of California, where nitrates are the largest single contributor to light extinction. The
26 contribution to light extinction by these compounds will vary based on the particle composition
27 and size distribution. Once the particles are formed, their size can change, resulting in a change
28 in their light extinction efficiency. Model calculations take into consideration the mass of the
29 particulate constituents and the relative humidity.
30 The model requirements for regional-scale, multiple-source haze models are nearly
31 identical to the model requirements for simulations of regional-scale, multiple-source
March 2001 4-102 DRAFT-DO NOT QUOTE OR CITE
-------
1 fine-particle impacts. Hence, the Eulerian-based grid models currently under development to
2 support fine particle impact assessments will be relied on to provide a means for assessing
3 large-scale, multiple-source haze impacts.
4 Middleton (1997) described the findings of a Eulerian-based grid model, the Denver Air
5 Quality Model (DAQM). The DAQM is the principal component of the Brown Cloud n study
6 that is part of earlier work investigating visibility in Denver over the last 20 years. The DAQM
7 is derived from the Regional Acid Deposition Model (RADM) and includes aerosol processes,
8 meteorological modeling analysis, and visibility analysis procedures. The DAQM has been used
9 to determine the relationship between emissions and concentrations of fine and coarse particles
10 and all major gaseous pollutants under various emission scenarios and meteorological
11 conditions. The results of the study demonstrated an association between visibility and air
12 quality issues in the Colorado Front Range area.
13 Neff (1997), in his evaluation of the DAQM model, suggested that the meteorological
14 model does not address adequately mesoscale structures responsible for the initiation and
15 maintenance of the brown cloud episodes or cloud systems and surface moisture fluxes. Given
16 these model uncertainties, it was suggested that there may be errors in the quantification of
17 emissions and in the calculated optical extinction and scattering.
18 The Visibility Assessment Scoping Model (VASM) uses Monte Carlo techniques to
19 generate multiple realizations of daily concentrations of sulfates, nitrates, elemental carbon,
20 organic carbon, fine and coarse dust, and the relative humidity to determine particle effects on
21 regional haze. Species-specific light attenuation is calculated based on particle concentration and
22 relative humidity, producing short-term haze intensity or visual range information (Shannon
23 etal., 1997).
24 The Elastic Light Scattering and Interactive Efficiency (ELSIE) model was used by Omar
25 et al. (1999) to determine the species concentrations and to relate apportionment to the extinction
26 coefficient in an aerosol mixture. The model assumes the aerosol is an internal inhomogeneous
27 mixture of chemical species and size distributions. Model input parameters included the size
28 distributions, prevailing relative humidity, refractive indices of the constituents, percent
29 solubility of the aerosol components, and the growth function of the aerosol particles. The model
30 assumes that the particles grow with increasing relative humidity according to a predetermined
31 growth function.
March 2001 4-103 DRAFT-DO NOT QUOTE OR CITE
-------
1 Several source-oriented models have been developed to evaluate the effects of pollutants on
2 regional haze. The U.S. Environmental Protection Agency, in cooperation with the U.S. Forest
3 Service, the Fish and Wildlife Service, and the National Park Service (the Interagency
4 Workgroup for Air Quality Modeling), developed the MESOPUFF II system of assessing
5 regional haze impacts. The MESOPUFF II system uses the light extinction for sulfates and
6 nitrates for an estimated 3- to 24-h average concentration (U.S. Environmental Protection
7 Agency, 1995b). The CALPUFF modeling system can process mesoscale meteorological data
8 and address dispersive processes of a regional nature. Simulated long-range pollutant trajectories
9 have been compared successfully to results from a field study involving transport to 1000 km
10 downwind (U.S. Environmental Protection Agency, 1995c). However, Lagrangian puff
11 dispersion modeling involving transport of 200 km or more tend to underestimate the horizontal
12 extent of the dispersion, causing the surface concentration to be overestimated (Moran and
13 Pielke, 1994). Another source-oriented Lagrangian trajectory model capable of computing light
14 extinction and scattering and estimating visual range from gas phase and primary particle phase
15 air pollutant emissions directly from sources was reported by Eldering and Cass (1996). The
16 model is comprised of several modules that take into consideration particle size distribution and
17 chemical composition, the speciation of organic vapor emissions, atmospheric chemical
18 reactions, transport of condensible material between the gas and particle phase, fog chemistry,
19 dry deposition, and light scattering and absorption. The model is, however, not suitable for
20 predicting visibility over great distances through nonuniform hazes and for visualization of
21 pollutant effects of isolated major point source plumes. Single line Lagrangian trajectory models
22 cannot represent horizontal turbulent diffusion, the effects of wind shear, and advection by
23 turbulent wind components. Error in transport calculations have been reported of up to ± 50%
24 (Eldering and Cass, 1996).
25 Gray and Cass (1998) developed a lagrangian particle-in-cell model for predicting source
26 class contributions of fine particle total carbon and elemental carbon. The model simulates the
27 motion and deposition of pollutants in an air basin with varying meteorological conditions. The
28 model also takes into consideration the vertical mixing characteristics of pollutants in areas
29 located near the source. The model is useful in determining changes in long-term average
30 pollutant concentrations from implementing specific emission control measures.
March 2001 4-104 DRAFT-DO NOT QUOTE OR CITE
-------
1 The Regional Particulate Model (RPM) simulates secondary fine particulate matter (PM2 5)
2 formation and long-range transport. The RPM is used with the Regional Acid Deposition Model
3 (RADM), a comprehensive acid rain model. Predictions from the RADM are used to simulate
4 the formation of sulfate and nitrate, ammonium particles, and secondary organic aerosols. The
5 external RADM includes particle physics from the RPM and operates at an 80- and 20-km
6 resolution. Additional work currently is being done that will incorporate the RADM/RPM and
7 external RADM models into a more comprehensive air quality modeling system,
8 Models-3/Community Multi-Scale Air Quality (CMAQ). This modeling system simulates the
9 processes involved in primary and secondary PM10 and PM2 5 and ozone formation, regional haze,
10 acid deposition, and nutrient deposition. The modeling system includes a mesoscale
11 meteorological model, emission model, and a version of the CMAQ.
12 The Regulatory Modeling System for Aerosols and Deposition (REMSAD) also simulates
13 PM25 formation. The REMSAD was derived from the Urban Airshed Model Version V
14 (UAM-V) for primary and secondary PM2 5 and PM10 formation, and acid nutrient and toxic
15 deposition. The REMSAD system consists of a meteorological data preprocessor, the core
16 aerosol and toxic deposition model (ATOM), and postprocessing programs. The ATOM is a
17 three-dimensional Eulerian grid model designed to calculate the concentrations of both inert and
18 chemically reactive pollutants by simulating the physical and chemical processes in the
19 atmosphere that affect pollutant concentrations. The basis for the model is the atmospheric
20 diffusion or species continuity equation. This equation represents a mass balance in which all of
21 the relevant emissions, transport, diffusion, chemical reactions, and removal processes are
22 expressed in mathematical terms (Systems Applications International, Inc., 1998).
23 Zannetti et al. (1990, 1993) and Fox et al. (1997) described a semi-empirical model that
24 could be used to estimate the visibility impact on one region resulting from sulfur dioxide
25 emission controls in a different region. The model combined four different input parameters:
26 (1) chemical transport; (2) possible nonlinearity of pollutant chemical transformation; (3) sulfate
27 fraction of fine particulate matter, including the amount of water absorbed by the fine particles;
28 and (4) the fraction of light extinction caused by fine particles. The model uses physically
29 realistic concepts of atmospheric transport, chemical transformation, and physical effects.
30 However, actual data sets, mathematical constructs, or expert opinions also may be used. Models
31 also have been developed that predict the downwind concentration of smoke particulate and other
March 2001 4-105 DRAFT-DO NOT QUOTE OR CITE
-------
1 combustion products from the burning of crude oil from accidental spills (McGrattan et al., 1995,
2 1996).
3
4 4.3.7.2 Plume Models
5 Several plume visibility models are currently available. Plume visibility models estimate
6 the value of optical parameters related to human perception, such as contrast and color
7 differences, and compare these values with perception thresholds to determine whether the plume
8 is likely to be perceptible under various simulated conditions (U.S. Environmental Protection
9 Agency, 1988; Latimer, 1988). An empirical algorithm, Probability of Detection Algorithm
10 (PROBDET), allows the prediction of the lower limit of plume contrast that can be detected
11 visually. The PROBDET can be used to estimate the detection level for plumes that fall within
12 the bounds defined by the full-length, oval, and circular plume stimuli (Ross et al., 1997).
13 A simplified dispersion model using a second-order turbulence closure scheme to account
14 for averaging time effects on the dispersion rate was described by Sykes and Gabruk (1997). The
15 lateral and vertical spread is estimated using a Gaussian plume framework. A simplified
16 representation of the turbulence spectrum is used to predict the reduced spread rate for short
17 averaging times.
18 Earlier plume models included PLUVUE I and n, used during the preparation of a permit
19 application to determine whether or not a proposed new facility would cause visibility
20 impairment in a Class I area (Latimer et al., 1978; Johnson et al., 1980; White et al., 1985; U.S.
21 Environmental Protection Agency, 1992). Seigneur et al. (1997) developed a plume visibility
22 model, the Reactive and Optics Model Emissions (ROME), that improves on the existing plume
23 visibility models. The model simulates the momentum and buoyancy forces of the plume rise,
24 the dispersion and chemistry, and condensation and evaporation of the aqueous phase.
25 A second-order closure algorithm is used to estimate instantaneous plume concentrations, or the
26 time-averaged plume concentration may be estimated using a first-order closure algorithm.
27 A comprehensive chemical kinetic mechanism simulates chemical transformation processes in
28 the gas, aqueous, and particle phases. Particle dynamics and chemical composition is based on
29 sectional representation of the particle size distribution. The model includes a radioactive
30 transfer module that provides optical properties using sectional particle size distributions.
31 Deposition velocities based on atmospheric stability, surface type, chemical type, and particle
March 2001 4-106 DRAFT-DO NOT QUOTE OR CITE
-------
1 size are derived using a resistance-based, dry deposition algorithm. The ROME can be used with
2 other models to estimate a stack plume opacity, the percentage of light intensity attenuated by the
3 plume near the stack after any condensed water has evaporated (Meng et al., 2000). When
4 compared with the PLUVUE n, the ROME, with the second-order dispersion algorithm, was
5 found to present a more accurate estimate of plume height, width, nitrogen oxide concentration,
6 nitrogen dioxide/nitrogen oxide ratio, and visibility. Error, bias, correlation coefficients, and
7 simulations were within a factor of two of that observed (Gabruk et al., 1999).
8
9 4.3.7.3 Photographs
10 Computer-generated photographs are sometimes used to illustrate the effects of pollution
11 on visibility. To begin, a photograph is taken on a very clean, cloud-free day to serve as the
12 initial scene image. As previously indicated, the appearance of an object is determined by the
13 path radiance and the transmitted radiance. To determine the transmitted radiance, an estimate of
14 the light-extinction coefficient from the photograph is used to determine the initial radiance for
15 each element in the scene. The transmitted radiance is equal to the initial radiance of the
16 element in the scene multiplied by the transmittance of the atmosphere in the sight path. Because
17 the path radiance changes over the distance of the sight path, the source function, the rate of
18 change over the distance of the sight path, also must be determined.
19 Eldering et al. (1996) proposed the use of a model that uses simulated photographs from
20 satellite and topographic images to evaluate the effect of atmospheric aerosols and gases on
21 visibility. Use of this model requires ground-based photography and size distribution and
22 chemical composition of atmospheric aerosols, NO2 concentration, temperature, and relative
23 humidity for a clear day, for comparison purposes. Light extinction and sky color are then
24 calculated based on differences in aerosol size distribution, NO2 concentration, temperature, and
25 relative humidity. The images created represent natural landscape elements.
26 Molenar et al. (1994) provides a discussion of existing visual air quality simulation
27 methods based on techniques under development for the past 20 years. The WinHaze visual air
28 quality modeling system is one tool that has been developed using techniques to simulate
29 changes in visibility due to changes in air quality.
30 One of the limitations in using photographic models for representation of haze is that haze
31 is assumed to be uniformly distributed throughout the scene and selected conditions are
March 2001 4-107 DRAFT-DO NOT QUOTE OR CITE
-------
1 idealized, so the full range of conditions that occur in a scene are not represented. Photographs
2 are also expensive to produce. More detailed information on the use of photographic
3 representation of haze may be found in the U.S. Environmental Protection Agency (1996b),
4 Trijonis et al. (1991), Molenar et al. (1994), and Eldering et al. (1993).
5
6 4.3.8 Trends in Visibility Impairment
7 Trends in visibility impairment or haziness, visual range, often are associated with fine
8 mass concentrations (<2.5 //g/m3). Observations of visual range, obtained by the National
9 Weather Service and available through the National Climatic Data Center of the National
10 Oceanic and Atmospheric Administration, provide one of the few truly long-term, daily records
11 of any parameter related to air pollution. After some manipulation, the visual range data can be
12 used as an indicator of fine mode particle pollution. The data reduction process and analyses of
13 resulting trends have been reported by Husar et al. (1994), Husar and Wilson (1993), and Husar
14 etal. (1981).
15 Generally, visibility impairment is greatest in the eastern United States and southern
16 California. Haziness in the southeastern United States is greatest in the humid summer months
17 because of its affinity to atmospheric water vapor, followed by the spring and fall, and winter.
18 Summer haziness in the southeastern United States has increased by approximately 80% since
19 the 1950s (Husar and Wilson, 1993) because of increased sulfate from increased SO2 emissions
20 (Husar et al., 1994). The resulting sulfate, considered to be ammonium sulfate, accounts for
21 40 to 70% of the fine particle mass (Husar and Wilson, 1993). Sulfate-related effects on
22 visibility in the southeast is a factor of 20 higher than the Great Basin area and 10 higher than the
23 desert southwest, central Rocky Mountains, and Sierra Mountains (Malm et al., 1994). For most
24 rural eastern sites, sulfates accounts for >60% of the annual average light extinction on the best
25 days and >75% of the light extinction on the worst days. A statistically significant increase in
26 summer sulfate concentrations was noted in two class I areas in the eastern United States
27 (Shenandoah and the Great Smoky Mountains) from 1982 to 1992 (Eldred et al., 1993; Cahill
28 etal., 1996). The increase was largest in the summer and decreased in the winter. The majority
29 of the southwest showed decreasing sulfur (Eldred et al., 1993; Eldred and Cahill, 1994). White
30 (1997) suggested that the increase in fine-particle sulfur may be the result of the measurement
31 method and not an upward trend in fine particle concentrations in those Class I areas. However,
March 2001 4-108 DRAFT-DO NOT QUOTE OR CITE
-------
1 Iyer et al. (2000), using the Spearman correlation of trend, reported an increased trend in hazy
2 days during the summer months in Shenandoah and the Great Smoky Mountains based on
3 monitoring data for the period 1979 to 1996 showing high sulfur concentrations.
4 Sulfates also may be a significant contributor to total light extinction in the rural western
5 United States, accounting for 30 to 40% of the total light extinction on the best days and 35 to
6 45% of the total light extinction on the haziest days. In several areas of the west, sulfates account
7 for over 50% of the annual average aerosol extinction (U.S. Environmental Protection Agency,
8 2000b).
9 Organics are the second largest contributor to light extinction in most areas in the United
10 States. Extinction caused by organic carbon is greatest in the Pacific Northwest, Oregon, Idaho,
11 and Montana, accounting for 40 to 45% of the total extinction. Organic carbon can contribute
12 between 20 to 30% to the total extinction in most of the western United States and 10 to 15% in
13 the remaining areas of the United States. Light absorption by carbon is relatively insignificant
14 but is highest in the Pacific Northwest (up to 15%) and in the eastern United States (up to 6%)
15 (Malm et al., 1994; U.S. Environmental Protection Agency, 2000b).
16 Some of the visibility impairment in northern California and Nevada, including Oregon,
17 southern Idaho and western Wyoming, results from coarse mass and soil, primarily considered
18 natural extinction. In some areas of the United States, extinction from coarse mass is almost
19 negligible because the overall extinction is so high. High dust concentrations from southern
20 California have contributed to regional haze in the Grand Canyon and other class I areas in the
21 southwestern United States (Vasconcelos et al., 1996). White et al. (1999) reported that some of
22 the worst haze near the Grand Canyon is associated with pollutant transport from southern
23 California and the subtropics.
24 Visibility impairment in southern California is primarily caused by light extinction by
25 nitrates. Nitrates contribute about 40% to the total light extinction in Southern California and
26 10 to 20% of the total extinction in other areas of the United States.
27 The average haze patterns across the continental United States, for five-season averages for
28 the years 1980 to 1985 and 1990 to 1995 are shown in Figure 4-23. Haze is indicated by the
29 75th percentile of the extinction coefficient that is calculated from the visual range, corrected to
30 60% relative humidity by the Koschmeider relationship.
31
March 2001 4-109 DRAFT-DO NOT QUOTE OR CITE
-------
g.
to
o
o
1980-85 Winter
;==;;|;;E=;;|P||E
1
1/km
1980-85 Summer
\*s
\
H
6
o
O
H
O
c
o
H
W
O
^
O
HH
H
W
,' 1990-95 Summer
Figure 4-23. Five-season average haze patterns (75th percentile of the extinction coefficient) for the periods 1980 to 1985 and
1990 to 1995. The relationship between haze, expressed as extinction coefficient (corrected to 60% relative
humidity) is 1.9/extinction coefficient in km"1 = visual range in km (i.e., the intensity of shading increases as visual
range decreases: 0.2 km'1 = 9.5 km, 0.17 km~l = 11.2 km, 0.14 km'1 = 13.6 km, 0.11 km~l = 17.3 km,
0.08 km J = 23.8 km, and 0.05 km J = 38.0 km). Winter = Dec., Jan., Feb.; Summer = June, July, Aug).
Source: Husar et al., 2000.
-------
1 The trends graphs in Figure 4-24 for regions in the United States represent the 75th
2 percentile of the light extinction coefficient for the stations located within the designated region
3 over a 30-year period (1940 to 1990). The trends are presented for quarters 1 (winter) and
4 3 (summer). The northeastern United States exhibited an increase in haze during quarter 3
5 between 1960 and 1970 and a steady decline between 1973 (0.22 km4) and 1992 (0.12 km4).
6 In quarter 1 the haziness steadily declined from 0.15 to 0.10 km4 in the 30-year period. The
7 Mid-Atlantic region, the Virginias and Carolinas, shows a strong increase in haziness in quarter 3
8 between 1960 and 1973, followed by a decline. The winter haze was virtually unchanged over
9 the 30-year period. The haziness over the Gulf states increased between 1960 and 1970 and
10 remained virtually unchanged since then. The central Midwest, including Missouri and
11 Arkansas, exhibit virtually no change during the winter season and a slight increase in the
12 summer (1960 to 1970). The upper Midwest shows an opposing trend for summer and winter.
13 Although summer haze has increased, mostly from 1960 to 1973, the winter haze has declined.
14 Based on PM2 5 concentrations and changes in the deciview scale, calculated from
15 reconstructed extinction coefficients, Sisler and Malm (2000) reported no significant
16 deterioration in air quality and visibility conditions at 30 IMPROVE network sites for the years
17 1988 to 1996. The sites were divided into eastern and western regions. Averaged PM25 mass
18 and extinction summaries for the sites appear in Table 4-7. The annual best visibility
19 (10th percentile) and median visibility days (50th percentile) are improving at approximately
20 70% of the sites. However, several sites are not showing steady improvements in either visibility
21 or PM25, particularly in the number of worst visibility days (90th percentile). The sites included
22 the Badlands, Big Bend, Crater Lake, Great Smoky Mountains, Mesa Verde, Shenandoah and
23 Yosemite National Parks, Chiricahua National Monument, and the District of Columbia.
24
25 4.3.9 Economics of Particulate Matter Visibility Effects
26 Given the evidence of potential economically significant effects of visibility impairtment,
27 economic analysis proceeds by quantifying in monetary terms the costs associated with different
28 ambient levels of PM. Where possible, direct economic valuation can take place using prices
29 that are determined in the marketplace. There are a variety of ways to estimate costs/benefits.
30 Avoided cost methods estimate the costs of pollution by using the expenditures that are made
31 necessary by pollution damage. For example, if ambient levels of particulate matter results in
March 2001 4-111 DRAFT-DO NOT QUOTE OR CITE
-------
g.
to
o
o
fe
H
6
o
o
H
O
O
H
W
O
O
H
W
0.06
194019501960197019801990
0.06L
194019501960197019801990
0.06
1940 1950 1960 1970 1980 1990
0.06
1940 1950 1960 1970 1980 1990
0.06
1940 1950 1960 1970 1980 1990
0.06L
1940 1950 1960 1970 1980 1990
Figure 4-24. Secular haze trends for 1940 to 1990 for six regions, summer (Quarter 1) and winter (Quarter 3) in the
75th percentile.
-------
1 increased frequency of building cleaning or repainting, then the appropriately calculated increase
2 in these costs is a reasonable estimate of true economic damage. Benefits associated with
3 reductions in the pollution levels then are represented by the avoided costs of these damages.
4 Estimating benefits for visibility is a more difficult and less precise exercise because the
5 effects are not valued in the marketplace. There are several methods that economists have
6 developed to estimate changes in environmental effects that are not valued in the marketplace
7 (Freeman, 1993). These include hedonic price analysis, stated preference models (including
8 contingent valuation, contingent choice, and contingent ranking), and travel cost models.
9 Hedonic price analysis works by analyzing the way that market prices change when an associated
10 environmental effect changes. Part of the economic costs imposed by the reduced visibility
11 caused by PM can be estimated by looking at the differences in sales price between otherwise
12 identical houses that have different degrees of visibility impairment.
13 The contingent valuation method (CVM) has been used to determine estimated value
14 changes in both visibility and ecosystem functions (Hanley and Spash, 1993; Chestnut, 1997).
15 The CVM determines pollutant-related effects by using carefully structured surveys to estimate
16 the amount of compensation equivalent to a given change in environmental quality or
17 equivalently, how much they would be willing to pay to obtain a given change in environmental
18 quality. There is an extensive scientific literature and body of practice on both this theory and
19 technique.
20 Other valuation methods include stated preference models, including contingent choice and
21 contingent ranking (also known as conjoint analysis), as well as travel cost models (Johnson and
22 Desvousges, 1997; Hanley and Spash, 1993). However, the primary methods used to date for
23 valuation of visibility have been the hedonic price and contingent valuation methods (Hanley and
24 Spash, 1993).
25 The effects of PM on visibility may differ widely between urban residential and
26 recreational areas. Separate estimates are needed to account for welfare changes associated with
27 improvements in visibility in class I areas. Chestnut and Dennis (1997) developed a method for
28 estimating the value to the public of visibility improvements in class I areas using the results of a
29 1990 cooperative agreement project jointly funded by the EPA and the National Park Service:
30 "Preservation Values For Visibility Protection at the National Parks." Using the contingent
31 valuation method, Chestnut and Davis calculated a household willingness to pay for visibility
March 2001 4-113 DRAFT-DO NOT QUOTE OR CITE
-------
1 improvements in class I areas, capturing both use and nonuse recreational values. This analysis
2 also accounts for geographic variations in the willingness to pay. The results indicate a
3 willingness to pay per deciview improvement in visibility of between $5 and $17 per household.
4
5
6 4.4 EFFECTS ON MATERIALS
7 Effects of air pollution on materials are related to both aesthetic appeal and physical
8 damage. Studies have demonstrated that particles, primarily consisting of carbonaceous
9 compounds, cause soiling of commonly used building materials and culturally important items,
10 such as statutes and works of art. Physical damage from the dry deposition of air pollutants, such
11 as PM (especially sulfates and nitrates) and SO2, and the absorption or adsorption of corrosive
12 agents on deposited particles also can result in the acceleration of naturally occurring weathering
13 processes of man-made building and cultural materials.
14 In the atmosphere, PM may be "primary", existing in the same form in which it was
15 emitted, or "secondary", formed by the chemical reactions of free, absorbed, or dissolved gases.
16 The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen ions;
17 particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
18 material. A substantial fraction of the fine particle mass, particularly during the warmer months,
19 is secondary sulfate and nitrate. Sulfates may be formed by the gas-phase conversion of SO2 to
20 H2SO4 by OH radicals and aqueous-phase reactions of SO2 with H2O2, O3, or O2. During the day,
21 NO2 may be converted to nitric acid (HNO3) by reacting with OH radicals. Nitrogen dioxide also
22 can be oxidized to HNO3 by a sequence of reactions initiated by O3. A more detailed discussion
23 of the atmospheric chemistry of PM appears in Chapter 2 of this document.
24 Limited new studies have been published that better define the role of air pollution in
25 materials damage. This section briefly summarizes information on particle and sulfur-containing
26 pollutants (formed by the chemical reactions of SO2 with other atmospheric pollutants) exposure-
27 related effects on materials addressed in the 1996 PM AQCD (U.S. Environmental Protection
28 Agency, 1996a) and presents relevant information published since completion of that document.
29 The effects of nitrates on manmade building materials and naturally occurring cultural materials
30 was discussed in the criteria document on nitrogen oxides (U.S. Environmental Protection
31 Agency, 1993).
March 2001 4-114 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.4.1 Effects of Particles and Sulfur Dioxide on Man-Made Surfaces
2 4.4.1.1 Metals
3 Metals under go natural weathering processes in the absence of environmental pollutants.
4 The additive effect of pollutants on the natural weathering processes will depend on the nature of
5 the pollutant and the deposition rate (the uptake of a pollutant by the material's surface), and the
6 presence of moisture. The influence of the metal protective corrosion film, the presence of other
7 surface electrolytes, the orientation of the metal surface, the presence of surface moisture, and the
8 variability in the electrochemical reactions will also contribute to the affect of pollutant exposure
9 on metal surfaces.
10 Several studies demonstrate the importance of time of surface wetness (caused by dew and
11 fog condensation and rain) on metals. Surface moisture facilitates the deposition of pollutants,
12 especially SO2, and promotes corrosive electrochemical reactions on metals (Haynie and Upham,
13 1974; Sydberger and Ericsson, 1977). Of critical importance is the formation of hygroscopic
14 salts on the metal that increases the time of surface wetness and, thereby, enhances the corrosion
15 process.
16 Pitchford and McMurry (1994) and Zhang et al. (1993) demonstrated particle size-related
17 effects of relative humidity. The effect of temperature on the rate of corrosion is complex.
18 Under normal temperature conditions, temperature would not have an affect on the rate of
19 corrosion. When the temperature decreases the relative humidity increases and the diffusivity
20 decreases. The corrosion rate decreases as the temperature approaches freezing because ice
21 prohibits the diffusion of SO2 to the metal surface and minimizes electrochemical processes
22 (Haynie, 1980; Biefer, 1981; Sereda, 1974).
23 The metal protective corrosion film (i.e., the rust layer on metal surfaces) provides some
24 protection against further corrosion. The effectiveness of the corrosion film in slowing down the
25 corrosion process is affected by the solubility of the corrosion layer, and the concentration and
26 deposition rate of pollutants. If the metal protective corrosion film is insoluble, it may add some
27 protection against acidic pollutants. An atmospheric corrosion model that considers the
28 formation and dissolution of the corrosion film on galvanized steel was proposed by Spence et al.
29 (1992). The model considers the effects of SO2, rain acidity, and the time of wetness on the rate
30 of corrosion. Although the model does not characterize specifically particle effects, the
31 contribution of particulate sulfate was considered in model development.
March 2001 4-115 DRAFT-DO NOT QUOTE OR CITE
-------
1 Whether suspended particles actually impact on the corrosion of metals is not clear.
2 Several studies suggest that suspended particles will promote the corrosion of metals (Goodwin
3 et al., 1969; Barton, 1958; Sanyal and Singhania, 1956; Baedecker et al., 1991); however, other
4 studies have not demonstrated a correlation between particle exposure and metal corrosion
5 (Mansfeld, 1980; Edney et al., 1989). Walton et al. (1982) suggested that catalytic species within
6 several species in fly ash promote the oxidation of SOX to a corrosive state. Still other
7 researchers indicate that the catalytic effect of particles is not significant, and that the corrosion
8 rate is dependent on the conductance of the thin-film surface electrolytes during periods of
9 wetness. Soluble particles likely increase the solution conductance (Skerry et al., 1988; Askey
10 etal., 1993).
11 The corrosion of most ferrous metals (iron, steel, and steel alloys) is increased by
12 increasing SO2 exposure. Steels are susceptible to corrosion when exposed to SO2 in the absence
13 of protective organic or metallic coatings. Studies on the corrosive effects of SO2 on steel
14 indicate that the rate of corrosion increases with increasing SO2 and is dependent on the
15 deposition rate of the SO2 (Baedecker et al., 1991; Butlin et al., 1992a). The corrosive effects of
16 SO2 on aluminum is exposure-dependent, but appears to be insignificant (Haynie, 1976; Fink
17 et al., 1971; Butlin et al., 1992a). The rate of formation of the patina on copper (protective
18 covering) can take as long as 5 years and is dependent on the SO2 concentration, deposition rate,
19 temperature, and relative humidity (Simpson and Horrobin, 1970). Further corrosion is
20 controlled by the availability of copper to react with deposited pollutants (Graedel et al., 1987).
21 Butlin et al. (1992a), Baedecker et al. (1991), and Cramer et al. (1989) reported an average
22 corrosion rate of 1 //m/year for copper; however, less than a third of the corrosion was attributed
23 to SO2 exposure, suggesting that the rate of patina formation was more dependent on factors
24 other than SO2. A recent report by Strandberg and Johansson (1997) showed relative humidity to
25 be the primary factor in copper corrosion and patina formation. The results of the studies on
26 particles and SO2 corrosion of metals are summarized in Table 4-8.
27
28 4.4.1.2 Painted Finishes
29 Exposure to air pollutants affect the durability of paint finishes by promoting discoloration,
30 chalking, loss of gloss, erosion, blistering, and peeling. Evidence exists that indicates particles
31 can damage painted finishes by serving as carriers for corrosive pollutants (Cowling and Roberts,
March 2001 4-116 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-8. CORROSIVE EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON METALS
g.
to
o
o
Metal
Exposure Conditions
Comments
Bibliography
\*s
\
H
6
o
O
H
O
c
o
H
W
O
V
O
HH
H
W
Mild Steel
Galvanized Steel
Zinc
Zinc
Carbon Steel
Weathering Steel
Aluminum
Aluminum
Specimens exposed to SO2 and O3 under natural
and artificial conditions, and to NO2 under natural
conditions. SO2 concentrations ranged from 2.1 to
60 ^g/m3. Annual average concentrations were
about 20 /j,g/m3. Meteorological conditions were
unaltered. Specimens exposed at 29 sites for
2 years for mild steel and 1 y for galvanized steel.
Rolled zinc specimens exposed at various sites
around the country (rural, industrialized, marine)
for up to 20 years. Actual pollutant exposures not
reported.
Specimens exposed at 5 sites for 1 to 5 years.
Average SO2 concentrations ranged from 2 ± 4 to
15 ± 17 ppb (5.2 ± 10.4 to 39.3 ± 44.5
PM concentrations ranged from 14 to 60
Highest pollutant concentrations recorded at
1 year exposure site.
See Baedecker et al. (1991) above for exposure
conditions.
See Baedecker et al. (1991) above for exposure
conditions.
See Butlin et al. (1992a) above for exposure
conditions.
Steel corrosion was dependent on long-term SO2
exposure. The corrosion rate was about 50 ^m/year
for mild steel specimens for most industrial sites, but
ranged from 21 to 71 ^m/year. The corrosion rate
ranged from 1.45 to 4.25 ^m/year for galvanized
steel. The authors concluded that rainfall also may
have a significant effect on galvanized steel based on
a corrosion rate of 3.4 ^m/year seen at a very wet
site.
The highest corrosion rates were associated with
industrialized environments and marine environments
in direct contact with salt spray.
Average corrosion rate ranged from 0.63 to
1.33 jum/y. The highest corrosion was noted in the
most industrialized area. However, the corrosion
rates did not differ significant regardless of the SO2
concentration, suggesting that SO2 exposure may not
be the dominant factor in zinc corrosion.
Average corrosion rate for samples exposed for
5 years ranged from 6.6 to 12.8 ^m/year for carbon
steel and 3.7 to 5.0 ^m/year for weathering steel.
Highest corrosion rate noted for samples exposed for
1 year.
Corrosion rate was very low at all sites and ranged
from 0.036 to 0. 106 ^on/year.
Corrosion greater on the under side of specimens,
possibly because of lack of washoff and increased
PM in area. Maximum corrosion rate was
0.85 ^m/year. Pit depths of up to 72 /j.m were noted
after 2 years of exposure.
Butlin etal. (1992a)
Showak and Dunbar
(1982)
Baedecker etal. (1991)
Cramer etal. (1989)
Baedecker etal. (1991)
Cramer etal. (1989)
Baedecker et al. (1991)
Butlin etal. (1992a)
-------
TABLE 4-8 (cont'd). CORROSIVE EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON METALS
g.
to
o
o
Metal
Exposure Conditions
Comments
Bibliography
oo
\*s
\
H
6
o
O
H
O
c
o
H
W
O
V
O
HH
H
W
Copper
Copper
Copper
Copper
See Baedecker et al. (1991) above for exposure
conditions.
See Butlin et al. (1992a) above for exposure
conditions.
Specimens exposed to 4 to 69 ppb (10.4 to
180.7 ,wg/m3) and 1.0 ppm (2,618.7 ,wg/m3)
SO, for 20h at various relative humidities.
Specimens exposed artificially to 0.49 ± 0.01 ppm
(187 ± 3.8 Mg/m3) SO2 for 4 weeks at 70 and 90%
relative humidity.
Average corrosion rate for 3- and 5-year exposures
was about 1 Aim/year but the soluble portion was less
than a third of that which could be contributed to SO2
exposure. Dry deposition of SO2 was not as
important in patina formation as wet deposition of H+.
Majority of test sites showed a corrosion rate of
1 ± 0.2 ^m/year. The corrosion rate was
1.48 ^m/year at the site receiving the most rainfall.
The lowest corrosion rate, 0.66 ^m/year, was
associated with low rainfall, low SO2.
SO2 had no effect on copper when relative humidity
was -<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.
Baedecker et al. (1991)
Butlin etal. (1992a)
Strandberg and
Johansson (1997)
Eriksson et al. (1993)
-------
1 1954) or by staining and pitting of the painted surfaces (Fochtman and Langer, 1957; Wolff et al.,
2 1990).
3 The erosion rate of oil-base house paint has been reported to be enhanced by exposure to
4 SO2 and high humidity. In a study by Spence et al. (1975), an erosion rate of 36.71 ±
5 8.03 //m/year was noted for oil-base house paint samples exposed to SO2 (78.6 //g/m3), O3
6 (156.8 //g/m3), and NO2 (94 //g/m3) and low humidity (50%). The erosion rate increased with
7 increased SO2 and humidity. The authors concluded that SO2 and humidity accounted for 61% of
8 the erosion. Acrylic coil coating and vinyl coil coating shows less pollutant-related erosion.
9 Erosion rates range from 0.7 to 1.3 //m/year and 1.4 to 5.3 //m/year, respectively. Similar
10 findings on SO2-related erosion of oil-base house paints and coil coatings have been reported by
11 other researchers (Davis et al., 1990; Yocom and Grappone, 1976; Yocom and Upham, 1977;
12 Campbell et al., 1974). Several studies suggest that the effect of SO2 is caused by its reaction
13 with extender pigments such as calcium carbonate and zinc oxide (Campbell et al., 1974; Xu and
14 Balik, 1989; Edney, 1989; Edney et al., 1988, 1989). However, Miller et al. (1992) suggested
15 that calcium carbonate acts to protect paint substrates. Another study indicated that exposure to
16 SO2 can increase the drying time of some paints by reacting with certain drying oils and will
17 compete with the auto-oxidative curing mechanism responsible for crosslinking the binder
18 (Holbrow, 1962).
19
20 4.4.1.3 Stone and Concrete
21 Numerous studies suggest that air pollutants can enhance the natural weathering processes
22 on building stone. The development of crusts on stone monuments have been attributed to the
23 interaction of the stone's surface with sulfur-containing pollutants, wet or dry deposition of
24 atmospheric particles, and dry deposition of gypsum particles from the atmosphere. Because of a
25 greater porosity and specific surface, mortars have a greater potential for reacting with
26 environmental pollutants (Zappia et al., 1998). Details on these studies are discussed in
27 Table 4-9. The stones most susceptible to the deteriorating effects of sulfur-containing pollutants
28 are the calcareous stones (limestone, marble, and carbonated cement). Exposure-related damage
29 to building stones result from the formation of salts in the stone that are subsequently washed
30 away during rain events leaving the stone surface more susceptible to the effects of pollutants.
31 Dry deposition of sulfur-containing pollutants promotes the formation of gypsum on the stone's
March 2001 4-119 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-9. EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON STONE
g.
to
o
o
Stone
Exposure Conditions
Comments
Bibliography
to
o
\*s
\
H
6
o
O
H
O
c
o
H
W
O
V
O
HH
H
W
Vermont marble
Marble sandstone
Limestone
Portland limestone
White Mansfield
dolomitic sandstone
Monk's Park
limestone
Sandstones (calcite
and noncalcite
stones)
Limestones
Sandstones
Marble
Granite
Basalt
Runoff water was analyzed from seven summer
storms. SO2 concentration stated to be low.
Analysis of runoff water for five slabs test
exposed to ambient conditions at a angle of
30° to horizontal.
Ambient air conditions. Exposure ranged from
70 to 1065 days. Averaged pollutant exposure
ranged from 1.4 to 20.4 ppb (3.7 to 53.4 Mg/m3)
SO2; 4.1 to 41.1 ppb NOX; 2.4 to 17.4 ppb
(4.5 to 32.7 Mg/m3) NO2; 10.1 to 25.6 ppb
(19.8 to 50.2 ,wg/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.
Between 10 to 50% of calcium in runoff water estimated from Schuster et al.
gypsum formation from dry deposition of SO2. (1994)
Pollutant exposure related erosion was primarily caused by dry Baedecker
deposition of SO2 and nitric acid between rain events and wet et al. (1992)
deposition of hydrogen ion. Recession estimates ranged from
15 to 30 ^m/year for marble and 25 to 45 ^m/year for limestone.
A large portion of the erosion results from the reaction of CO2
with the calcium in the stone.
Increased stone weight loss with increased SO2. Rainfall did not Webb et al.
significantly affect stone degradation. Stone loss associated with (1992)
SO2 exposure estimated to be 24 ^m/year. Slight trend in
decreasing stone loss with increasing length of exposure.
Significant correlations existed between the mean annual SO2 Butlin et al.
concentration, rainfall volume, and hydrogen ion loading and the (1992b)
weight changes.
Insignificant differences in erosion rate found between calcite and Petuskey et al.
noncalcite sandstone. Moisture affected the rate of pollutant (1995)
deposition and enhanced susceptibility to pollutant related
erosion. Rain events given as primary factor affecting stone
erosion. Pollutant related erosion judged to be insignificant.
Crusts on stones were found to contain two layers; top layer, Garcia-Valles
usually black in color, composed of gypsum between 40 and et al. (1998)
400 jam thick. Innermost layer, ranging from brown to orange in
color, primarily consisted of calcite, between 10 and 600 ,wm
thick. Gypsum-rich layer thought to be the result of sulphation of
the calcitic layer by atmospheric pollutants or dry or wet
deposition of atmospheric dust.
-------
TABLE 4-9 (cont'd). EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON STONE
g.
to
o
o
Stone
Exposure Conditions
Comments
Bibliography
to
\*s
\
H
6
o
O
H
O
c
o
H
W
O
V
O
HH
H
W
Portland limestone
Massangis Jaune
Roche limestone
White Mansfield
dolomitic
Monk's Park
Portland limestone
Carrara marble
Travertine
Tranistone
Carrara marble
Georgia marble
Carrara marble
Monk's Park
limestone
Portland limestone
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.
Samples exposed for 2 mo under both sheltered and
unsheltered conditions. Mean daily atmospheric SO2
concentration was 68.7 ,wg/m3 and several heavy
rainfalls.
Sample exposed in laboratory to 3 ppm SO2 and 95%
RH at 25 °C for 150 days. Samples were coated with
three carbonaceous particle samples from combustion
sources, and with activated carbon and graphite.
Samples exposed in sheltered ambient environment
for 6, 12, or 20 mo.
Samples exposed for 6 mo (cold and hot conditions)
in ambient environment. PM concentrations ranged
from 57.3 to 116.7 Mg/m3 (site 1) and 88 to
189.8 Mg/m3 (site 2). Some exposures also were
associated with high SO2, NO, and NO2.
Samples artificially exposed to fly-ash containing
1309.3 ,wg/m3 SO2 (0.5 ppm), at 95% RH and 25 °C
for 81 or 140 days. Fly-ash samples from five
different sources were used in study.
In the absence of moisture, little reaction is seen. SO2 is Haneef et al.
oxidized to sulfates in the presence of moisture. The effect (1993)
is enhanced 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.
Significant amounts of gypsum were noted on the Portland Viles (1990)
stone. Sheltered stones also showed soiling by
carbonaceous particles and other combustion products.
Etch holes and deep etching was noted in some of the
exposed unsheltered samples.
Exposure to particles from combustion processes enhanced Sabbioni et al.
sulfation of calcareous materials by SO2 because of metal (1996)
content of particles.
Carrara marble found to be more reactive with SO2 than Yerrapragada
Georgia marble possibly because of the compactness of the et al. (1994)
Georgia marble. Greater effects noted when samples were
also exposed to NO2.
Pollutant exposed samples showed increased weight gain Realini et al.
over that expected from natural weathering processes. (1995)
There was a blackening of stone samples exposed to
carbonaceous rich paniculate matter.
Exposure to fly-ash did not enhance oxidation of SO2 to Hutchinson
sulfates. Mineral oxides in fly ash contributed to sulphation et al. (1992)
ofCaCO3.
-------
fa
g.
to
TABLE 4-9 (cont'd). EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON STONE
Stone
Exposure Conditions
Comments
Bibliography
to
to
\*/
\
H
6
o
O
H
O
c
o
H
W
O
V
O
HH
H
W
Lime mortar
Pozzolan mortar
Cement mortar
Limestone
Travertine marble
Limestone
Quartz-cemented
sandstone
Calcite-cemented
sandstone
Granite
Brick
Limestone
Sandstone
Carrara marble
Travertine marble
Trani limestone
Portland limestone
Lime mortar
Pozzolan mortar
Cement mortar
Samples exposed to 7,856 ^g/m3 (3 ppm) SO2 at 100%
RH and 25 °C for 30, 60, or 90 days; samples sprayed
with bidistilled water every 7 days to simulate rainfall.
Samples exposed under actual ambient air conditions at
two locations in Rome. Monitoring data obtained for
SO2, NO, NO2, and total suspended particulates (TSP)
but not reported. Exposure was for four seasons.
Samples from structures exposed for varying periods of
time under ambient air conditions. Samples selected
because of black layer on surface.
Samples of ancient grey crust formed between 1180
and 1636 on the Church of Saint Trophime in Arks and
formed between 1530 and 1187 on the Palazz
d'Accursio in Bolonga.
Samples of the stones and mortars were representative
of those used in the past and currently for new
construction and restorations. Samples were exposed
for 6, 12, and 24 mo under ambient conditions in
Milan.
Exposure to SO2 produced significant quantities of
calcium sulfite and calcium sulfate on specimens;
however, the amount produced was dependent of the
porosity, specific surface, and alkalinity of the sample.
Zappia et al.
(1994)
TSP exposure increased the cleaning frequency for stone Lorusso et al.
monuments. Monuments are soiled proportionately (1997)
overtime, based on brightness values. Horizontal surfaces
showed higher graying values because of particle
sediment.
Black layers were found to be primarily comprised of iron Nord and
compounds, quartz, silicate, soot, and dirt. Ericsson (1993)
Crust samples contained calcite, soil dust, carbonaceous Ausset et al.
particles, and gypsum crystals. (1998)
Mortars were more reactive than the stones. Of the Zappia et al.
mortars, cement and pozzolan mortar were more reactive (1998)
than the lime mortar. Carrara marble was the least
reactive of the stones. The maximum amount of
degradation was found in areas sheltered from rain.
-------
g.
to
o
o
TABLE 4-9 (cont'd). EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON STONE
Stone
Exposure Conditions
Comments
Bibliography
Lime mortar
Sample of black crust taken from Zamboni Tower Gate.
Exposure to environmental pollutants caused the
formation of two separate layers on the mortar: an outer
thin surface black crust composed of gypsum and
carbonaceous particles and the inner composed of
products from the dissolution and sulphation of the
carbonate matrix in the mortar.
Sabbioni et al.
(1998)
to
oo
\*s
\
H
6
o
O
H
O
c
o
H
W
O
V
O
HH
H
W
Carrara marble Samples of crust removed from Milan General
Hospital, built around 1937.
Gypsum main component of crust followed by Bugini et al.
carbonaceous particles and iron oxides. Estimated rate of (2000)
crust formation was 2-5 /^m/year. Total amount of
gypsum formed over the lifetime of exposure was 5 to
13 mg/cm2, an estimated 0.2 mg/cmVyear.
-------
1 surface. Gypsum is a gray to black crusty material comprised mainly of calcium sulfate
2 dihydrate from the reaction of calcium carbonate (calcite) in the stone with atmospheric SO2 and
3 moisture (relative humidities exceeding 65%). Approximately 99% of the sulfur in gypsum is
4 sulfate because of the sulphation process caused by the deposition of SO2 aerosol. Sulphites also
5 are present in the gypsum layer as an intermediate product (Sabbioni et al., 19961; Ghedini et al.,
6 2000; Gobbi et al., 1998; Zappia et al., 1998). Gypsum is more soluble than calcite and is known
7 to form on limestone, sandstones, and marble when exposed to SO2. Gypsum also has been
8 reported to form on granite stone by replacing silicate minerals with calcite (Schiavon et al.,
9 1995). Gypsum occupies a larger volume than the original stone, causing the stone's surface to
10 become cracked and pitted. The rough surface serves as a site for deposition of airborne
11 particles.
12 The dark colored gypsum is caused by surface deposition of carbonaceous particles
13 (noncarbonate carbon) from combustion processes occurring in the area (Sabbioni, 1995;
14 Saiz-Jimenez, 1993; Ausset et al., 1998), trace metals contained in the stone, dust, and numerous
15 other anthropogenic pollutants. After analyzing damaged layers of several stone monuments,
16 Zappia et al. (1993) found that the dark-colored damaged surfaces contained 70% gypsum and
17 20% noncarbonate carbon. The lighter colored damaged layers were exposed to rain and
18 contained 1% gypsum and 4% noncarbonate carbon. It is assumed that rain removes reaction
19 products, permitting further pollutant attack of the stone monument, and likely redeposits some
20 of the reaction products at rain runoffs sites on the stone. Following sulfur compounds, carbon
21 was reported to be the next highest element in dark crust on historical monuments in Rome.
22 Elemental carbon and organic carbon accounted for 8 and 39% of the total carbon in the black
23 crust samples. The highest percentage of carbon, carbonate carbon, was caused by the carbonate
24 matrix in the stones. The high ratio of organic carbon to elemental carbon indicates the presence
25 of a carbon source other than combustion processes (Ghedini et al., 2000). Cooke and Gibbs
26 (1994) suggested that stones damaged during times of higher ambient pollution exposure likely
27 would continue to exhibit a higher rate of decay, termed the "memory effect", than newer stones
28 exposed under lower pollution conditions. Increased stone damage also has been associated with
29 the presence of sulfur oxidizing bacteria and fungi on stone surfaces (Garcia-Valles et al., 1998;
30 Young, 1996; Saiz-Jimenez, 1993; Diakumaku et al., 1995).
March 2001 4-124 DRAFT-DO NOT QUOTE OR CITE
-------
1 Dissolution of gypsum on the stone's surface initiates structural changes in the crust layer.
2 Garica-Valles et al. (1998) proposed a double mechanism; the dissolution of the gypsum, in the
3 presence of sufficient moisture, followed by recrystallization inside fissures or pores. In the
4 event of limited moisture, the gypsum in dissolved and recrystallizes at its original location.
5 According to the authors, this would explain the gypsum-rich crustal materials on stone surfaces
6 sheltered from precipitation.
7 Moisture was found to be the dominant factor in stone deterioration for several sandstones
8 (Petuskey et al., 1995). Dolske (1995) reported that the deteriorative effects of sulfur-containing
9 rain events, sulfates, and SO2 on marble were largely dependent on the shape of the monument or
10 structure rather than the type of marble. The author attributed the increased fluid turbulence over
11 a nonflat vertical surface versus a flat surface to the increased erosion. Sulfur-containing
12 particles also have been reported to enhance the reactivity of Carrara marble and Travertine and
13 Trani stone to SO2 (Sabbioni et al., 1992). Particles with the highest carbon content had the
14 lowest reactivity.
15 The rate of stone deterioration is determined by the pollutant and the pollutant
16 concentration, the stone's permeability and moisture content, and the pollutant deposition
17 velocity. Dry deposition of SO2 between rain events has been reported to be a major causative
18 factor in pollutant-related erosion of calcareous stones (Baedecker et al., 1991; Dolske, 1995;
19 Cooke and Gibbs, 1994; Schuster et al., 1994; Hamilton et al., 1995; Webb et al., 1992). Sulfur
20 dioxide deposition increases with increasing relative humidity (Spiker et al., 1992), but the
21 pollutant deposition velocity is dependent on the stone type (Wittenburg and Dannecker, 1992),
22 the porosity of the stone, and the presence of hygroscopic contaminants.
23 Although it is clear from the available information that gaseous pollutants, in particular dry
24 deposition of SO2 will promote the decay of some types of stones under the specific conditions,
25 carboneous particles (noncarbonate carbon) may help to promote the decay process by aiding in
26 the transformation of SO2 to a more acidic species (Del Monte and Vittori, 1985). Several
27 authors have reported enhanced sulfation of calcareous material by SO2 in the presence of
28 particles containing metal oxides (Sabbioni et al., 1996; Hutchinson et al., 1992).
29
30
31
March 2001 4-125 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.4.2 Soiling and Discoloration of Man-Made Surfaces
2 Ambient particles can cause soiling of man-made surfaces. Soiling has been defined as the
3 deposition of particles of less than 10 //m on surfaces by impingement. Soiling generally is
4 considered an optical effect, that is, soiling changes the reflectance from opaque materials and
5 reduces the transmissions of light through transparent materials. Soiling can represent a
6 significant detrimental effect requiring increased frequency of cleaning of glass windows and
7 concrete structures, washing and repainting of structures, and, in some cases, reduction in the
8 useful life of the object. Particles, in particular carbon, also may help catalyze chemical reactions
9 that result in the deterioration of materials during exposure.
10 It is difficult to determine the accumulated particle levels that cause an increase in soiling;
11 however, soiling is dependent on the particle concentration in the ambient environment, particle
12 size distribution, and the deposition rate and the horizontal or vertical orientation and texture of
13 the surface being exposed (Haynie, 1986). The chemical composition and morphology of the
14 particles and the optical properties of the surface being soiled will determine the time at which
15 soiling is perceived (Nazaroff and Cass, 1991). Carey (1959) reported that the average observer
16 could observe a 0.2% surface coverage of black particles on a white background. A recent study
17 suggest that it would take a 12% surface coverage by black particles before there is 100%
18 accuracy in identifying soiling (Bellan et al., 2000). The rate at which an object is soiled
19 increases linearly with time; however, as the soiling level increases, the rate of soiling decreases.
20 The buildup of particles on a horizontal surface is counterbalanced by an equal and opposite
21 depletion process. The depletion process is based on the scouring and washing effect of wind
22 and rain (Schwar, 1998).
23
24 4.4.2.1 Stones and Concrete
25 Most of the research evaluating the effects of air pollutants on stone structures have
26 concentrated on gaseous pollutants. The deposition of the sulfur-containing pollutants are
27 associated with the formation of gypsum on the stone (see Section 4.4.1.3). The dark color of
28 gypsum is attributed to soiling by carbonaceous particles from nearby combustion processes.
29 A lighter gray colored crust is attributed to soil dust and metal deposits (Ausset et al., 1998;
30 Camuffo, 1995; Moropoulou et al., 1998). Realini et al. (1995) found the formation of a dark
31 gypsum layer and a loss of luminous reflection in Carrara marble structures exposed for 1 year
March 2001 4-126 DRAFT-DO NOT QUOTE OR CITE
-------
1 under ambient air conditions. Dark areas of gypsum were found by McGee and Mossitti (1992)
2 on limestone and marble specimens exposed under ambient air conditions for several years. The
3 black layers of gypsum were located in areas shielded from rainfall. Particles of dirt were
4 concentrated around the edges of the gypsum formations. Lorusso et al. (1997) attributed the
5 need for frequent cleaning and restoration of historic monuments in Rome to exposure to total
6 suspended particulates. They also concluded that, based on a decrease in brightness (graying),
7 surfaces are soiled proportionately over time; however, graying is higher on horizontal surfaces
8 because of sedimented particles. Davidson et al. (2000) evaluated the effects of air pollution
9 exposure on a limestone structure on the University of Pittsburgh campus using estimated
10 average TSP levels in the 1930s and 1940s and actual values for the years 1957 to 1997.
11 Monitored levels of SO2 were available for the years 1980 to 1998. Based on the available data
12 on pollutant levels and photographs, it was thought that soiling began while the structure was
13 under construction. With decreasing levels of pollution, the soiled areas have been slowly
14 washed away, the process taking several decades, leaving a white, eroded surface. Studies
15 describing the effects of particles on stone surfaces are discussed in Table 4-9.
16
17 4.4.2.2 Household and Industrial Paints
18 Few studies are available that evaluate the soiling effects of particles on painted surfaces.
19 Particles composed of elemental carbon, tarry acids, and various other constituents are
20 responsible for soiling of structural painted surfaces. Coarse-mode particles (>2.5 //m) initially
21 contribute more soiling of horizontal and vertical painted surfaces than do fine-mode particles
22 (<2.5 //m), but are more easily removed by rain (Haynie and Lemmons, 1990). The
23 accumulation of fine particles likely promotes remedial action (i.e., cleaning of the painted
24 surfaces). Coarse-mode particles are primarily responsible for soiling of horizontal surfaces.
25 Rain interacts with coarse particles, dissolving the particle and leaving stains on the painted
26 surface (Creighton et al., 1990; Haynie and Lemmons, 1990). Haynie and Lemmons (1990)
27 proposed empirical predictive equations for changes in surface reflectance of gloss-painted
28 surfaces that were exposed protected and unprotected from rain and oriented horizontally and
29 vertically.
30 Early studies by Parker (1955) and Spence and Haynie (1972) demonstrated an association
31 between particle exposure and increased frequency of cleaning of painted surfaces. Particle
March 2001 4-127 DRAFT-DO NOT QUOTE OR CITE
-------
1 exposures also caused physical damage to the painted surface (Parker, 1955). Unsheltered
2 painted surfaces are initially more soiled by particles than sheltered surfaces but the effect is
3 reduced by rain washing. Reflectivity is decreased more rapidly on glossy paint than on flat paint
4 (Haynie and Lemmons, 1990). However, surface chalking of the flat paint was reported during
5 the exposure. The chalking interfered with the reflectance measurements for particle soiling.
6 Particle composition measurements that were taken during exposure of the painted surfaces
7 indicated sulfates to be a large fraction of the fine mode and only a small fraction of the coarse
8 mode. Although no direct measurements were taken, fine mode particles likely also contained
9 large amounts of carbon and possibly nitrogen or hydrogen (Haynie and Lemmons, 1990).
10
11
12 4.5 EFFECTS OF ATMOSPHERIC PARTICIPATE MATTER ON
13 CLIMATE CHANGE PROCESSES AND THEIR POTENTIAL
14 HUMAN HEALTH AND ENVIRONMENTAL IMPACTS
15 Global climate change processes and their potential human health and environmental
16 impacts have been accorded extensive attention during the past several decades, and they still
17 continue to be of broad national and international concern. This is reflected by extensive
18 research and assessment efforts undertaken since the mid-1970s by U.S. Federal Government
19 Agencies (e.g., NOAA, EPA, CDC, etc.) or via U.S. Federal Interagency programs (e.g., the U.S.
20 Global Climate Change Research Program [USGCRP]) and by analogous extensive research and
21 assessment efforts undertaken by numerous other national governments or international
22 collaborative activities (e.g., those coordinated by the Intergovernmental Panel on Climate
23 Change [IPCC], established in the 1980s under the joint auspices of the World Meteorological
24 Organization [WMO], and the United Nations Environment Programme [UNEP]).
25 Atmospheric particles play important roles in two key types of global climate change
26 processes or phenomena: (1) alterations in the amount of solar radiation in the ultraviolet range
27 (especially UV-B) penetrating through the Earth's atmosphere and reaching its surface, where it
28 can exert a variety of effects on human health, plant and animal biota, and other environmental
29 components; and (2) alterations in the amount of solar radiation in the visible range being
30 transmitted through Earth's atmosphere and either being reflected back into space or absorbed
31 (together with trapping of infrared radiation emitted by the Earth's surface by certain gases),
March 2001 4-128 DRAFT-DO NOT QUOTE OR CITE
-------
1 which enhances heating of the Earth's surface and lower atmosphere (i.e., the widely-known
2 "greenhouse effect") and leads to consequent "global warming" impacts on human health and the
3 environment. Atmospheric particles also play a lesser role by absorbing infrared radiation
4 emitted by the Earth's surface.
5 The effects of atmospheric PM on the transmission of electromagnetic radiation emitted by
6 the sun at ultraviolet and visible wavelengths and by the earth at infrared wavelengths depend on
7 the radiative properties (extinction efficiency, single scattering albedo, and asymmetry
8 parameter) of the particles, which are, in turn, dependent on the size and shape of the particles,
9 the composition of the particles and the distribution of components within individual particles.
10 In general, the radiative properties of particles are size and wavelength dependent. In addition,
11 the extinction cross-section tends to be at a maximum when the particle radius is similar to the
12 wavelength of the incident radiation. Thus, fine particles present mainly in the accumulation
13 mode would be expected to exert a greater influence on the transmission of electromagnetic
14 radiation than would coarse particles. The composition of particles can be crudely summarized
15 in terms of the broad classes identified in Chapter 6 of the 1996 PM AQCD and recapitulated in
16 Chapter 2 of this document (e.g., fine particles mainly consisting of nitrate, sulfate, mineral dust,
17 elemental carbon, organic carbon compounds [e.g., PAHs], and metals derived from high
18 temperature combustion or smelting processes). The major sources of these components are
19 shown in Table 2.1 of Chapter 2 in this document.
20 Knowledge of the factors controlling the transfer of solar radiation in the ultraviolet
21 spectral region is needed for assessing the potential biological and environmental impacts
22 associated with exposure to UV-B radiation (290 to 315 nm). Knowledge of the effects of PM
23 on the transfer of radiation in the visible and infrared spectral regions is needed for assessing the
24 relation between particles and global warming and its environmental and biological impacts.
25 Key information regarding important conceptual aspects and factors related to solar ultraviolet
26 radiation processes and effects is summarized first below and atmospheric PM roles noted,
27 followed by summarization of global warming processes, their potential human health and
28 environmental impacts, and potential relationships to atmospheric PM.
29
30
31
March 2001 4-129 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.5.1 Solar Ultraviolet Radiation Transmission Impacts on Human Health
2 and the Environment: Atmospheric Particulate Matter Effects
3 4.5.1.1 Bases for Concern Regarding Increased Ultraviolet Radiation Transmission
4 The transmission of solar UV-B radiation through the earth's atmosphere is controlled by
5 ozone, clouds, and particles. The depletion of stratospheric ozone caused by the release of
6 anthropogenically produced chlorine (Cl)-and bromine (Br)-containing compounds has resulted
7 in heightened concern over potentially serious increases in the amount of solar UV-B radiation
8 (SUVB) reaching the Earth's surface. SUVB is also responsible for initiating the production of
9 OH radicals that oxidize a wide variety of volatile organic compounds, some of which can
10 deplete stratospheric ozone (e.g., CH3C1, CH3Br), absorb terrestrial infrared radiation (e.g., CH4),
11 and contribute to photochemical smog formation (e.g., C2H4, C5H8 ).
12 Increased penetration of SUVB to the Earth's surface as the result of stratospheric ozone
13 depletion continues to be of much concern because of proj ections of consequent increased
14 surface-level SUVB exposure and associated potential negative impacts on human health, plant
15 and animal biota, and man-made materials. Several summary overviews (Kripke, 1989; Grant,
16 1989; Kodama and Lee, 1993; Van der Leun et al., 1995, 1998) of salient points related to
17 stratospheric ozone depletion processes and bases for concern provide a concise introduction to
18 the subject, as does Figure 4-25. As shown to the left in the figure, stratophospheric ozone
19 depletion results from: (a) anthropogenic production and associated emission into the lower
20 atmosphere of certain trace gases having long atmospheric residence times (e.g.,
21 chlorofluorocarbons [CFCs], carbon tetrachloride [CC14], and Halon 1211 [CF2C1 Br] and 1301
22 [CF3Br], which have atmospheric residence times of 75 to 100 years, 50 years, 25 years, and
23 110 years, respectively); (b) their tropospheric accumulation and gradual transport, over decades,
24 up to the stratosphere, where (c) photodissociation processes release Cl and Br, that (especially
25 under very cold subzero upper atmospheric conditions) catalyze ozone reduction; leading to
26 (d) stratospheric ozone depletion that is most marked over Antarctica during Southern
27 Hemisphere wintertime, to a less marked but still significant extent over the Arctic Polar Region
28 during Northern Hemisphere wintertime, and to a lesser extent over mid-latitude regions during
29 any season.
30 Given the long time involved in transport of such gases to the stratosphere and their long
31 residence times there, any effects already seen on stratospheric ozone are likely caused by the
March 2001 4-130 DRAFT-DO NOT QUOTE OR CITE
-------
BASES FOR CONCERN ABOUT STRATOSPHERIC OZONE DEPLETION
DUE TO CFC's, HALONS, AND OTHER TRACE GASES
Stratospheric
Ozone Depletion
Cl, Br Catalyze
Ozone Reduction
Photodissociation
Releases Cl and Br
Slow Transport
to Stratosphere
Tropospheric
Accumulation
Air Emissions of
CFC's, Halons, etc.
OZONE DEPLETION EFFECTS
/CFC's &O3Column\
\ Reorganization j
I
Climate Changes:
Temp., Winds, etc.
Increase of Air
Stagnation Periods
Accumulation of
Tropospheric O3
and Acid Aerosols
Increased UV-B Light
Penetration to Surface
Altered Bio-
geochemical
Cycling
UV-B Radiation Direct
Human Health Impacts
Immune
Suppresssion
Natural Ecosystem
and Agriculture
Impacts
Skin Damage
(Sunburn)
Damage
to Eye
Terrestrial
Ecosystem Shifts
Lower Crop Yields
AND
Man's Production
of CFC's, Halons,
Other Trace Gasses
Environmental
Effects: Crop,
Forest Damage
Human
Health
Effects
Infectious
Diseases
Increased
Skin Cancer
Premature
Skin Aging
Cataracts
Incidence
Increased
Aquatic
Ecosystem Shifts
Less Plankton &
Seafood
Figure 4-25. Processes involved in stratospheric ozone depletion because of man's
production of CFCs, halons, and other trace gases are shown to the left. The
types of effects caused by stratospheric ozone depletion and consequent
increased UV-B penetration to the Earth's surface are hypothesized to include
both direct effects on human health (e.g., increased cancer rates, immune
suppression, etc.) and other terrestrial and aquatic ecological effects resulting
from increased UV-B alterations of biogeochemical cycles.
Source: Adapted from Grant (1989).
1 atmospheric loadings of trace gases from anthropogenic emissions several decades ago, and those
2 gases already in the atmosphere may continue to exert stratospheric ozone depletion effects well
3 into the 21st century. Shorter lived gases, such as CH3Br, also exert significant ozone depletion
4 effects.
March 2001
4-131
DRAFT-DO NOT QUOTE OR CITE
-------
1 The main types of effects hypothesized as likely to result from stratospheric ozone
2 depletion and consequent increased SUVB penetration through the Earth's atmosphere include
3 the following.
4 (1) Direct Human Health Effects, such as skin damage (sunburn), leading to more rapid aging
5 and increased incidence of skin cancer; ocular effects (retinal damage and increased cataract
6 formation possibly leading to blindness); and suppression of some immune system
7 components (contributing to skin cancer induction and spread to nonirradiated skin areas, as
8 well as possibly increasing susceptibility to certain infectious diseases or decreasing
9 effectiveness of vacinations).
10 (2) Agricultural/Ecological Effects, mediated largely through altered biogeochemical cycling
11 resulting in consequent damaging impacts on terrestrial plants (leading to possible reduced
12 yields of rice, other food crops, and commercially important trees, as well as to biodiversity
13 shifts in natural terrestrial ecosystems); and deleterious effects on aquatic life (including
14 reduced ocean zooplankton and phytoplankton, as important base components of marine
15 food-chains supporting the existence of commercially important, edible fish and other
16 seafood, as well as to other aquatic ecosystem shifts).
17 (3) Indirect Human Health and Ecological Effects, mediated through increased tropospheric
18 ozone formation (and consequent exacerbation of surface-level, ozone-related health and
19 ecological impacts) and alterations in the concentrations of other important trace species,
20 most notably the hydroxyl radical and acidic aerosols.
21 (4) Other Types of Effects, such as faster rates of polymer weathering because of increased
22 UV-B radiation and other effects on man-made commercial materials and cultural artifacts,
23 secondary to climate change or exacerbation of air pollution problems.
24 Extensive qualitative and quantitative characterizations of stratospheric ozone depletion
25 processes and projections of their likely potential impacts on human health and the environment
26 have been the subjects of periodic (1988, 1989, 1991, 1994, 1998) international assessments
27 carried out under WMO and UNEP auspices since the 1987 signing of the Montreal Protocol on
28 Substances that Deplete the Ozone Layer. The reader is referred for more detailed up-to-date
29 information to the two most recently completed international assessments of processes
30 contributing to stratospheric ozone depletion and the status of progress towards ameliorating the
31 problem (WMO, 1999) and revised qualitative and quantitative projections of likely consequent
March 2001 4-132 DRAFT-DO NOT QUOTE OR CITE
-------
1 human health and environmental effects (UNEP, 1998). (See Appendices 4A and 4B for
2 synopses of key points abstracted from the executive summaries of these assessments).
3 Of considerable importance is the growing recognition, as reflected in these newer
4 assessments, of impacts of enhanced solar radiation on biogeochemical cycles (see, for example,
5 Zepp et al., 1998, and earlier discussions in this chapter [Sections 4.2.2.1 and 4.2.2.2]). As noted
6 in the Zepp et al. paper, the effects of UV-B radiation (both in magnitude and direction) on trace
7 gas (e.g., CO) emissions and mineral nutrient cycling are species specific and can affect a variety
8 of processes. These include, for example, changes in the chemical composition of living plant
9 tissue, photodegradation of dead plant matter (e.g., ground litter), release of CO from vegetation
10 previously charred by fire, changes in microbial decomposer communities, and effects on
11 nitrogen-fixing microorganisms and plants. Also, studies of natural acquatic ecosystems indicate
12 that organic matter is the primary determinant of UV-B penetration through water. Organic
13 matter changes, caused by enhanced UV-B penetration and augmented by acidification and
14 climate change, contribute to clarification of water and changes in light quality that broadly
15 impact the effects of UV-B on aquatic biogeochemical cycles. Enhanced UV-B levels have both
16 positive and negative impacts on aquatic ecosystem microbial activities that can affect nutrient
17 cycling and the uptake or release of greenhouse gases. Thus, there are emerging complex issues
18 regarding interactions and feedbacks between climate change and changes in terrestrial and
19 marine biogeochemical cycles because of increased UV-B penetration to the Earth's surface.
20 As noted in the above detailed assessments, since the signing of the Montreal Protocol,
21 much progress has been made in reducing emissions of ozone depleting gases, leading to
22 estimates of the maximum extent of stratospheric ozone depletion as likely having been reached
23 in the year 2000, to be followed by gradual lessening of the problem and its impacts during the
24 next half-century. However, the assessments also note that the modeled projections are subject
25 to considerable uncertainty. The role of atmospheric particles, discussed below, is one of
26 numerous salient factors complicating modeling efforts.
27
28 4.5.1.2 Airborne Particle Impacts on Atmospheric Ultraviolet Radiation Transmission
29 A given amount of ozone in the lower troposphere has been shown to absorb more solar
30 radiation than an equal amount of ozone in the stratosphere because of the increase in its
31 effective optical path produced by Rayleigh scattering in the lower atmosphere (Bruehl and
March 2001 4-133 DRAFT-DO NOT QUOTE OR CITE
-------
1 Crutzen, 1988). The effects of particles are more complex. The impact of particles on the SUVB
2 flux throughout the boundary layer are highly sensitive to the altitude of the particles and to their
3 single scattering albedo. Even the sign of the effect can reverse as the composition of the particle
4 mix changes from scattering to absorbing types (e.g., from sulfate to elemental carbon or PAHs)
5 (Dickerson et al., 1997). In addition, scattering by particles also may increase the effective
6 optical path of absorbing molecules, such as ozone, in the lower atmosphere.
7 The effects of particles present in the lower troposphere on the transmission of SUVB have
8 been examined both by field measurements and by radiative transfer model calculations. The
9 presence of particles in urban areas modifies the spectral distribution of solar irradiance at the
10 surface. Shorter wavelength radiation (i.e., in the ultraviolet) is attenuated more than visible
11 radiation (e.g., Peterson et al., 1978; Jacobson, 1999). Wenny et al. (1998) also found greater
12 attenuation of SUVB than SUVA (315 to 400 nm). However, this effect depends on the nature
13 of the specific particles involved and, therefore, is expected to depend strongly on location.
14 Lorente et al. (1994) observed an attenuation of SUVB ranging from 14 to 37%, for solar zenith
15 angles ranging from about 30° to about 60°, in the total (direct and diffuse) SUVB reaching the
16 surface in Barcelona during cloudless conditions on very polluted days (aerosol scattering optical
17 depth at 500 nm, 0.46 < T500nm < 1.15) compared to days on which the turbidity of urban air was
18 similar to that for rural air (T500nm < 0.23). Particle concentrations that can account for these
19 observations can be estimated roughly by combining Koschmeider's relation for expressing
20 visual range in terms of extinction coefficient with one for expressing the mass of PM2 5 particles
21 in terms of visual range (Stevens et al., 1984). By assuming a scale height (i.e., the height at
22 which the concentration of a substance falls off to 1/e of its value at the surface) of 1 km for
23 PM2 5, an upper limit of 30 //g/ m3 can be derived for the clear case and between 60 and
24 150 //g/m3 for the polluted case. Estupinan et al. (1996) found that summertime haze under clear
25 sky conditions attenuates SUVB between 5 and 23% for a solar zenith angle of 34°, compared to
26 a clear sky day in autumn. Minis (1996) measured a decrease in SUVB by about 80% downwind
27 of major biomass burning areas in Amazonia in 1995. This decrease in transmission
28 corresponded to optical depths at 340 nm ranging from three to four. Justus and Murphey (1994)
29 found that SUVB reaching the surface decreased by about 10% because of changes in aerosol
30 loading in Atlanta, GA, from 1980 to 1984. Also, higher particle levels in Germany (48 °N) may
March 2001 4-134 DRAFT-DO NOT QUOTE OR CITE
-------
1 be responsible for greater attenuation of SUVB than in New Zealand (Seckmeyer and McKenzie,
2 1992).
3 In a study of the effects of nonurban haze on SUVB transmission, Wenny et al. (1998)
4 derived a very simple regression relation between the measured aerosol optical depth at 312 nm,
5
6 ln( SUVB transmission at solar noon) = -0.1422 T312nm - 0.138, R2 = 0.90,
7
8 and the transmission of SUVB to the surface. In principle, values of T312nm could be found from
9 knowledge of the aerosol optical properties and visual range values. Wenny et al. (1998) also
10 found that absorption by particles accounted for 7 to 25% of the total (scattering + absorption)
11 extinction. Relations such as the above one are strongly dependent on local conditions and
12 should not be used in other areas without knowledge of the differences in aerosol properties.
13 Although all of the above studies reinforce the idea that particles play a major role in modulating
14 the attenuation of SUVB, none included measurements of ambient PM concentrations, so direct
15 relations between PM levels and SUVB transmission could not be determined.
16 Liu et al. (1991) estimated, roughly, overall effects of increases of anthropogenic airborne
17 particles that have occurred since the beginning of the industrial revolution on atmospheric
18 transmission of SUVB. Based on (a) estimates of the reduction in visibility from about 95 km to
19 about 20 km over nonurban areas in the eastern United States and in Europe, (b) calculations of
20 optical properties of airborne particles found in rural areas to extrapolate the increase in
21 extinction at 550 to 310 nm, and (c) radiative transfer model calculations, Liu et al. concluded
22 that the amount of SUVB reaching Earth's the surface likely has decreased from 5 to 18% since
23 the beginning of the industrial revolution. This was attributed mainly to scattering of SUVB
24 back to space by sulfate containing particles. Radiative transfer model calculations have not
25 been done for urban particles.
26 Although aerosols are expected to decrease the flux of SUVB reaching the surface,
27 scattering by particles is expected to result in an increase in the actinic flux within and above the
28 aerosol layer. However, when the particles significantly absorb SUVB, a decrease in the actinic
29 flux is expected. Actinic flux is the radiant energy integrated over all directions at a given
30 wavelength incident on a point in the atmosphere, and is the quantity needed to calculate rates of
31 photolytic reactions in the atmosphere. Blackburn et al. (1992) measured attenuation of the
March 2001 4-135 DRAFT-DO NOT QUOTE OR CITE
-------
1 photolysis rate of ozone and found that aerosol optical depths near unity at 500 nm reduced
2 ozone photolysis rate by as much as a factor of two. Dickerson et al. (1997) showed that the
3 photolysis rate for NO2, a key parameter for calculating the overall intensity of photochemical
4 activity, could be increased within and above a scattering aerosol layer extending from the
5 surface, although it would be decreased at the surface. This effect is qualitatively similar to what
6 is seen in clouds, where photolysis rates are increased in the upper layers of a cloud and above
7 the cloud (Madronich, 1987). For a simulation of an ozone episode that occurred during July
8 1995 in the Mid-Atlantic region, Dickerson et al. (1997) calculated ozone increases of up to
9 20 ppb compared to cases that did not include the radiative effects of particles in urban airshed
10 model (UAM-IV) simulations. In contrast, Jacobson (1998) found that particles may have
11 caused a 5 to 8% decrease in O3 levels during the Southern California Air Quality Study in 1987.
12 Absorption by organic compounds and nitrated inorganic compounds was hypothesized to
13 account for the reductions in UV radiation intensity.
14 The photolysis of ozone in the Hartley bands also leads to production of electronically
15 excited oxygen atoms, O(JD) that then react with water vapor to form OH radicals. Thus,
16 enhanced photochemical production of ozone is accompanied by the scavenging of species
17 involved in greenhouse warming and stratospheric depletion. However, these effects may be
18 neutralized or even reversed by the presence of absorbing material in the particles. Any
19 evaluation of the effects of particles on photochemical activity therefore will depend on the
20 composition of the particles and also will be location-specific.
21 Also complicating any straightforward evaluation of UV-B penetration to specific areas of
22 the Earth's surface are the influences of clouds, as discussed by Erlick et al. (1998), Frederick
23 et al. (1998), and Soulen and Fredrick (1999). Varying estimations of atmospheric transmission
24 of UV and visible spectrum light are obtained for cloudy atmospheres, depending on presence of
25 aerosols and the extent of their external or internal mixing with cloud droplets. Even in
26 situations of very low atmospheric PM (e.g., over Antarctica), interannual variations in
27 cloudiness over specific areas can be as important as ozone levels in determining UV surface
28 irradiation, with net impacts varying from a month or season to another (Soulen and Fredrick,
29 1999).
30 Given the above considerations, quantitation of projected effects of variations in
31 atmospheric PM on human health or the environment because of particle impacts on transmission
March 2001 4-136 DRAFT-DO NOT QUOTE OR CITE
-------
1 of solar UV-B would require location-specific evaluations, taking into account composition,
2 concentration, and internal structure of the particles; temporal variations in atmospheric mixing
3 heights and depths of layers containing the particles; and consequent impacts on surface-level
4 exposures of humans, ecosystem constituents, or man-made materials. The outcome of such
5 modeling effects would likely vary from location to location in terms of increased or decreased
6 surface level UV-B exposures because of location-specific changes in atmospheric PM
7 concentrations or composition. For example, to the extent that any location-specific scattering by
8 airborne PM were to affect the directional characteristics of UV radiation at ground level, and
9 thereby enhance radiation incident from low angles (Dickerson, 1997), the biological
10 effectiveness of resulting ground-level UV-B exposures could be enhanced. Airborne PM also
11 can reduce the ground-level ratio of photorepairing radiation (UV-A and short-wavelength
12 visible) to damaging UV-B radiation. Lastly, PM deposition is a major source of PAH in certain
13 freshwater lakes and coastal areas, and the adverse effects of solar UV are enhanced by uptake of
14 PAH by aquatic organisms. Thus, although airborne PM may, in general, tend to reduce ground-
15 level UV-B, its net effect in some locations may be to increase UV damage to certain aquatic and
16 terrestrial organisms, as discussed by Cullen and Neale (1997).
17
18 4.5.2 Global Warming Processes, Human Health and Environmental
19 Impacts, and Atmospheric Particle Roles
20 4.5.2.1 Bases for Concern Regarding Global Warming and Climate Change
21 Various trace gases emitted because of man's activities, including several noted above as
22 contributing to stratospheric ozone depletion, can act as "greenhouse gases" (GHG). That is, as
23 their tropospheric concentrations increase, they retard the escape of infrared radiation from the
24 earth's surface and thereby contribute to the trapping of heat near the surface (the "greenhouse
25 effect") and, ultimately, to consequent global warming and climate change. Much concern has
26 evolved with regard to increases in the naturally very low concentrations in the atmosphere of
27 some of these gases, especially carbon dioxide (CO2), nitrous oxide (N2O), methane (CH4),
28 chloroflurocarbons (CFCs), and tropospheric ozone (O3).
29 Atmospheric processes involved in mediating global warming and its likely consequent
30 effects have been reviewed extensively previously (United Nations Environment Programme,
31 1986; World Meteorological Organization, 1988; U.S. Environmental Protection Agency, 1987;
March 2001 4-137 DRAFT-DO NOT QUOTE OR CITE
-------
1 IPCC, 1996, 1998; US GCRP, 2000) and more concisely summarized by others (e.g., Grant
2 1989; Patz et al., 2000a; Patz et al., 2000b). The main focus here is (a) to provide first a very
3 brief summary of key points regarding processes involved and types of effects proj ected as likely
4 to be associated with global warming and climate change and, then, (b) to undertake discussion
5 of salient considerations regarding potential impacts of atmospheric PM on such processes and
6 effects.
7 All of the above noted assessments and summaries emphasize that estimating likely future
8 global warming trends and associated climate change caused by greenhouse gases is extremely
9 complex, with modeling results being highly dependent on key assumptions about the rates of
10 future increases in various gases and numerous other factors (including particle effects).
11 Modeling of the magnitude of the warming directly associated with radiative forcing by
12 greenhouse gases (without feedback enhancement) projects temperature increases, for example,
13 of about 1.2 °C for a doubling of CO2; another 0.45 °C for a simultaneous doubling of N2O and
14 CH4; and an additional 0.15 °C from a uniform 1-ppb increase in atmospheric concentrations of
15 CFC-11 and CFC-12. Indirect effects (feedbacks) that likely would increase temperatures further
16 are expected to occur. Increased water vapor (trapping heat) and snow and ice melting (reducing
17 reflection of radiation back into space) are two examples of such feedback factors expected to
18 increase temperatures. However, major uncertainties exist with regard to feedbacks between
19 global warming and clouds, which could either amplify or, perhaps, reduce a temperature rise.
20 Taking assumptions about rates of increase (or decrease) in GHG concentrations, consequent
21 initial warming effects, feedback effects, and accompanying uncertainties into account, numerous
22 modeling efforts have attempted to project likely future trends in global warming. Despite the
23 complexity and uncertainties inherent in such modeling efforts, all typically agree that some
24 global warming has occurred and will continue to occur during the coming decades, but the
25 ranges of quantitative estimates vary considerably depending on specific assumptions
26 incorporated into the models. Thus, for example, "low" scenarios assuming stabilization or
27 reductions in GHG emissions (resulting from implementation of the 1987 Montreal Protocol)
28 project lower temperature changes than other scenarios assuming higher rates of increase in GHG
29 emissions or differing feedback-effect patterns.
30 Given the wide range of estimates of global warming trends and patterns of associated
31 climate change emerging from modeling efforts, the estimation of likely human health and
March 2001 4-138 DRAFT-DO NOT QUOTE OR CITE
-------
1 ecological effects associated with global warming on any quantitative basis is extremely difficult.
2 The onset of any notable global warming effect is also important, with various analyses
3 indicating that global temperatures for the past century have been rising (and now appear to be
4 beyond average levels within the range of variation seen with cycles of global warming or
5 cooling over the past several centuries before marked anthropogenic emissions of greenhouse
6 gases occurred). Also posing difficulties for the quantitative estimation of human health and
7 other effects are expected wide regional variations in temperature and climate characteristics
8 (e.g., rain and snowfall amounts) that may be projected reasonably to result from various global
9 warming trend scenarios. Lastly, it should be noted that, despite general warming trends in
10 long-term average temperatures, wide extremes in both high and low temperatures also are
11 expected to occur more frequently in some areas.
12 A Special Report of the IPCC Working Group II on Regional Impacts of Climate Change:
13 An Assessment of Vulnerabilities (IPCC, 1998) assesses global warming processes and identifies
14 several types of vulnerabilities likely to occur because of climate change resulting from global
15 warming. Such general types of vulnerabilities include impacts on terrestrial and aquatic
16 ecosystems, hydrology and water resources, food and fiber production, coastal systems, and
17 human health. Appendix 4C provides excerpts of materials from the executive summary of the
18 IPCC (1998) report that comprise a helpful overview of key points regarding projected global
19 warming processes, likely climate change patterns, and their consequent impacts in terms of the
20 types of vulnerabilities noted above.
21 The IPCC (1998) report notes that human activities resulting in emissions of long-lived
22 GHCs are projected by General Circulation Models (GCMs) to lead to global and regional
23 changes in temperature, precipitation and other climate variables—resulting in increases in
24 global mean sea level; prospects for more extreme weather events, floods, and droughts in some
25 areas; and consequent changes in soil moisture. Based on various scenarios of current and
26 plausible future emissions of GHGs and aerosols and the range of sensitivities of climate change
27 to atmospheric levels (and residence time) of GHGs, GCMs project mean annual global surface
28 temperature increases in the range of 1 to 3.5 °C by 2100, a global mean sea level rise of 15 to
29 95 cm, and significant changes in spatial and temporal patterns of precipitation. The average rate
30 of warming will be more rapid than any seen in the past 10,000 years, although regional changes
31 could differ substantially from mean global rates.
March 2001 4-139 DRAFT-DO NOT QUOTE OR CITE
-------
1 Human health, ecosystems, and socioeconomic sectors (e.g., hydrology and water
2 resources, food and fiber production, etc.) are projected to be vulnerable to the magnitude and
3 rate of climate change, as well as increased climate variability. Wide variations in the courses
4 and net impacts of climate change in different geographic areas can be expected, and, although
5 many regions are likely to experience severe adverse impacts (some possibly irreversible) of
6 climate change, some climate change impacts may be locally beneficial in some regions.
7 In general, projected climate change impacts can be expected to represent additional stresses on
8 those natural ecosystems and human societal systems already impacted by increasing resource
9 demands, unsustainable resource management practices, and pollution, with wide variation likely
10 across regions and nations in their ability to cope with consequent alterations in ecological
11 balances, in availability of adequate food, water, and clean air, and in human health and safety.
12 Appendix 4C also includes excerpts from the executive summary of the IPCC 1998 special report
13 regarding the assessment of different types of vulnerabilities to climate change projected for each
14 of 10 different geographic regions of the Earth, with emphasis being placed in Appendix 4C on
15 those projected for two regions (North America and Polar) of most relevance to the continental
16 United States and Alaska.
17 Appendix 4C notes that (a) the characteristics of subregions and sectors of North America
18 suggest that neither impacts of climate change nor response options will be uniform, and (b)
19 many systems of North America are moderately to highly sensitive to climate change, with the
20 range of estimated effects including the potential for substantial damage or, conversely, the
21 potential for some beneficial outcomes. The most vulnerable continental United States sectors
22 and regions include long-lived natural forest ecosystems in the East and interior West, water
23 resources in the southern plains, agriculture in the Southeast and southern plains, northern
24 ecosystems and habitats, estuaries and beaches in developed areas, and low-latitude cool and cold
25 water fisheries. Other sectors or subregions may benefit from warmer temperatures or increased
26 CO2 fertilization (e.g., west coast coniferous forests; some western rangelands; reduced energy
27 costs for heating in northern latitudes; reduced road salting and snow-clearance costs; longer
28 open-water seasons in norther channels and ports; and agriculture in the northern latitudes, the
29 interior West, and the west coast). For Alaska, substantial shifts in ecosystems (with possible
30 major declines or loss of some sensitive species like bear and caribou or of other ice-dependent
31 animals) may occur in parallel to beneficial effects such as opening of ice-bound water
March 2001 4-140 DRAFT-DO NOT QUOTE OR CITE
-------
1 transportation routes or possible expanded agricultural viability secondary to longer growing
2 seasons. On the other hand, for North America, the potential for mainly deleterious direct or
3 indirect effects on human health is likely to increase (e.g., increased mortality directly linked to
4 temperature extremes, increases in incidence and spread of vector-borne infectious diseases,
5 impacts secondary to sea-level rise, and impacts secondary to increased tropospheric air pollution
6 [as depicted in Figure 4-26]).
7 More detailed evaluations of possible global climate change impacts on various U.S.
8 geographic areas are being conducted by the United States Global Change Research Program
9 (USGCRP). An overview report on the assessment results and key findings from a series of
10 workshops convened by the USGCRP National Assessment Synthesis team (NAST) has been
11 prepared (USGCRP, 2000). Selected highly salient key points from the report and subsidiary
12 regional assessments are presented in Appendix 4D. Overall key findings from the USGCRP
13 (2000) report are noted below.
14 (1) Increased Warming. Assuming continued growth in world GHG emissions, the primary
15 climate models used in the USGCRP assessment project that temperatures in the United
16 States will rise by 5 to 10 °F (3 to 6 °C) on average during the next 100 years.
17 (2) Differ ing Regional Impacts. Climate change will vary widely across the United States.
18 Temperature increases will vary somewhat from region to region. Heavy and extreme
19 precipitation events are likely to become more frequent, yet some regions will get drier.
20 The potential impacts of climate change will vary widely across the nation.
21 (3) Vulnerable Ecosystems. Many ecosystems are highly vulnerable to the projected rate and
22 magnitude of climate change. A few, such as alpine meadows in the Rocky Mountains and
23 some barrier islands, are likely to disappear entirely in some areas, with others, such as
24 some forests of the Southeast, being likely to experience major species shifts or break up.
25 Goods and services lost through disappearance or fragmentation of certain ecosystems are
26 likely to be costly or impossible to replace.
27 (4) Widespread Water Concerns. Water is an issue in every region, but the nature of the
28 vulnerabilities varies, with different nuances in each. Drought is an important concern in
29 every region. Floods and water quality are concerns in many regions. Snowpack changes
30 are especially important in the West, the Pacific Northwest, and Alaska.
March 2001 4-141 DRAFT-DO NOT QUOTE OR CITE
-------
BASES FOR CONCERN ABOUT GLOBAL WARMING AND CLIMATE
CHANGE EFFECTS ON THE ENVIRONMENT AND HUMAN HEALTH
GLOBAL WARMING
AND CLIMATE CHANGE
I Temperature I
I Extremes I
Geographic Variations in
Temperature Increases
(Avg. & Extremes), Rainfall,
and Ecosystem Shifts
Sea-Level Rise I
(Increased Frequency
of
Air Stagnation Periods
Elderly (> 65 yrs) Most at
Risk — Also Infants
Key: Acclimatization
Geographic Distribution &
Abundance of Vectors/Hosts
Depend on Temperature,
Moisture, Habitat effects
Near-term: Storm Surges,
Costal Flooding
Long Term: Inland Advance
of Saltwater Oceans/Seas
Increased Tropospheric
Air Pollution
(PM, 03, CO, etc)
HEAT-STRESS
MORTALITY
Lower
threshold
temperatures
in North than
in South
If only partial
acclimatization
then heat
deaths rise
COLD-INDUCED
MORTALITY
Higher
threshold
temperatures
in South
(>0°C) than
North (<0°C)
Prob. drop in
cold-related
mortality
7
\
TICK-BORNE
DISEASES
USA
EXAMPLES:
Lyme Disease
Rocky
Mountain
Spotted
Fever
MOSQUITO-
BORNE
DISEASES
USA
EXAMPLES:
Malaria
Dengue Fever
Arbovirus-
Related
Encephalitis
HUMAN
HEALTH
IMPACTS
Loss of Life
Nutrition
Vector-Borne
Diseases
Other
Communicable
Diseases
OTHER
TYPES OF
IMPACTS
Damaaeto:
Industries
Agriculture
Aquatic and
Land
Ecosystems
HUMAN
HEALTH
IMPACTS
Acute Pulmon.
Function
Decrements
Impaired Lung
Defenses
Increased
Respiratory
Disease
Susceptibility
OTHER
TYPES OF
IMPACTS
Forest and
Agriculture
Damage
Ecosystem
Effects
Materials
Damage
Figure 4-26. Bases for concern about global warming and climate change effects on the
environment and human health. Types of hypothesized likely human health
effects include (1) increases in mortality directly linked to temperature
extremes, (2) increases in incidence and spread of vector-borne infectious
diseases, (3) impacts secondary to projected sea-level rise, and (4) impacts
secondary to increased tropospheric air pollution. Additional impacts can be
expected because of shifting agricultural sustainability in various U.S. regions
consequent to extreme weather patterns leading to inland flooding or
droughts.
Source: Adapted from Grant (1989).
1 (5) Secure Food Supply. At the national level, the U.S. agriculture sector is likely to be able to
2 adapt to climate change. Overall, U.S. crop productivity is very likely to increase over the
3 next few decades, but the gains will not be uniform across the nation. Falling prices and
4 competitive pressures are very likely to stress some farmers, while benefiting consumers.
March 2001
4-142
DRAFT-DO NOT QUOTE OR CITE
-------
1 (6) Near-Term Increases in Forest Growth. Forest productivity is likely to increase over the
2 next several decades in some areas as trees respond to higher CO2 levels. Over the longer
3 term, changes in larger scale processes such as fire, insects, droughts, and disease will
4 possibly decrease forest productivity. Also, climate change is likely to cause long-term
5 shifts in forest species (e.g., distribution of sugar maple stands more northward, out of the
6 United States).
7 (7) Increased Damage in Coastal and Permafrost Areas. Climate change and the resulting rise
8 in sea level are likely to exacerbate threats to building, roads, powerlines, and other
9 infrastructure in climatically sensitive places, such as low-lying coastlines and the
10 permafrost regions of Alaska.
11 (8) Other Stresses Magnified by Climate Change. Climate change will very likely magnify the
12 cumulative impacts of other stresses, such as air and water pollution and habitat destruction
13 caused by human development patterns. For some systems, such as coral reefs, the
14 combined effects of climate change and other stresses are very likely to exceed a critical
15 threshold, bringing large, possibly irreversible impacts.
16 (9) Surprises Expected. It is likely that some aspects and impacts of climate change will be
17 totally unanticipated as complex systems respond to ongoing climate change in
18 unforeseeable ways.
19 (10) Uncertainties Remain. Significant uncertainties remain in the science underlying regional
20 climate changes and their impacts. Further research is needed to improve understanding
21 and predictive ability about societal and ecosystem impacts and to provide the public with
22 additional useful information about adaptation strategies.
23 The selected findings highlighted in Appendix 4D from the USGCRP (2000) report and
24 subsidiary regional reports illustrate well the considerable uncertainties and difficulties in
25 projecting likely climate change impacts on regional or local scales. The findings presented in
26 Appendix 4D also reflect well the mixed nature of projected potential climate change impacts
27 (combinations of mostly deleterious, but other possible beneficial effects) for U.S. regions and
28 their variation across the different regions. Difficulties in assessing regional-specific potential
29 impacts also can be illustrated by discussion below of determinants of the potential extent of
30 direct or indirect impacts of global warming on human health, as abstracted from various
31 published assessments cited above or alluded to in Appendices 4C, 4D, and 4E.
March 2001 4-143 DRAFT-DO NOT QUOTE OR CITE
-------
1 Modeling efforts and published analyses by Kalkstein and others have helped to identify
2 important factors that affect the magnitude of temperature-dependent mortality and provide
3 bases for projecting future temperature-related mortality trends (see Appendix 4E for recently
4 published projections for U.S. cities by Kalkstein and Greene, 1997). Examples of key
5 determinants of temperature-related mortality include (1) weather-sensitive mortality occurs
6 mainly as a function of extremes of temperature beyond certain threshold points (for increasing
7 or decreasing temperatures) that are characteristic of any particular city; (2) the extent of the
8 mortality is generally more dependent on the duration of the periods (days) during which
9 threshold points are exceeded than on maximum temperatures and also varies as a function of
10 combined relative humidity, temperature, and barometric pressure conditions that constitute
11 "oppressive" weather events that vary for different locales; and (3) the major population segment
12 typically most severely affected are the elderly (>65 years old).
13 Threshold temperature findings for summer and winter in U.S. cities suggest that weather
14 effects on mortality are relative (i.e., they vary in relation to the typical conditions to which local
15 residents have become acclimatized). Thus, the highest summer threshold temperatures for
16 mortality are found for the South and Southeast and the lowest in the Pacific and Northeast U.S.
17 regions. Conversely, lowest threshold temperatures for winter mortality are found for cities in
18 the coldest regions, whereas notably higher thresholds for cold-associated deaths occur for
19 warmer region cities, with threshold values for some being well above the freezing point. Also,
20 the total accumulated times of occurrence in the season of particular oppressive weather events
21 are important determinants of mortality levels (e.g., hot conditions early in the spring and
22 summer have a larger impact than similar conditions later in the summer, and length of a
23 heat-stress period also has a larger impact than maximum temperatures reached).
24 Acclimatization is a key determinant of weather-related mortality, and the greatest initial
25 increases in heat-related mortality might be expected in cities where temperatures are normally
26 cooler or in areas where global-warming-induced climate changes lead to increased frequency
27 and durations of high-temperature episodes. Eventual acclimatization may occur over the years,
28 however, when higher-than-usual temperatures become the new norm for cities in currently
29 cooler, more northern regions. As for cold-associated deaths, if average winter temperatures
30 were not to drop as low as usual in various regions, then winter mortality might generally
31 decrease because of fewer days falling below existing winter threshold levels for many cities.
March 2001 4-144 DRAFT-DO NOT QUOTE OR CITE
-------
1 But, if acclimatization occurs to higher average winter temperatures and wider variations in
2 temperature extremes occur in some areas because of global-warming-induced weather changes,
3 then those periods of lower maximum temperatures (especially of several days duration) could
4 cause even higher than past mortality rates previously observed with comparable winter
5 conditions. More sophisticated modeling also is needed to take into account combined effects of
6 temperature extremes and weather-related increases in air pollutants as possible mortality
7 determinants; for example, increased mortality or morbidity effects because of temperature
8 extremes may be exacerbated or added to by higher surface level atmospheric PM derived from
9 increased coal or oil combustion to generate more heat (in winter) or electricity (in summer for
10 air conditioning) during extreme temperature periods.
11 In addition to concern about possible mortality increases because of temperature extremes,
12 global warming, and consequent climate change also may impact human health through increases
13 in some infectious diseases. For many parts of the world, infectious diseases remain among the
14 leading causes of death, as occurred earlier in industrialized or "developed" countries (where
15 diseases such as influenza, pneumonia, and tuberculosis were among the leading causes of death
16 in 1900). Since then, the incidence and associated mortality for these and other infectious
17 diseases such as diphtheria, typhus, and polio have been reduced dramatically in industrialized
18 countries. In developed countries, it is not clear to what extent global-warming-induced climate
19 change may cause general increases in the incidence of such diseases, unless serious disruptions
20 of social structures occur or, in some coastal areas, breakdowns in sanitation systems happen as a
21 consequence of sea-level rise. The spread of infectious diseases is likely of greater concern for
22 many less developed countries, where inadequate medical care systems, immunization programs,
23 housing conditions, and nutrition make them more vulnerable to the spreading of such diseases.
24 Of particular shared concern for both developed and less developed countries with regard to
25 potential global warming impacts are infectious diseases spread by climate-dependent vectors.
26 Vector-borne diseases are those for which the infectious microbial agent is transmitted to humans
27 via another agent (the vector), such as the flea, tick, or mosquito. Well known examples of
28 vector-borne diseases are malaria (transmitted to humans via mosquitos) and bubonic plague
29 (transmitted via fleas or, at times, via animals directly to man as a respiratory disease). Climate
30 change can affect vector-borne diseases by various direct impacts on the infectious agent, the
31 vector, or intermediate hosts through variations in temperature, humidity, rainfall, or storm
March 2001 4-145 DRAFT-DO NOT QUOTE OR CITE
-------
1 patterns that alter (1) the multiplication rates of the infectious agent or the vector, (2) the biting
2 rate of the vector, (3) the geographic distribution of the intermediate animal hosts, or (4) the
3 amount of time that intermediate hosts or human hosts are exposed to the vector. Climate change
4 also can affect indirectly the rates or incidences of vector-borne diseases via impacts on
5 agricultural practices, ecosystem mixes (of grasses, trees, underbrush, etc.), surface water levels,
6 or other factors that determine intermediate host or vector distribution or survival. The variety of
7 vector-borne diseases is considerable, with some being of more concern than others for particular
8 countries, depending on specific climatic conditions and existing pools of infected hosts (both
9 human and intermediate animal hosts). Examples of vector-borne diseases illustrative of
10 concerns that apply to the United States for potential spread of vector-borne diseases include
11 Lyme disease, Rocky Mountain spotted fever, dengue fever, malaria, and viral encephalitis.
12 Lyme disease (initially recognized in Lyme, CT) is an inflammatory disease caused by a
13 spirochete, Borrelia burgdorferi, transmitted by several subspecies oflxodes racinus ticks.
14 Numerous species of birds and mammals can be hosts for various subspecies of the tick vector,
15 with varying geographic distributions. Lyme disease has four major U.S. foci, is spreading
16 rapidly, and has been found in Europe (Germany, Switzerland, France, and Austria). The U.S.
17 distribution of human cases of the disease tends to match areas where the tick vector is abundant,
18 and deer populations, along with factors such as temperature, humidity and local vegetation,
19 represent key determinants of tick abundance. The precise impact of global warming and climate
20 change on the distribution of Lyme disease is difficult to estimate. Lengthening of warm weather
21 periods and shortening of winter weather could enhance tick vector abundance and its potential
22 spread into adjoining areas if the weather changes (temperature, precipitation, etc.) were to favor
23 wider distribution of deer or other animal or bird hosts. Shifts of human populations into or out
24 of affected areas in response to changes in local climate also would help determine location-
25 specific alterations in Lyme disease rates.
26 Rocky Mountain spotted fever (initially identified in western mountain areas but actually
27 much more prevalent in southeastern U.S. states) is a highly fatal disease if not promptly
28 diagnosed and treated. Caused by the occobacillus, Rickettsiae rickettsii, the disease is spread by
29 ticks and is also known as tick fever, with analogous diseases occurring in many other countries.
30 The main North America vectors are the dog tick, D. variabilis, and the wood ticks, D. andersoni
31 and D. occidentalis, with varying geographic distributions. Geographic tick distributions parallel
March 2001 4-146 DRAFT-DO NOT QUOTE OR CITE
-------
1 closely the typical U.S. distribution of disease cases (highest incidence across the South).
2 Crucial for the spread of Rocky Mountain spotted fever is the wide variety of intermediate hosts
3 available to the ticks (i.e., many woodland mammals and birds) and temperature. Certain
4 optimum ranges of high temperatures (24 to 30 °C) likely speed the rickettsial growth in the
5 ticks, and ambient temperatures are important in determining tick breeding season length and
6 cycles, as well as their activity levels and biting rates. Each are enhanced by higher temperatures,
7 and the abundance of the vector is held in check, in part, by frequency and length of time that
8 winter temperatures drop well below freezing, thus killing overwintering adults. Lastly, relative
9 humidity conditions and rainfall are important as well, in that hot dry weather results in
10 desiccation of ticks and their eggs, reducing reproduction rates. Global warming-induced climate
11 change might increase the range of Rocky Mountain spotted fever tick vectors into more
12 northward U.S. areas and, possibly, into Canada, assuming the climate change includes sufficient
13 rainfall to sustain adequate habitats for host species and adequate moisture for survival of ticks
14 and eggs. Hot, dry periods caused by any prolonged drought conditions in the United States or
15 Canada predicted by some global warming scenarios, conversely, would not be conducive to
16 increased incidence of the disease in drought-affected areas.
17 Malaria, once widespread in the southern United States, remains endemic in many areas of
18 the world and is caused by four agents: (1) Plasmodium vivax, (2) P. malariae, (3) P. ovale, and
19 (4) P.falciparum. The agents cause clinical syndromes of varying severity, the most serious
20 being caused by P. falciparum, which can progress to death (>10% fatality in untreated children
21 and nonimmune adults). The other forms, although less severe, are still debilitating and are
22 typified by recurring episodes of fever, chills, and sweating. Malarial agents are transmitted from
23 infected humans, as the main host pool, by the bite of various subspecies of anopheles mosquitos.
24 Ambient temperatures of at least 15 to 18 °C are crucial for development of the malarial agents
25 within the mosquitos, and ambient temperature levels determine breeding season length and
26 survival rates (higher tropical temperatures being most favorable). Man's agricultural activities,
27 in providing irrigation ditches and more stagnant water habitats, has contributed to spread and
28 abundance of the anopheles mosquito in many areas of the world. Malaria is now rarely
29 endogenously transmitted in the United States, the pool of infected humans as hosts having been
30 reduced very substantially, owing to mosquito eradication programs. Prior to such programs, the
31 disease was endemic in widespread southern U.S. areas up to the 1940s, but, since then,
March 2001 4-147 DRAFT-DO NOT QUOTE OR CITE
-------
1 outbreaks mainly have occurred because of infected immigrants entering the country or U.S.
2 military veterans returning from overseas endemic areas. Global warming leading to higher
3 temperatures in more northerly U.S. areas and Europe could enhance conditions for the spread of
4 the disease. Both the range and abundance of competent vectors (various anopheles subspecies)
5 likely would be increased, especially if increased irrigation were required to support agriculture,
6 owing to higher temperatures. Also, higher temperatures in more northerly areas could extend
7 the range of adequate temperatures (>15 to 18 °C) needed for development of malarial agents in
8 the mosquitos. The remaining key factor in determining the likelihood of the spread of malaria,
9 however, is the infected host pool, with numbers of infected human hosts moving into or out of
10 areas of enhanced vulnerability being of crucial importance, as emphasized by Longstreth (1999).
11 Dengue fever, another mosquito-borne disease, is caused by four serotypes of a Group B
12 arbovirus. Fever, general muscle ache, severe headache, and retroorbital pain typify dengue fever
13 (usually not fatal); but it can progress to dengue haemorrhagic fever or dengue shock syndrome
14 (often fatal). Once endemic along the U.S. Gulf and South Atlantic coasts, dengue fever is now
15 rarely endogenously transmitted in the United States. The Aedis aegypti mosquito is the primary
16 vector, with wide southern U.S. distribution. The breeding season of the A. aegypti mosquito is
17 temperature-dependent, with breeding year-round in southern Florida, nearly year-round
18 elsewhere in Florida and along the Gulf Coast, and much shorter for successively more
19 northward bands of geographic distribution. Another potential vector, Aedes triseriatus, is
20 endogenous to states east of the Mississippi, and Aedes albopictus., a proven dengue vector
21 introduced from northern Asia, has been found in scattered U.S. sites. Higher temperatures are
22 also crucial for dengue transmission; transmission of dengue occurred experimentally only if
23 A. aegypti mosquitos were kept at 30 °C, and the incubation period for the virus to develop in the
24 mosquitos was shortened at 32 to 35 °C. Consistent with this, cases of dengue haemorrhagic
25 fever increased at non-U.S. sites when daily mean temperatures were 28 to 30 °C during hot
26 seasons, but decreased at the sites during cooler seasons with 25 to 28 °C temperatures.
27 Temperature increases in temperate ozone areas with A Aegypti or A. albopictus present would
28 tend to expand the range of these dengue fever vectors, including potential spread especially of
29 A. albopictus farther north in the United States and, perhaps, into Canada, in view of its
30 adaptation to cold weather as well. Whether or not increases in dengue will actually occur,
31 however, likely will depend on the distribution of rainfall and moisture content, the effects of
March 2001 4-148 DRAFT-DO NOT QUOTE OR CITE
-------
1 agricultural practices (e.g., increased irrigation), and movements of infected human hosts into or
2 out of areas with increased vector density.
3 Arbovirus-induced encephalitis syndromes vary in severity but include several that can be
4 highly fatal and are related to several other types of arbovirus-related syndromes (e.g., yellow
5 fever, dengue and other haemorrhagic fevers, hepatitis, arthritis, rashes, various tropical fevers).
6 Different types of mosquitos that serve as competent vectors for various types of
7 arbovirus-induced encephalitis of concern for the United States display different patterns of
8 distribution and differentially infect other hosts besides man (e.g., birds and large vertebrates
9 [horses, etc.] for some, birds and swine for another, and small woodland animals for others).
10 All have temperature-dependent components involved in development or transmission of the
11 viruses, but specific effects vary for different types. For example, the maximum temperatures
12 allowing the western equine encephalitis (WEE) vector to transmit the virus effectively are below
13 25 °C, and this allows for earlier spread of the disease in warm periods and the possible more
14 northern spread of the disease. In contrast, St. Louis encephalitis (SLE) arbovirus development
15 and transmission are markedly enhanced by temperatures exceeding 25 °C. Rainfall and
16 moisture patterns are also important, with most vectors (e.g., Cx tarsalis) benefitting from higher
17 rainfall; but at least one (Cxpipiens) is enhanced by less rainfall, with outbreaks of its
18 encephalitis syndrome being more common during high-temperature drought periods. Thus,
19 effects of global warming and climate change on the incidence and spread of arbovirus-related
20 encephalitis syndromes are difficult to predict. However, it generally appears that higher
21 temperatures should enhance the abundance and wider U.S. geographic distribution of most of
22 the competent mosquito vectors. All of this again assumes that higher temperatures and rainfall
23 patterns will be such to allow adequate habitats for other hosts besides humans in the potential
24 new range areas. Lastly, as noted before for the other infectious diseases discussed, the
25 movement of populations into or out of the affected areas also will be important in determining
26 any location-specific increased (or decreased) incidence of arbovirus-related encephalitises.
27 Of special concern would be the introduction of any new arboviruses not now currently endemic
28 in the United States (e.g., Japanese B encephalitis [VBE], not currently found in the United
29 States but closely related to SLE in terms of involving Culex mosquitos and birds, with several
30 Culex subspecies in the United States found to be effective vectors for the virus).
March 2001 4-149 DRAFT-DO NOT QUOTE OR CITE
-------
1 The above discussion of potential effects of global warming and climate change on the
2 incidence and spread of infectious diseases is further complicated by considerations of possible
3 impacts of expected sea-level rise in response to global warming. Some low-lying coastal areas
4 now serving as excellent habitats for certain mosquitos, for example, might come to be inundated
5 by seawater and no longer be available breeding areas. Or, increased storm surges or expansion
6 of marshy areas reaching farther inland might contribute to creation of conditions in some areas
7 more favourable to enhance mosquito breeding. Also of concern is the potential for disruption of
8 sanitation systems. The spread of infectious diseases, besides the vector-borne types discussed
9 above, could be increased because of flooding of coastal cities secondary to heavy precipitation
10 events (e.g., hurricanes). Inundation of sewage treatment facilities and sewage lines might not
11 only result in immediate spread of disease-containing fecal or other material, but damage to such
12 sanitation system components could result in longer term disruption of waste-removal
13 capabilities and the spread of disease.
14 Lastly, another concern with climate-induced heavy precipitation events or sea-level rise is
15 the potential for flooding of inland or coastal waste disposal sites. This could result in increased
16 spread of waterborne infectious diseases, depending on the specific materials present in such
17 dumps and the extent of their dispersal caused by flooding. The flooding of dump sites
18 containing hazardous chemical wastes represents yet another potential concern associated with
19 sea-level rise. The spread of various toxic chemicals from such waste disposal sites could carry
20 with it increased threats of many types of possible health effects, as well as potential
21 environmental effects (natural vegetation and ecosystem damage, contamination of crop lands by
22 toxic chemicals, etc.).
23 Difficulties in projecting region-specific climate change impacts are complicated further by
24 the need to evaluate potential effects of local- or regional-scale changes in key air pollutants not
25 only on global scale temperature trends but also in terms of potentially more local- or regional -
26 scale impacts on temperature and precipitation patterns. Of much importance for this are varying
27 roles played by atmospheric particles.
28
29 4.5.2.2 Airborne Particle Relationships to Global Warming and Climate Change
30 Atmospheric particles both scatter and absorb incoming solar radiation at visible light
31 wavelengths. The scattering of solar radiation back to space leads to a decrease in transmission
March 2001 4-150 DRAFT-DO NOT QUOTE OR CITE
-------
1 of visible radiation to the Earth's surface and, hence, to a decrease in the heating rate of the
2 surface and the atmosphere. The absorption of either incoming solar radiation or outgoing
3 terrestrial infrared radiation by atmospheric particles results in heating of the lower atmosphere.
4 Interactions of atmospheric particles with electromagnetic radiation from the visible through the
5 infrared spectral regions are responsible for their direct effects on climate, which are the result of
6 the same physical processes responsible for visibility degradation. Visibility reduction is caused
7 by particle scattering in all directions, whereas climate effects result mainly from scattering in the
8 upward direction. The net effect of the above processes can be expressed as a radiative forcing,
9 which is the change in the average net radiation at the top of the troposphere because of a change
10 in solar (shortwave, or visible) or terrestrial (longwave, or infrared) radiation (Houghton et al.,
11 1990). The radiative forcing drives the climate to respond, but because of uncertainties in a
12 number of feedback mechanisms involving climate response, radiative forcing is used as a first-
13 order estimate of the potential importance of various substances. Sulfate particles scatter solar
14 radiation effectively and do not absorb at visible wavelengths, whereas they absorb weakly at
15 infrared wavelengths (IPCC, 1995). Nitrate particles exhibit grossly similar properties. The
16 effects of mineral dust particles are complex; they weakly absorb solar radiation but their overall
17 effect on solar radiation depends on particle size and the reflectivity of the underlying surface.
18 They absorb infrared radiation and thus contribute to greenhouse warming (Tegen et al., 1996).
19 Organic carbon particles mainly reflect solar radiation, whereas elemental carbon and other black
20 carbon particles (e.g., PAHs with H:C ratios of <0.3) are strong absorbers of solar radiation
21 (IPCC, 1995). However, the optical properties of carbonaceous particles are modified if they
22 become coated with water or sulfuric acid. Particles containing black carbon also can exert a
23 direct effect after deposition onto surfaces that are more reflective (e.g., snow and ice). In this
24 case, additional solar radiation is absorbed by the surface; conversely, more reflective particles
25 deposited on a dark surface result in additional solar radiation being reflected back to space.
26 Anthropogenic (Twomey, 1974; Twomey, 1977) and biogenic (Charlson et al., 1987)
27 sulfate particles also exert indirect effects on climate by serving as cloud condensation nuclei,
28 which results in changes in the size distribution of cloud droplets by producing more particles
29 with smaller sizes. The same mass of liquid water in smaller particles leads to an increase in
30 amount of solar radiation that clouds reflect back to space because the total surface area of the
31 cloud droplets is increased. This has been supported by satellite observations indicating that the
March 2001 4-151 DRAFT-DO NOT QUOTE OR CITE
-------
1 effective radius of cloud droplets is smaller in the Northern Hemisphere than in the Southern
2 Hemisphere (Han et al., 1994). Smaller cloud droplets also have a lower probability of
3 precipitating and, thus, have a longer lifetime than larger ones. Although the effects of sulfate
4 have been considered most widely, interactions with other aerosol components also may be
5 important. Novakov and Penner (1993) have provided evidence that carbonaceous particles can
6 modify the nucleation properties of sulfate particles.
7 The amount of solar radiation incident on the earth-atmosphere system, or the solar
8 constant, is 1370 W m"2, or 342.5 W m"2 on a globally averaged basis (calculated by dividing the
9 solar constant by 4). The addition of sulfate and organic carbon as airborne PM results in
10 enhanced scattering and net cooling, whereas the addition of particles containing elemental
11 carbon results in absorption of solar and terrestrial radiation and net heating. The estimated
12 raditive forcing because of the scattering of solar radiation back to space caused mainly by
13 sulfate particles is -0.4 W m"2 (IPCC, 1995), with an uncertainty range of a factor of two. The
14 uncertainty range reflects uncertainties in the emissions of SO2, the amount of SO2 that is
15 oxidized to sulfate, the atmospheric lifetime of sulfate, and the optical properties of the sulfate
16 particles. These values may be compared to the radiative forcing exerted by greenhouse gases of
17 about + 2.4 W m"2, with an uncertainty factor of 1.15 from the preindustrial era (ca. 1800) to
18 1994. Since the latter part of the 19th century, the mean surface temperature of the earth has
19 increased from 0.3 to 0.6 °C according to the IPCC (1995) assessment. Estimates of the indirect
20 effects of particles range from 0 to -1.5 W m"2 (IPCC, 1995). Because of a lack of quantitative
21 knowledge, no central value could be given. Therefore, on a globally averaged basis, the direct
22 and indirect effects of anthropogenic sulfate particles likely have offset partially the warming
23 effects caused by increases in levels of greenhouse gases (Charlson et al., 1992).
24 Much of the work investigating the effects of particles on climate has focused on sulfate
25 particles. However, particles containing elemental carbon (EC) from fossil fuel combustion and
26 biomass burning or mineral dust may exert radiative forcing, with spatial distributions very
27 different than for sulfate. Tegen et al. (1996) and Tegen and Lacis (1996) used a global scale
28 three-dimensional model to evaluate the radiative forcing caused by mineral dust particles.
29 Tegen and Lacis (1996) found that the sign and the magnitude of the radiative forcing depends on
30 the height distribution of the dust and the effective radius of the particles. In particular, for a dust
31 layer extending from 0 km to 3 km, positive radiative forcing at visible wavelengths is found for
March 2001 4-152 DRAFT-DO NOT QUOTE OR CITE
-------
1 particle radii greater than 1.8 //m, whereas negative forcing is found for smaller particles. They
2 calculated a global mean radiative forcing caused by mineral dust from all sources of 0.14 W m"2
3 and from mineral dust from lands disturbed by human activity of 0.09 W m"2. This value
4 represents a near cancellation between a much larger solar forcing of -0.25 W m"2 and a thermal
5 forcing of 0.34 W m"2. Uncertainty factors could not be estimated for these calculations because
6 they were judged to be largely unknown. Haywood and Shine (1995) estimated a global mean
7 radiative forcing of 0.1 W m"2, with an uncertainty factor >3, caused by the absorption of solar
8 radiation by EC released by fossil fuel combustion. The IPCC (1995) estimated a global mean
9 radiative forcing of -0.1 W m"2 caused by particles produced by biomass burning, with an
10 uncertainty factor of three. The global mean radiative forcing exerted by particles would then be
11 -0.5 W m"2, with an uncertainty of about a factor of 2.4. Figure 4-27 summarizes estimates of
12 global mean radiative forcing exerted by greenhouse gases and various types of particles.
13 Deviations from the global mean values can be very large on the regional scale. For
14 instance, Tegen et al. (1996) found that local radiative forcing exerted by dust raised from
15 disturbed lands ranges from -2.1 W m"2 to 5.5 W m"2 over desert areas and their adjacent seas.
16 The largest regional values of radiative forcing caused by anthropogenic sulfate are about
17 -3 W m"2 in the eastern United States, south central Europe, and eastern China (Kiehl and
18 Briegleb, 1993). These regional maxima in aerosol forcing are at least a factor of 10 greater than
19 their global mean values shown in Figure 4-27. By comparison, regional maxima in forcing by
20 the well-mixed greenhouse gases are only about 50% greater than their global mean value (Kiehl
21 and Briegleb, 1993). Thus, the estimates of local radiative forcing by particles also are large
22 enough to completely cancel the effects of greenhouse gases in many regions and to cause a
23 number of changes in the dynamic structure of the atmosphere that still need to be evaluated.
24 A number of anthropogenic pollutants whose distributions are highly variable are also effective
25 greenhouse absorbers. These gases include O3 and, possibly, HNO3, C2H4, NH3, and SO2, all of
26 which are not commonly considered in radiative forcing calculations (Wang et al. 1976). High
27 ozone values are found downwind of urban areas and areas where there is biomass burning.
28 However, Van Borland et al. (1997) found that there may not be much cancellation between the
29 radiative effects for ozone and for sulfate, because both species have different seasonal cycles
30 and show significant differences in their spatial distribution.
March 2001 4-153 DRAFT-DO NOT QUOTE OR CITE
-------
3-
IE
> 2-
^^
O)
'o
CD
CO
^
c
CD
CD
i
CD -1 -
.Q
O
CD
N^'
2
— Halocarbons
— N20
— CH4
~~ C°2 Sulfate
i-. /
/
r
I Tropospheric -p
Ozone
^\
\
4
Direct
Effect
Fossil Solar
Fuel Biomass Mineral Variability
Soot Burning Qust -p
T i I
^^^^^n ^^^^^^ 1
T
Indirect
Effect
Confidence level
High Low Low Low
Very Very Very Very Very
low low low low low
Figure 4-27. Estimated global mean radiative forcing exerted by gas and various particle
phase species from 1850 to 1950.
Source: Adapted from IPCC (1995) and Tegen and Lacis (1996).
1 Observational evidence for the climatic effects of particles is sparse. Haywood et al. (1999)
2 found that the inclusion of anthropogenic aerosols results in a significant improvement between
3 calculations of reflected sunlight at the top of the atmosphere and satellite observations in
4 oceanic regions close to sources of anthropogenic PM.
5 Uncertainties in calculating the direct effect of airborne particles arise from a lack of
6 knowledge of their vertical and horizontal variability, their size distribution, chemical
7 composition and the distribution of components within individual particles. For instance,
8 gas-phase sulfur species may be oxidized to form a layer of sulfate around existing particles in
9 continental environments, or they may be incorporated in sea-salt particles (e.g., Li-Jones and
10 Prospero, 1998). In either case, the radiative effects of a given mass of the sulfate will be much
March 2001
4-154
DRAFT-DO NOT QUOTE OR CITE
-------
1 lower than if pure sulfate particles were formed. It also must be stressed that the overall radiative
2 effect of particles at a given location is not simply determined by the sum of effects caused by
3 individual classes of particles because of interactions between particles with different radiative
4 characteristics and with gases.
5 Calculations of the indirect effects of particles on climate are subject to much larger
6 uncertainties than are calculations of their direct effects, reflecting uncertainties in a large
7 number of chemical and microphysical processes in describing the effects of sulfate on the size
8 distribution and number of droplets within a cloud. A complete assessment of the radiative
9 effects of PM will require supercomputer calculations that incorporate the spatial and temporal
10 behavior of particles of varying composition that have been emitted or formed from precursors
11 emitted from different sources. Refining values of model input parameters (such as improving
12 emissions estimates) may be as important as improving the models per se in calculations of direct
13 radiative forcing (Pan et al., 1997) and indirect radiative forcing (Pan et al., 1998) caused by
14 sulfate. However, uncertainties associated with the calculation of radiative effects of particles
15 likely will remain much larger than those associated with well-mixed greenhouse gases.
16 This means that, although on a global scale atmospheric particles likely exert an overall net
17 effect of slowing global warming, much uncertainty would apply to any modeling efforts aimed
18 at projecting net effects on global warming processes, resulting climate change, and any
19 consequent human health or environmental effects because of location-specific increases or
20 decreases in anthropogenic emissions of atmospheric particles or their precursors. For example,
21 any net impacts of regional sulfates in reducing global-climate-change-induced increases in local
22 temperatures may well be offset partially by local surface level heating because of carbonaceous
23 particles from diesel emissions or coal combustion energy generation being deposited on snow or
24 ice covered surfaces or contributing to more rapid evaporation or rainout of water from overhead
25 clouds.
26
27
28
29
March 2001 4-155 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.6 SUMMARY
2 4.6.1 Particulate Matter Effects on Vegetation and Ecosystems
3 Human existence on this planet depends on ecosystems and the services and products they
4 provide. Both ecosystem structure and function play an essential role in providing societal
5 benefits. Society derives two types of benefits from the structural aspects of an ecosystem:
6 (1) products with market value such as fish, minerals, forage, forest products, biomass fuels,
7 natural fiber, and many pharmaceuticals, and the genetic resources of valuable species (e.g.,
8 plants for crops and timber and animals for domestication); and (2) the use and appreciation of
9 ecosystem for recreation, aesthetic enjoyment, and study.
10 Ecosystem functions that maintain clean water, pure air, a green earth, and a balance of
11 creatures, are functions that enable humans to survive. They are the dynamics of ecosystems.
12 The benefits they impart include absorption and breakdown of pollutants, cycling of nutrients,
13 binding of soil, degradation of organic waste, maintenance of a balance of gases in the air,
14 regulation of radiation balance, climate, and the fixation of solar energy. Concern has risen in
15 recent years concerning the integrity of ecosystems because there are few ecosystems on the
16 Earth today that are not influenced by humans. For this reason, the deposition of PM and its
17 impact on vegetation and ecosystems is of great importance.
18 The PM whose effects on vegetation and ecosystems are considered in this chapter is not a
19 single pollutant but represents a heterogeneous mixture of particles differing in origin, size, and
20 chemical constituents. The effects of exposure to a given mass concentration of PM of particular
21 size (measured as PM10; PM25, etc.) may, depending on the particular mix of deposited particles,
22 lead to widely differing phytotoxic responses. This has not been characterized adequately.
23 Atmospheric deposition of particles to ecosystems takes place via both wet and dry
24 processes through the three major routes indicated below.
25 (1) Precipitation scavenging, in which particles are deposited in rain and snow
26 (2) Fog, cloud water, and mist interception
27 (3) Dry deposition, a much slower, yet more continuous removal to surfaces
28 Deposition of heavy metal particles to ecosystems occurs by wet and dry processes. Dry
29 deposition is considered more effective for coarse particles of natural origin and elements such as
30 iron and manganese, whereas wet deposition generally is more effective for fine particles of
March 2001 4-156 DRAFT-DO NOT QUOTE OR CITE
-------
1 atmospheric origin and elements such as cadmium, chromium, lead, nickel, and vanadium. The
2 actual importance of wet versus dry deposition, however, is highly variable, depending on the
3 type of ecosystem, location, and elevation.
4 Deposition of PM on above-ground plant parts can have either a physical and or chemical
5 impact, or both. Particles transferred from the atmosphere to plant surfaces may cause direct
6 effects if they (1) reside on the leaf, twig, or bark surface for an extended period; (2) be taken up
7 through the leaf surface; or (3) are removed from the plant via resuspension to the atmosphere,
8 washing by rainfall, or litter-fall with subsequent transfer to the soil.
9 Chemical effects include excessive alkalinity or acidity. The effects of "inert" PM are
10 mainly physical, whereas the effects of toxic particles are both chemical and physical. The
11 effects of dust deposited on plant surfaces or on soil are more likely to be associated with their
12 chemistry than with the mass of deposited particles and are usually of more importance than any
13 physical effects. The majority of the easily identifiable direct and indirect effects, other than
14 climate-change impacts, occur in severely polluted areas around heavily industrialized point
15 sources such as limestone quarries; cement kilns; and iron; lead, and various smelting factories.
16 Studies of the direct effects of chemical additions to foliage in particulate deposition have found
17 little or no effects of PM on foliar processes; however, both conifers and deciduous species have
18 shown significant effects on leaf surface structures after exposure to simulated acid rain or mist
19 at pH 3.5. Many experimental studies indicate that epicuticular waxes (which function to prevent
20 water loss from plant leaves) can be destroyed by acid rain in a few weeks. This function is
21 particularly crucial in conifers because of the longevity of evergreen foliage.
22 Though there has been no direct evidence of a physiological association between tree injury
23 and exposure to metals, heavy metals have been implicated because their deposition pattern is
24 correlated with forest decline. The role of heavy metals has been indicated by phytochelatin
25 measurements. Phytochelatins are intracellular metal-binding peptides that act as indicator of
26 metal stress. Because they are produced by plants as a response to sublethal concentrations of
27 heavy metals, they can be used to indicate that heavy metals are involved in forest decline.
28 Concentrations of the phytochelatins increased with altitude, as did forest decline, and they also
29 increased across regions showing increased levels of forest injury.
30 Secondary organics formed in the atmosphere have been referred to under the following
31 terms: toxic substances, pesticides, hazardous air pollutants (HAPS), air toxics, semivolatile
March 2001 4-157 DRAFT-DO NOT QUOTE OR CITE
-------
1 organic compounds (SOCs), and persistent organic pollutants (POPS). The chemical substances
2 listed under the above headings are not criteria pollutants controlled by NAAQS as cited under
3 CAA Sections 108 and 109 (U.S. Code, 1991), but rather are controlled under CAA Sect. 112,
4 Hazardous Air Pollutants. Their possible effects in the environment on humans and ecosystems
5 are discussed in many other government documents and publications. They are mentioned in this
6 chapter because, in the atmosphere many of the chemical compounds are partitioned between gas
7 and particle phases and are deposited as particulate matter. As particles, they become airborne
8 and can be distributed over a wide area and impact remote ecosystems. Some of the chemical
9 compounds are of concern to humans because they may reach toxic levels in food chains of both
10 animals and humans, whereas others tend to decrease or maintain the same toxicity as they move
11 through the food chain.
12 An important characteristic of fine particles is their ability to affect the flux of solar
13 radiation passing through the atmosphere directly, by scattering and absorbing solar radiation,
14 and indirectly, by acting as cloud condensation nuclei that, in turn, influence the optical
15 properties of clouds. Regional haze has been estimated to diminish surface solar visible radiation
16 by approximately 8%. Crop yields have been reported as being sensitive to the amount of
17 sunlight received, and crop losses have been attributed to increased airborne particle levels in
18 some areas of the world.
19 The transmission of solar UV-B radiation through the Earth's atmosphere is controlled by
20 ozone, clouds, and particles. The depletion of stratospheric ozone caused by the release of
21 chlorofluorcarbons and other ozone-depleting substances has resulted in heightened concern
22 regarding potentially serious increases in the amount of solar UV-B (SUVB) radiation reaching
23 the Earth's surface. Plant species vary enormously in their response to UV-B exposures, and
24 large differences in response also occur among different genotypes within a species. In general,
25 dicotyledonous plants are more sensitive than monocotyledons from similar environments.
26 In addition, plant responses may differ depending on stage of development. Because plants
27 evolved under the selective pressure of ambient UV-B radiation in sunlight, they have developed
28 adaptive mechanisms. Although inhibition of photosynthesis is a detrimental growth effect,
29 flavonoid synthesis represents acclimation. Plants growing under full light have been shown to
30 be protected against UV-B effects but not when growing under weak visible light. A common
March 2001 4-158 DRAFT-DO NOT QUOTE OR CITE
-------
1 adaptation is alteration in leaf transmission properties, which results in attenuation of UV-B in
2 the epidermis before it can reach the leaf interior.
3 Indirect effects of PM on plants are usually the most significant because they can alter
4 nutrient cycling in ecosystems and inhibit plant uptake of nutrients and, therefore, have a great
5 impact on ecosystem biodiversity. Indirect effects occur through the soil and result from the
6 deposition of heavy metals, nitrates, sulfates, or acidic precipitation and their impact on the soil
7 microbial community. The soil environment is one of the most dynamic sites of biological
8 interaction in nature. Bacteria in the soil are essential components of the nitrogen and sulfur
9 cycles that make these elements available for plant uptake. Fungi form mycorrhizae,
10 a mutualistic symbiotic relationship, that is integral in mediating plant uptake of mineral
11 nutrients. Changes in the soil environment that influence the role of the bacteria and fungi in
12 nutrient cycling and availability determine plant and ecosystem response.
13 Major impacts of PM on soil environments occur through deposition of nitrates and sulfates
14 and the acidifying effect of the FT ion associated with these compounds in wet and dry
15 deposition. Although the soils of most of North American forest ecosystems are nitrogen
16 limited, there are some forests that exhibit severe symptoms of nitrogen saturation. They include
17 the high-elevation, spruce-fir ecosystems in the Appalachian Mountains; the eastern hardwood
18 watersheds at the Fernow Experimental Forest near Parsons, WV; the mixed conifer forest and
19 chaparral watershed with high smog exposure in the Los Angeles Air Basin; the high-elevation
20 alpine watersheds in the Colorado Front Range; and a deciduous forest in Ontario, Canada.
21 Nitrogen saturation results when additions to soil background nitrogen (nitrogen loading)
22 exceed the capacity of plants and soil microorganisms to utilize and retain nitrogen. An
23 ecosystem no longer functions as a sink under these circumstances. Possible ecosystem
24 responses to nitrate saturation, as postulated by Aber and his coworkers, include (1) a permanent
25 increase in foliar nitrogen and reduced foliar phosphorus and lignin because of the lower
26 availability of carbon, phosphorus, and water; (2) reduced productivity in conifer stands caused
27 by disruptions of physiological function; (3) decreased root biomass and increased nitrification
28 and nitrate leaching; (4) reduced soil fertility, the results of increased cation leaching, increased
29 nitrate and aluminum concentrations in streams, and decreased water quality. Saturation implies
30 that some resource other than nitrogen is limiting biotic function. Water and phosphorus for
31 plants and carbon for microorganisms are the resources most likely to be the secondary limiting
March 2001 4-159 DRAFT-DO NOT QUOTE OR CITE
-------
1 factors. The appearance of nitrogen in soil solution is an early symptom of excess nitrogen. In
2 the final stage, disruption of forest structure becomes visible.
3 Changes in nitrogen supply can have a considerable impact on an ecosystem's nutrient
4 balance. Increases in soil nitrogen play a selective role. Plant succession patterns and
5 biodiversity are affected significantly by chronic nitrogen additions in some ecosystems.
6 Long-term nitrogen fertilization studies in both New England and Europe suggest that some
7 forests receiving chronic inputs of nitrogen may decline in productivity and experience greater
8 mortality. For example, long-term fertilization experiments at Mount Ascutney, VT, suggest that
9 declining coniferous forest stands with slow nitrogen cycling may be replaced by deciduous
10 fast-growing forests that cycle nitrogen rapidly. Excess nitrogen inputs to unmanaged heathlands
11 in the Netherlands also have been found to result in nitrophilous grass species replacing slower
12 growing heath species. Over the past several decades, the composition of plants in the forest
13 herb layers had been shifting toward species commonly found on nitrogen-rich areas. It also was
14 observed that the fruiting bodies of mycorrhizal fungi had decreased in number.
15 Notable impacts of excess nitrogen deposition also have been observed with regard to
16 aquatic systems. For example, atmospheric nitrogen deposition into soils in watershed areas
17 feeding into estuarine sound complexes (e.g., the Pamlico Sound of North Carolina) appear to
18 contribute to excess nitrogen flows in runoff (especially during and after heavy rainfall events
19 such as hurricanes). Together with excess nitrogen runoff from agricultural practices or other
20 uses (e.g., fertilization of lawns or gardens), massive influxes of such nitrogen into watersheds
21 and sounds can lead to dramatic decreases in water oxygen and increases in algae blooms that
22 can cause extensive fish kills and damage to commercial fish and sea food harvesting.
23 Acidic deposition has played a major role in soil acidification in some areas of Sweden,
24 elsewhere in Europe, and in eastern North America. Soil acidification and its effects result from
25 deposition of nitrates, sulfates, and associated FT ion. A major concern is that soil acidity will
26 lead to nutrient deficiency. Growth of tree species can be affected when high aluminum-to-
27 nutrient ratios limit uptake of calcium and magnesium and create a nutrient deficiency. Calcium
28 is essential in the formation of wood and the maintenance of cells (the primary plant tissues
29 necessary for tree growth), and it must be dissolved in soil water to be taken up by plants. Acidic
30 deposition can increase aluminum concentrations in soil water by lowering the pH in aluminum -
31 rich soils through dissolution and ion-exchange processes. Aluminum in soil can then be taken
March 2001 4-160 DRAFT-DO NOT QUOTE OR CITE
-------
1 up by roots more readily than calcium because of its greater affinity for negatively charged
2 surfaces. Tree species can be adversely affected if altered Ca/Al ratios impair Ca or Mg uptake.
3 Overall, then, PM produced by human activities has the potential to cause the loss of
4 ecosystem biodiversity in ways that reduces the ability of ecosystems to provide the services that
5 society requires to sustain life. The major impacts of PM on ecosystems are the indirect effects
6 that occur through the soil and affect plant growth, vigor, and reproduction. Mineral nutrient
7 cycling can be altered by the deposition of heavy metals. The deposition of nitrogen and sulfur
8 and the acidifying effects of the two in association with the H+ ion in precipitation also alter
9 biogeochemical cycling, cause soil acidification, alter the Ca/Al ratio, and impact the growth of
10 vegetation and forest trees, in particular. Leaching of nitrates and other minerals through runoff
11 can impact coastal and aquatic wetlands and, thus, influence their ability to produce the products
12 and services necessary for human society.
13
14 4.6.2 Particulate Matter-Related Effects on Visibility
15 Visibility is defined as the degree to which the atmosphere is transparent to visible light and
16 the clarity and color fidelity of the atmosphere. Visual range is the farthest distance a black
17 object can be distinquished against the horizontal sky. Visibility impairment is any humanly
18 perceptible change in visibility. For regulatory purposes, visibility impairment, characterized by
19 light extinction, visual range, contrast, and coloration, is classified into two principal forms:
20 (1) "reasonably attributable" impairment, attributable to a single source or small group of
21 sources, and (2) regional haze, any perceivable change in visibility caused by a combination of
22 many sources over a wide geographical area.
23 Visibility is measured by human observation, light scattering by particles, the light
24 extinction-coefficient and parameters related to the light-extinction coefficient (visual range and
25 deciview scale), the light scattering coefficient, and fine PM concentrations. The air quality
26 within a sight path will affect the illumination of the sight path by scattering or absorbing solar
27 radiation before it reaches the Earth's surface. The rate of energy loss with distance from a beam
28 of light is the light extinction coefficient. The light extinction coefficient is the sum of the
29 coefficients for light absorption by gases (oag), light scattering by gases (osg), light absorption by
30 particles (oap), and light scattering by particles (osp). Atmospheric particles are frequently divided
31 into fine and coarse particles. Corresponding coefficients for light scattering and absorption by
March 2001 4-161 DRAFT-DO NOT QUOTE OR CITE
-------
1 fine and coarse particles are osfp and oafp and oscp and oacp, respectively. Visibility within a sight
2 path longer than approximately 100 km (60 mi) is affected by change in the optical properties of
3 the atmosphere over the length of the sight path.
4 Visibility impairment is associated with airborne particle properties, including size
5 distributions (i.e., fine particles in the 0.1- to 1.0-//m size range) and aerosol chemical
6 composition, and with relative humidity. With increasing relative humidity, the amount of
7 moisture available for absorption by particles increases, thus causing the particles to increase in
8 both size and volume. As the particles increase in size and volume, the light scattering potential
9 of the particles also generally increases. Visibility impairment is greatest in the eastern United
10 States and Southern California. In the eastern United States, visibility impairment is caused
11 primarily by light scattering by sulfate aerosols and, to a lesser extent, by nitrate particles and
12 organic aerosols, carbon soot, and crustal dust. Haziness in the southeastern United States,
13 caused by increased atmospheric sulfate, has increased by ca. 80% since the 1950s and is greatest
14 in the summer months, followed by the spring and fall, and winter. Light scattering by nitrate
15 aerosols is the major cause of visibility impairment in Southern California. Nitrates contribute
16 about 40% to the total light extinction in Southern California and accounts for 10 to 20% of the
17 total extinction in other U.S. areas.
18 Organic particles are the second largest contributors to light extinction in most U.S. areas.
19 Organic carbon is the greatest cause of light extinction in the Pacific Northwest, Oregon, Idaho,
20 and Montana, accounting for 40 to 45% of the total extinction. Also, organic carbon contributes
21 between 15 to 20% to the total extinction in most of the western United States and 20 to 30% in
22 the remaining U.S. areas.
23 Coarse mass and soil, primarily considered "natural extinction", is responsible for some of
24 the visibility impairment in northern California and Nevada, Oregon, southern Idaho, and
25 western Wyoming. Dust transported from Southern California and the subtropics has been
26 associated with regional haze in the Grand Canyon and other southwestern U.S. class I areas.
27
28 4.6.3 Particulate Matter-Related Effects on Materials
29 Building materials (metals, stones, cements, and paints) undergo natural weathering
30 processes from exposure to environmental elements (wind, moisture, temperature fluctuations,
31 sun light, etc.). Metals form a protective film that protects against environmentally induced
March 2001 4-162 DRAFT-DO NOT QUOTE OR CITE
-------
1 corrosion. The natural process of metal corrosion from exposure to natural environmental
2 elements is enhanced by exposure to anthropogenic pollutants, in particular SO2, rendering the
3 protective film less effective.
4 Dry deposition of SO2 enhances the effects of environmental elements on calcereous stones
5 (limestone, marble, and cement) by converting calcium carbonate (calcite) to calcium sulfate
6 dihydrate (gypsum). The rate of deterioration is determined by the SO2 concentration, the stone's
7 permeability and moisture content, and the deposition rate; however, the extent of the damage to
8 stones produced by the pollutant species apart from the natural weathering processes is uncertain.
9 Sulfur dioxide also has been found to limit the life expectancy of paints by causing discoloration
10 and loss of gloss and thickness of the paint film layer.
11 A significant detrimental effect of particle pollution is the soiling of painted surfaces and
12 other building materials. Soiling changes the reflectance of a material from opaque and reduces
13 the transmission of light through transparent materials. Soiling is a degradation process that
14 requires remediation by cleaning or washing, and, depending on the soiled surface, repainting.
15 Available data on pollution exposure indicates that particles can result in increased cleaning
16 frequency of the exposed surface and may reduce the life usefulness of the material soiled.
17 Attempts have been made to quantify the pollutants exposure levels at which materials damage
18 and soiling have been perceived. However, to date, insufficient data are available to advance our
19 knowledge regarding perception thresholds with respect to pollutant concentration, particle size,
20 and chemical composition.
21
22 4.6.4 Effects of Particulate Matter on the Transmission of Solar Ultraviolet
23 Radiation and Global Warming Processes
24 Extensive potential future impacts on human health and the environment are projected to
25 occur because of increased transmission of solar ultraviolet radiation (UV-B) through the Earth's
26 atmosphere, secondary to stratospheric ozone depletion resulting from anthropogenic emissions
27 of chlorofluorcarbons (CFCs), halons, and certain other gases. However, the estimation of the
28 likely future extent of detrimental effects caused by increased penetration of solar UV-B to the
29 Earth's surface is complicated by atmospheric particle effects, which vary depending on size and
30 composition of particles that can differ substantially over different geographic areas and from
31 season to season over the same area. Also, atmospheric particles greatly complicate projections
March 2001 4-163 DRAFT-DO NOT QUOTE OR CITE
-------
1 of future trends in global warming processes because of emissions of greenhouse gases;
2 consequent increases in global mean temperature, and resulting changes in regional and local
3 weather patterns; and mainly deleterious (but some beneficial) location-specific human health
4 and environmental impacts.
5 The physical processes (i.e., scattering and absorption) responsible for airborne particle
6 effects on transmission of solar ultraviolet and visible radiation are the same as those responsible
7 for visibility degradation. Scattering of solar radiation back to space and absorption of solar
8 radiation determine the effects of an aerosol layer on solar radiation. The transmission of solar
9 UV-B radiation is affected strongly by atmospheric particles. Measured attenuations of UV-B
10 under hazy conditions range up to 37% of the incoming solar radiation. Measurements relating
11 variations in PM mass directly to UV-B transmission are lacking. Particles also can affect the
12 rates of photochemical reactions occurring in the atmosphere. Depending on the amount of
13 absorbing substances in the particles, photolysis rates either can be increased or decreased.
14 In addition to direct climate effects through the scattering and absorption of solar radiation,
15 particles also exert indirect effects on climate by serving as cloud condensation nuclei, thus
16 affecting the abundance and vertical distribution of clouds. The direct and indirect effects of
17 particles appear to have significantly offset the global warming effects caused by the buildup of
18 greenhouse gases because the onset of the Industrial Revolution, on a globally averaged basis.
19 However, because the lifetime of particles is much shorter than that required for complete mixing
20 within the Northern Hemisphere, the climate effects of particles generally are felt much less
21 homogeneously than are the effects of long-lived greenhouse gases.
22 Any effort to model the impacts of local alterations in particle concentrations on projected
23 global climate change or consequent local and regional weather patterns would be subject to
24 considerable uncertainty. This also would be the case for any projections of impacts of location-
25 specific airborne PM alterations on potential human health or environmental effects associated
26 with either increased atmospheric transmission of solar UV radiation or global warming
27 secondary to accumulation of stratospheric ozone-deleting substances or "greenhouse gases."
March 2001 4-164 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Abdullah, U. M; Iqbal, M. Z. (1991) Response of automobile, stone and cement paniculate matters on stomatal
3 clogging of plants. Geobios (Jodhpur, India) 18: 196-202.
4 Aber, J. D.; Nadelhoffer, K. J.; Steudler, P.; Melillo, J. M. (1989) Nitrogen saturation in northern forest ecosystems:
5 excess nitrogen from fossil fuel combustion may stress the biosphere. Bioscience 39: 378-386.
6 Aber, J.; McDowell, W.; Nadelhoffer, K.; Magill, A.; Berntson, G.; Kamakea, M.; McNulty, S.; Currie, W.;
7 Rustad, L.; Fernandez, I. (1998) Nitrogen saturation in temperate forest ecosystems. BioScience 48: 921-934.
8 Adams, K. M.; Kavis, L. L, Jr.; Japar, S. M.; Pierson, W. R. (1989) Real-time, in situ measurements of atmospheic
9 optical absorption in the visible via photoacoustic spectroscopy - II. validation for atmospheric elemental
10 carbon aerosol. Atmos. Environ. 23: 693-700.
11 Alexander, M. (1977) Introduction to soil microbiology. 2nd ed. New York, NY: John Wiley & Sons; pp. 225-230;
12 239-246; 272-286; 355-356.
13 Allen, M. F. (1991) The ecology of mycorrhizae. Cambridge, MA: Cambridge University Press.
14 Allen, E. R.; Cabrera, N.; Kim, J.-C. (1994) Atmospheric deposition studies in rural forested areas of the
15 southeastern United States. Presented at: 87th annual meeting & exhibition of the Air & Waste Management
16 Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no.
17 94-WP88.02.
18 Altshuller, A. P.; Linthurst, R. A., eds. (1984) Acidic deposition phenomenon and its effects: critical assessment
19 review papers; volume I, atmospheric sciences; volume II, effects sciences. Washington, DC: U.S.
20 Environmental Protection Agency, Office of Research and Development; EPA report nos.
21 EPA/600/8-83/016aF-bF. Available from: NTIS, Springfield, VA; PB85-100030 and PB85-100048.
22 Andersen, C. P.; Rygiewicz, P. T. (1991) Stress interactions and mycorrhizal plant response: understanding carbon
23 allocation priorities. Environ. Pollut. 73: 217-244.
24 Antonovics, J.; Bradshaw, A. D.; Turner, R. G. (1971) Heavy metal tolerance in plants. In: Cragg, J. B., ed.
25 Advances in ecological research. Volume 7. London, United Kingdom: Academic Press; pp. 1-85.
26 April, R.; Newton, R. (1992) Mineralogy and mineral weathering. In: Johnson, D. W.; Lindberg, S. E., eds.
27 Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
28 Springer-Verlag; pp. 378-425. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
29 Ecological studies: analysis and synthesis: v. 91).
30 Askey, A.; Lyon, S. B.; Thompson, G. E.; Johnson, J. B.; Wood, G. C.; Sage, P. W.; Cooke, M. J. (1993) The effect
31 of fly-ash particulates on the atmospheric corrosion of zinc and mild steel. Corros. Sci. 34: 1055-1081.
32 Ausset, P.; Bannery, F.; Del Monte, M.; Lefevre, R. A. (1998) Recording of pre-industrial atmospheric environment
33 by ancient crusts on stone monuments. Atmos. Environ. 32: 2859-2863.
34 Ayensu, E.; Van R. Claasen, D.; Collins, M.; Dearing, A.; Fresco, L.; Gadgil, M.; Gitay, H.; Glaser, G.; Juma, C.;
35 Krebs, J.; lenton, R.; Lubchenco, J.; McNeely, J. A.; Mooney, H. A.; Pinstrup-Andersen, P.; Ramos, M.;
36 Raven, P.; Reid, W. V; Samper, C.; Sarukhan, J.; Schei, P.; Tundisi, J. G.; Watson, R. T.; Guanhua, X.;
37 Zakri, A. H. (1999) International ecosystem assessment. Science 286: 685-686.
3 8 Babich, H.; Stotzky, G. (1978) Effects of cadmium on the biota: influence of environmental factors. In: Permian, D.,
39 ed. Advances in applied microbiology, volume 23. New York, NY: Academic Press; pp. 55-117.
40 Baedecker, P. A.; Edney, E. O.; Moran, P. J.; Simpson, T. C.; Williams, R. S. (1991) Effects of acidic deposition on
41 materials. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume III: terrestrial,
42 materials, health and visibility effects. Washington, DC: The U.S. National Acid Precipitation Assessment
43 Program. (State of science and technology report no. 19).
44 Baedecker, P. A.; Reddy, M. M.; Reimann, K. J.; Sciammarella, C. A. (1992) Effects of acidic deposition on the
45 erosion of carbonate stone-experimental results from the U.S. National Acid Precipitation Assessment
46 Program (NAPAP). Atmos. Environ. Part B 26: 147-158.
47 Barone, J. B.; Cahill, T. A.; Eldred, R. A.; Flocchini, R. G.; Shadoan, D. J.; Dietz, T. M. (1978) A multivariate
48 statistical analysis of visibility degradation at four California cities. Atmos. Environ. 12: 2213-2221.
49 Barton, K. (1958) The influence of dust on atmospheric corrosion of metals. Werkst. Korros. 8/9: 547-549.
50 Bellan, L. M.; Salmon, L. G.; Cass, G. R. (2000) A study on the human ability to detect soot deposition onto works
51 of art. Environ. Sci. Technol. 34: 1946-1952.
52 Bewley, R. J. F.; Parkinson, D. (1984) Effects of sulphur dioxide pollution on forest soil microorganisms. Can. J.
53 Microbiol. 30: 179-185.
54 Biefer, G. J. (1981) Atmospheric corrosion of steel in the Canadian Arctic. Mater. Perform. 20: 16-19.
March 2001 4-165 DRAFT-DO NOT QUOTE OR CITE
-------
1 Billings, W. D. (1978) Plants and the ecosystem. 3rd ed. Belmont, CA: Wadsworth Publishing Company, Inc.;
2 pp. 1-62, 83-108.
3 Binenko, V. I.; Harshvardhan, H. (1993) Aerosol effects in radiation transfer. In: Jennings, S. G., ed.
4 Aerosol effects on climate. Tucson, AZ: University of Arizona Press; pp. 190-232.
5 Binkley, D.; Valentine, D.; Wells, C.; Valentine, U. (1989) An empirical analysis of the factors contributing to
6 20-year decrease in soil pH in an old-field plantation of loblolly pine. Biogeochemistry 8: 39-54.
7 Bjorn, L. O. (1996) Effects of ozone depletion and increased UV-B on terrestrial ecosystems. Int. J. Environ. Stud.
8 51:217-243.
9 Blackburn, T. E. (1992) Solar photolysis of ozone to singlet D oxygen atoms. J. Geophys. Res. 97: 10,109-10,117.
10 Blandford, J. C. (1994) Transmissometer data reduction and validation (IMPROVE protocol). Fort Collins, CO:
11 Air Resource Specialists, Inc.; technical instruction no. 4400-5000.
12 Blaschke, H. (1990) Mycorrhizal populations and fine root development on Norway spruce exposed to controlled
13 doses of gaseous pollutants and simulated acidic rain treatments. Environ. Pollut. 68: 409-418.
14 Bobbink, R. (1998) Impacts of tropospheric ozone and airborne nitrogenous pollutants on natural and semi-natural
15 ecosystems: a commentary. New Phytol. 139: 161-168.
16 Bobbink, R.; Hornung, M; Roelofs, J. G. M. (1998) The effects of air-borne nitrogen pollutants on species diversity
17 in natural and semi-natural European vegetation. J. Ecol. 86: 717-738.
18 Bolle, H.-J.; Seiler, W.; Bolin, B. (1986) Other greenhouse gases and aerosols: assessing their role for atmospheric
19 radiative transfer. In: Bolin, B.; Doos, B. R.; Jager, J.; Warrick, R. A., eds. The greenhouse effect, climatic
20 change, and ecosystems. Chichester, United Kingdom: John Wiley and Sons; pp. 157-203. (SCOPE 29).
21 Bondietti, E. A.; McLaughlin, S. B. (1992) Evidence of historical influences of acidic deposition on wood and soil
22 chemistry. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient cycling:
23 a synthesis of the integrated forest study. New York, NY: Springer-Verlag: pp. 358-377. (Billings, W. D.;
24 Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological studies analysis and synthesis: v. 91).
25 Bowden, R. D.; Geballe, G. T.; Bowden, W. B. (1989) Foliar uptake of 15N from simulated cloud water by red
26 spruce (Picea rubens) seedlings. Can. J. For. Res. 19: 382-386.
27 Bradley, R.; Burt, A. J.; Read, D. J. (1981) Mycorrhizal infection and resistance to heavy metal toxicity in Calluna
28 vulgaris. Nature (London) 292: 335-337.
29 Brandt, C. J.; Rhoades, R. W. (1972) Effects of limestone dust accumulation on composition of a forest community.
30 Environ. Pollut. 3:217-225.
31 Brandt, C. J.; Rhoades, R. W. (1973) Effects of limestone dust accumulation on lateral growth of forest trees.
3 2 Environ. Pollut. 4: 207-213.
33 Briihl, C.; Crutzen, P. J. (1988) On the disproportionate role of tropo spheric ozone as a filter against solar UV-B
34 radiation. Geophys. Res. Lett. 16: 703-706.
3 5 Bruhn, J. N. (1980) Effects of oxidant air pollution on Ponderosa and Jeffrey pine foliage decomposition
36 [dissertation]. Berkeley, CA: University of California; Dissertation Abstracts International v. 42-07B.
3 7 Brumme, R.; Leimcke, U.; Matzner, E. (1992) Interception and uptake of NH4 and NO3 from wet deposition by
38 above-ground parts of young beech (Fagus silvaticaL.) trees. Plant Soil 142: 273-279.
39 Brunei, J.; Diekmann, M.; Falkengren-Grerup, U. (1998) Effects of nitrogen deposition onfield layer vegetation in
40 south Swedish oak forests. Environ. Pollut. (Oxford, U.K.) 102(suppl. 1): 35-40.
41 Bugini, R.; Tabasso, M. L.; Realini, M. (2000) Rate of formation of black crusts on marble. A case study. J. Cult.
42 Heritage 1: 111-116.
43 Butlin, R. N.; Coote, A. T.; Devenish, M.; Hughes, I. S. C.; Hutchens, C. M.; Irwin, J. G.; Lloyd, G. O.; Massey,
44 S. W.; Webb, A. H.; Yates, T. J. S. (1992a) Preliminary results from the analysis of metal samples from the
45 National Materials Exposure Programme (NMEP). Atmos. Environ. Part B 26: 199-206.
46 Butlin, R. N.; Coote, A. T.; Devenish, M.; Hughes, I. S. C.; Hutchens, C. M.; Irwin, J. G.; Lloyd, G. O.; Massey,
47 S. W.; Webb, A. H.; Yates, T. J. S. (1992b) Preliminary results from the analysis of stone tablets from the
48 National Materials Exposure Programme (NMEP). Atmos. Environ. Part B 26: 189-198.
49 Bytnerowicz, A.; Fenn, M. E. (1996) Nitrogen deposition in California forests: a review. Environ. Pollut.
50 92: 127-146.
51 Cahill, T. A.; Eldred, R. A.; Wakabayashi, P. H. (1996) Trends in fine particle concentrations at Great Smoky
52 Mountains National Park. For presentation at: Annual meeting of the Air & Waste Management Association;
53 June; Nashville, TN. Pittsburgh, PA: Air & Waste Management Association; submitted paper no.
54 96-MP1A.05.
March 2001 4-166 DRAFT-DO NOT QUOTE OR CITE
-------
1 Campbell, G. G.; Schurr, G. G.; Slawikowski, D. E.; Spence, J. W. (1974) Assessing air pollution damage to
2 coatings. J. Paint Technol. 46: 59-71.
3 Camuffo, D. (1995) Physical weathering of stones. In: Saiz-Jimenez, C., ed. The deterioration of monuments:
4 proceedings of the 2nd international symposium on biodeterioration and biodegradation; January 1994;
5 Sevilla, Spain. Sci. Total Environ. 167: 1-14.
6 Cape, J. N. (1993) Direct damage to vegetation caused by acid rain and polluted cloud: definition of critical levels
7 for forest trees. Environ. Pollut. 82: 167-180.
8 Cape, J. N.; Sheppard, L. J.; Fowler, D.; Harrison, A. F.; Parkinson, J. A.; Dao, P.; Paterson, I. S. (1992)
9 Contribution of canopy leaching to sulphate deposition in a Scots pine forest. Environ. Pollut. 75: 229-236.
10 Carey, W. F. (1959) Atmospheric deposits in Britain: a study of dinginess. Int. J. Air Pollut. 2: 1-26.
11 Ceccotti, S. P.; Messick, D. L. (1997) A global review of crop requirements, supply, and environmental impact on
12 nutrient sulphur balance. In: Cram, W. J.; De Kok, L. J.; Stulen, L; Brunold, C.; Rennenberg, H., eds.
13 Sulphur metabolism in higher plants. Leiden, The Netherlands: Backhuys Publishers; pp. 155-163.
14 Chameides, W. L.; Yu, H.; Liu, S. C.; Bergin, M; Zhou, X.; Mearns, L.; Wang, G.; Kiang, C. S.; Saylor, R. D.;
15 Luo, C.; Huang, Y.; Steiner, A.; Giorgi, F. (1999) Case study of the effects of atmospheric aerosols and
16 regional haze on agriculture: an opportunity to enhance crop yields in China through emission controls?
17 Proc. Natl. Acad. Sci. U. S. A. 96: 13626-13633.
18 Chapin, F. S., III. (1991) Integrated responses of plants to stress: a centralized system of physiological responses.
19 Bioscience 41: 29-36.
20 Chapin, F. S., Ill; Bloom, A. J.; Field, C. B.; Waring, R. H. (1987) Plant responses to multiple environmental
21 factors. BioScience 37: 49-57.
22 Chapin, F. S., Ill; Sala, O. E.; Burke, I. C.; Grime, J. P.; Hooper, D. U.; Lauenroth, W. K.; Lombard, A.; Mooney,
23 H. A.; Mosier, A. R.; Naeem, S.; Pacala, S. W.; Roy, J.; Steffen, W. L.; Tilman, D. (1998) Ecosystem
24 consequences of changing biodiversity. BioScience 48: 45-52.
25 Charlson, R. J.; Lovelock, J. E.; Andreae, M. O.; Warren, S. G. (1987) Oceanic phytoplankton, atmospheric
26 sulfphur, cloud albedo and climate. Nature (London) 326: 655-661.
27 Charlson, R. J.; Schwartz, S. E.; Hales, J. M.; Cess, R. D.; Coakley, J. A., Jr.; Hansen, J. E.; Hofmann, D. J. (1992)
28 Climate forcing by anthropogenic aerosols. Science (Washington, DC) 255: 423-430.
29 Chestnut, L. (1997) Methodology for estimating values for changes in visibility at national parks [draft
30 memorandum]. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
31 Planning and Standards; contract no. 68-D-98-001; EPA air docket A-96-56, document no. VI-B-09-(ooo).
32 Chestnut, L. G.; Dennis, R. L. (1997) Economic benefits of improvements invisibility: acid rain. Provisions of the
33 1990 clean air act amendments. J. Air Waste Manage. Assoc. 47: 395-402.
34 Chevone, B. I.; Herzfeld, D. E.; Krupa, S. V.; Chappelka, A. H. (1986) Direct effects of atmospheric sulfate
35 deposition on vegetation. J. Air Pollut. Control Assoc. 36: 813-815.
36 Choudhury, B. J. (1987) Relationships between vegetation indices, radiation absorption, and net photosynthesis
37 evaluated by a sensitivity analysis. Remote Sens. Environ. 22: 209-233.
38 Clarkson, D. T.; Sanderson, J. (1971) Inhibition of the uptake and long-distance transport of calcium by aluminum
39 and other polyvalent cations. J. Exp. Bot. 22: 837-851.
40 Clevenger, T. E.; Saiwan, C.; Koirtyohann, S. R. (1991) Lead speciation of particles on air filters collected in the
41 vicinity of a lead smelter. Environ. Sci. Technol. 25: 1128-1133.
42 Cole, D. W.; Johnson, D. W. (1977) Atmospheric sulfate additions and cation leaching in a Douglas fir ecosystem.
43 Water Resour. Res. 13: 313-317.
44 Cooke, R. U.; Gibbs, G. B. (1994) Crumbling heritage? Studies of stone weathering in polluted atmospheres.
45 Atmos. Environ. 28: 1355-1356.
46 Costanza, R.; d'Arge, R.; De Groot, R.; Farber, S.; Grasso, M.; Harmon, B.; Limburg, K.; Naeem, S.; O'Neill, R. V.;
47 Paruelo, J.; Raskin, R. G.; Sutton, P.; Van Den Belt, M. (1997) The value of the world's ecosystem services
48 and natural capital. Nature (London) 387: 253-259.
49 Cotrufo, M. F.; De Santo, A. V.; Alfani, A.; Bartoli, G.; De Cristofaro, A. (1995) Effects of urban heavy metal
50 pollution on organic matter decomposition in Quercus ilex L. woods. Environ. Pollut. 89: 81-87.
51 Cowling, J. E.; Roberts, M. E. (1954) Paints, varnishes, enamels, and lacquers. In: Greathouse, G. A.; Wessel, C. J.,
52 eds. Deterioration of materials: causes and preventive techniques. New York, NY: Reinhold Publishing
53 Corporation; pp. 596-645.
March 2001 4-167 DRAFT-DO NOT QUOTE OR CITE
-------
1 Cramer, S. D.; McDonald, L. G.; Bhagia, G.; Flinn, D. R.; Linstrom, P. J.; Carter, J. P. (1989) Effects of acid
2 deposition on the atmospheric corrosion of structural metals. Research Triangle Park, NC: U.S.
3 Environmental Protection Agency; internal report.
4 Creighton, P. J.; Lioy, P. J.; Haynie, F. H.; Lemmons, T. J.; Miller, J. L.; Gerhart, J. (1990) Soiling by atmospheric
5 aerosols in an urban industrial area. J. Air Waste Manage. Assoc. 40: 1285-1289.
6 Cronan, C. S.; Grigal, D. F. (1995) Use of calcium/aluminum ratios as indicators of stress in forest ecosystems.
7 J. Environ. Qual. 24: 209-226.
8 Cullen, J. J.; Neale, P. J. (1997) Biological weighting functions for describing the effects of ultraviolet radiation on
9 aquatic ecosystems. In: Hader, D.-P., ed. The effects of ozone depletion on aquatic ecosystems. Austin, TX:
10 Academic Press, pp. 97-118.
11 Dahlgren, R. A. (1994) Soil acidification and nitrogen saturation from weathering of ammonium-bearing rock.
12 Nature 368: 838-841.
13 Daily, G. C. (1997) Introduction: what are ecosystem services? In: Daily, G. C., ed. Nature's services: societal
14 dependence on natural ecosystems. Washington, DC: Island Press; pp. 1-10.
15 Daily, G. C.; Ehrlich, P. R. (1999) Managing Earth's ecosystems: an interdisciplinary challenge. Ecosystems
16 2:277-280.
17 Darley, E. F. (1966) Studies on the effect of cement-kiln dust on vegetation. J. Air Pollut. Control Assoc.
18 16: 145-150.
19 Dassler, H.-G.; Ranft, H.; Rehn, K.-H. (1972) Zur Widerstandsfahigkeit von Geholzen gegeniiber
20 Fluorverbindungen und Schwefeldioxid [The susceptibility of woody plants exposed to fluorine compounds
21 and SO2]. Flora (Jena) 161: 289-302.
22 Dattner, S. L. (1995) Urban visibility in the El Paso del Norte airshed. Presented at: 88th annual meeting &
23 exhibition of the Air & Waste Management Association, volume 3B. Basic sciences: air quality and noise;
24 June; San Antonio, TX. Pittsburgh, PA: Air & Waste Management Association; paper no. 95-WA76.02.
25 Davidson, C. L; Tang, W.; Finger, S.; Etyemezian, V.; Striegel, M. F.; Sherwood, S. I. (2000) Soiling patterns on a
26 tall limestone building: changes over 60 years. Environ. Sci. Technol. 34: 560-565.
27 Davis, G. D.; Shaw, B. A.; Arah, C. O.; Fritz, T. L.; Moshier, W. C.; Simpson, T. C.; Moran, P. J.; Zankel, K. L.
28 (1990) Effects of SO2 deposition on painted steel surfaces. Surf. Interface Anal. 15: 107-112.
29 Day, D.; Malm, W. C.; Kreidenweis, S. M. (1996) Aerosol light scattering measurements as a function of relative
30 humidity at Great Smoky Mountains National Park. In: 89th annual meeting & exhibition of the Air & Waste
31 Management Association; June; Nashville, TN. Pittsburgh, PA: Air & Waste Management Association;
32 paper no. 96-TA1B.06.
33 Day, D. E.; Malm, W. C.; Kreidenweis, S. M. (2000) Aerosol light scattering measurements as a function of relative
34 humidity. J. Air Waste Manage. Assoc. 50: 710-716.
35 Del Monte, M.; Vittori, O. (1985) Air pollution and stone decay: the case of Venice. Endeavour 9: 117-122.
36 Diakumaku, E.; Gorbushina, A. A.; Krumbein, W. E.; Panina, L.; Soukharjevski, S. (1995) Black fungi in marble
3 7 and limestones~an aesthetical, chemical and physical problem for the conservation of monuments.
38 In: Saiz-Jimenez, C., ed. The deterioration of monuments: proceedings of the 2nd international symposium
39 onbiodeterioration andbiodegradation; January 1994; Sevilla, Spain. Sci. Total Environ. 167: 295-304.
40 Dickerson, R. R.; Kondragunta, S.; Stenchikov, G.; Civerolo, K. L.; Doddridge, B. G.; Holben, B. N. (1997)
41 The impact of aerosols on solar ultraviolet radiation and photochemical smog. Science (Washington, DC)
42 278: 827-830.
43 Dobbins, R. A.; Mulholland, G. W.; Bryner, N. P. (1994) Comparison of a fractal smoke optics model with light
44 extinction measurements. Atmos. Environ. 28: 889-897.
45 Dolske, D. A. (1995) Deposition of atmospheric pollutants to monuments, statues, and buildings. Sci. Total
46 Environ. 167: 15-31.
47 Dreisinger, B. R.; McGovern, P. C. (1970) Monitoring atmospheric sulphur dioxide and correlating its effects on
48 crops and forests in the Sudbury area. In: Linzon, S. N., ed. Impact of air pollution on vegetation conference;
49 April; Toronto, ON, Canada. Pittsburgh, PA: Air Pollution Control Association.
50 Driscoll, C. T.; Van Dreason, R. (1993) Seasonal and long-term temporal patterns in the chemistry of Adirondack
51 lakes. Water Air Soil Pollut. 67: 319-344.
52 Driscoll, C. T.; Wyskowski, B. J.; DeStaffan, P.; Newton, R. M. (1989) Chemistry and transfer of aluminum in a
53 forested watershed in the Adirondack region of New York, USA. In: Lewis, T. E., ed. Environmental
54 chemistry and toxicology of aluminum. Chelsea, MI: Lewis Publishers, Inc.; pp. 83-105.
March 2001 4-168 DRAFT-DO NOT QUOTE OR CITE
-------
1 Driscoll, C. T.; Newton, R. M; Gubala, C. P.; Baker, J. P.; Christensen, S. W. (1991) Adirondack mountains.
2 In: Charles, D. F., ed. Acidic deposition and aquatic ecosystems: regional case studies. New York, NY:
3 Springer-Verlag; pp. 133-202.
4 Edney, E. O. (1989) Paint coatings: controlled field and chamber experiments. Research Triangle Park, NC:
5 U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment Laboratory; report
6 no. EPA/600/3-89/032. Available from: NTIS, Springfield, VA; PB89-189849.
7 Edney, E. O.; Stiles, D. C.; Corse, E. W.; Wheeler, M. L.; Spence, J. W.; Haynie, F. H.; Wilson, W. E., Jr. (1988)
8 Field study to determine the impact of air pollutants on the corrosion of galvanized steel. Mater. Perform.
9 27:47-50.
10 Edney, E. O.; Cheek, S. F.; Corse, E. W.; Spence, J. W.; Haynie, F. H. (1989) Atmospheric weathering caused by
11 dry deposition of acidic compounds. J. Environ. Sci. Health Part A 24: 439-457.
12 Eldering, A.; Cass, G. R. (1996) Source-oriented model for air pollutant effects on visibility. J. Geophys. Res.
13 [Atmos.]: 101: 19,343-19,369.
14 Eldering, A.; Larson, S. M.; Hall, J. R.; Hussey, K. J.; Cass, G. R. (1993) Development of an improved image
15 processing based visibility model. Environ. Sci. Technol. 27: 626-635.
16 Eldering, A.; Cass, G. R.; Moon, K. C. (1994) An air monitoring network using continuous particle size distribution
17 monitors: connecting pollutant properties to visibility via Mie scattering calculations. Atmos. Environ.
18 28:2733-2749.
19 Eldering, A.; Hall, J. R.; Hussey, K. J.; Cass, G. R. (1996) Visibility model based on satellite-generated landscape
20 data. Environ. Sci. Tech. 30: 361-370.
21 Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the
22 United States of America. Atmos. Environ. 28: 1009-1019.
23 Eldred, R. A.; Cahill, T. A.; Feeney, P. J. (1993) Comparison of independent measurements of sulfur and sulfate in
24 the IMPROVE network. Presented at: 86th annual meeting and exhibition of the Air & Waste Management
25 Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management Association; paper no.
26 93-RA-l 10.02.
27 Eldred, R. A.; Cahill, T. A.; Flocchini, R. G. (1997) Composition of PM25 and PM10 aerosols in the IMPROVE
28 network. In: Special issue on aerosols and atmospheric optics-part I, [published papers from the international
29 specialty conference; September 1994; Snowbird, UT]. J. Air Waste Manage. Assoc. 47: 194-203.
30 Ellenberg, H. (1987) Floristic changes due to eutrophication. In: Asman, W. A. H.; Diederen, S. M. A., eds.
31 Ammonia and acidification: proceedings of a symposium of the European Association for the Science of Air
32 Pollution (EURASAP); April; Bilthoven, The Netherlands. European Association for the Science of Air
33 Pollution; pp. 301-308.
34 Eller, B. M. (1977) Road dust induced increase of leaf temperature. Environ. Pollut. 13: 99-107.
35 Eriksson, P.; Johansson, L.-G.; Strandberg, H. (1993) Initial stages of copper corrosion in humid air containing SO2
36 andNO2. J. Electrochem. Soc. 140: 53-59.
37 Erlick, C.; Frederick, J. E.; Saxena, V. K.; Wenny, B. N. (1998) Atmospheric transmission in the ultraviolet and
38 visible: aerosols in cloudy atmospheres. J. Geophys. Res. 103: 31,541-31,556.
39 Estupinan, J. B.; Raman, S.; Crescenti, G. H.; Streicher, J. J.; Barnard, W. F. (1996) Effects of clouds and haze on
40 UV-B radiation. J. Geophys. Res. 101: 16,807-16,816.
41 Evans, E. G.; Pitchford, M. (1991) Federal visibility monitoring in the eastern United States. Presented at:
42 84th annual meeting and exhibition of the Air & Waste Management Association; June; Vancouver, BC,
43 Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-49.6.
44 Falkengren-Grerup, U. (1987) Long-term changes in pH of forest soils in southern Sweden. Environ. Pollut.
45 43:79-90.
46 Falkengren-Grerup, U. (1998) Nitrogen response of herbs and graminoids in experiments with simulated acid soil
47 solution. Environ. Pollut. 102(suppl. 1): 93-99.
48 Farber, R. J.; Welsing, P. R.; Rozzi, C. (1994) PM10 and ozone control strategy to improve visibility in the
49 Los Angeles Basin. Atmos. Environ. 28: 3277-3283.
50 Farmer, A. M. (1993) The effects of dust on vegetation~a review. Environ. Pollut. 79: 63-75.
51 Fenn, M. E.; Poth, M. A.; Johnson, D. W. (1996) Evidence for nitrogen saturation in the San Bernardino Mountains
52 in southern California. For. Ecol. Manage. 82: 211-230
53 Fenn, M. E.; Poth, M. A.; Aber, J. D.; Baron, J. S.; Bormann, B. T.; Johnson, D. W.; Lemly, A. D.; McNulty, S. G.;
54 Ryan, D. F.; Stottlemyer, R. (1998) Nitrogen excess in North American ecosystems: predisposing factors,
55 ecosystem responses, and management strategies. Ecol. Appl. 8: 706-733.
March 2001 4-169 DRAFT-DO NOT QUOTE OR CITE
-------
1 Fink, F. W.; Buttner, F. H.; Boyd, W. K. (1971) Final report on technical-economic evaluation of air-pollution
2 corrosion costs on metals in the U.S. Columbus, OH: Battelle Memorial Institute, Columbus Laboratories;
3 EPA contract no. CPA 70-86; EPA report no. APTD-0654. Available from: NTIS, Springfield, VA;
4 PB-198-453.
5 Flowers, E. C.; McCormick, R. A.; Kurfis, K. R. (1969) Atmospheric turbidity over the United States, 1961-1966.
6 J. Appl. Meteorol. 8: 955-962.
7 Fochtman, E. G.; Langer, G. (1957) Automobile paint damaged by airborne iron particles. J. Air Pollut. Control
8 Assoc. 6: 243-247.
9 Foster, J. R. (1991) Effects of organic chemicals in the atmosphere on terrestrial plants. In: Moser, T. J.; Barker,
10 J. R.; Tingey, D. T., eds. Ecological exposure and effects of airborne toxic chemicals: an overview. Corvallis,
11 OR: U.S. Environmental Protection Agency, Environmental Research Laboratory; pp. 60-89; EPA report no.
12 EPA-600/3-91-001. Available from: NTIS, Springfield, VA; PB91-148460.
13 Foster, N. W.; Nicolson, J. A.; Hazlett, P. W. (1989) Temporal variation in nitrate and nutrient cations in drainage
14 waters from a deciduous forest. J. Environ. Qual. 18: 238-244.
15 Fox, D.; McDonald, K.; Zannetti, P.; Nejedley, Z. (1997) Impact of north-western emission changes on visibility in
16 the Rocky Mountain parks. Presented at: Air & Waste Management Association's 90th annual meeting &
17 exhibition; June; Toronto, Ontario, Canada. Pittsburgh, PA: Air & Waste Management Association; paper
18 no. 97-WA70.03.
19 Fox, D. G.; Malm, W. C.; Mitchell, B.; Fisher, R. W. (1999) Where there's fire, there's smoke: fine paniculate and
20 regional haze. EM November: 15-24.
21 Frederick, J. E.; Qu, Z.; Booth, C. R. (1998) Ultraviolet radiation at sites on the Antarctic coast. J. Photochem.
22 Photobiol. B 68: 183-190.
23 Freedman, B.; Hutchinson, T. C. (1981) Sources of metal and elemental contamination of terrestrial environments.
24 In: Lepp, N. W., ed. Effect of heavy metal pollution on plants. Volume 2: metals in the environment. London,
25 United Kingdom: Applied Science Publishers; pp. 35-94.
26 Freeman, A. M., III. (1993) The measurement of environmental and resource values: theory and methods.
27 Washington, DC: Resources for the Future; pp. 1-27.
28 Freudenburg, W. R.; Alario, M. (1999) What ecologists can learn from nuclear scientists. Ecosystems 2: 286-291.
29 Fuller, K. A.; Malm, W. C.; Kreidenweis, S. M. (1999) Effects of mixing on extinction by carbonaceous particles.
30 J. Geophys. Res. 104: 15,941-15,954.
31 Furst, P.; Krause, G. H. M.; Hein, D.; Delschen, T.; Wilmers, K. (1993) PCDD/PCDF in cow's milk in relation to
32 their levels in grass and soil. Chemosphere 27: 1349-1357.
33 Gabruk, R. S.; Sykes, R. I.; Seigneur, C.; Pai, P.; Gillespie, P.; Bergstrom, R. W.; Saxena, P. (1999) Evaluation of
34 the reactive and optics model of emissions (ROME). Atmos. Environ. 33: 383-399.
35 Gaggi, C.; Bacci, E.; Calamari, D.; Fanelli, R. (1985) Chlorinated hydrocarbons in plant foliage: an indication of the
36 tropospheric contamination level. Chemosphere 14: 1673-1686.
37 Garcia-Valles, M.; Vendrell-Saz, M.; Molera, J.; Blazquez, F. (1998) Interaction of rock and atmosphere: patinas on
38 Mediterranean monuments. Environ. Geol. 36(1-2): 137-149.
39 Garner, J. H. B. (1991) Ozone exposure and nitrogen loading in the forests of eastern North America. In: Berglund,
40 R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
41 international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
42 Association; pp. 289-310. (A&WMA transactions series no. TR-19).
43 Garner, J. H. B. (1994) Nitrogen oxides, plant metabolism, and forest ecosystem response. In: Alscher, R. G.;
44 Wellburn, A. R., eds. Plant responses to the gaseous environment: molecular, metabolic and physiological
45 aspects, [3rd international symposium on air pollutants and plant metabolism]; June 1992; Blacksburg, VA.
46 London, United Kingdom: Chapman & Hall; pp. 301-314.
47 Garner, J. H. B.; Pagano, T.; Cowling, E. B. (1989) An evaluation of the role of ozone, acid deposition, and other
48 airborne pollutants in the forests of eastern North America. Asheville, NC: U.S. Department of Agriculture,
49 Forest Service, Southeastern Forest Experiment Station; general technical report SE-59.
50 Garten, C. T., Jr. (1988) Fate and distribution of sulfur-35 in yellow poplar and red maple trees. Oecologia
51 76:43-50.
52 Garten, C. T., Jr.; Hanson, P. J. (1990) Foliar retention of 15N-nitrate and 15N-ammoniumby red maple
53 (Acer rubrum) and white oak (Quercus alba) leaves from simulated rain. Environ. Exp. Bot. 30: 333-342.
54 Gawel, J. E.; Ahner, B. A.; Friedland, A. J.; Morel, F. M. M. (1996) Role for heavy metals in forest decline
5 5 indicated by phytochelatin measurements. Nature 3 81: 64-65.
March 2001 4-170 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ghedini, N.; Gobbi, G.; Sabbioni, C.; Zappia, G. (2000) Determination of elemental and organic carbon on
2 damaged stone monuments. Atmos. Environ. 34: 4383-4391.
3 Gildon, A.; Tinker, P. B. (1983) Interactions of vesicular-arbuscular mycorrhizal infection and heavy metals in
4 plants: I. the effects of heavy metals on the development of vesicular-arbuscular mycorrhizas. New Phytol.
5 95:247-261.
6 Gilliam, F. S.; Adams, M. B.; Yurish, B. M. (1996) Ecosystem nutrient responses to chronic nitrogen inputs at
7 Fernow Experimental Forest, West Virginia. Can. J. For. Res. 26: 196-205.
8 Gobbi, G.; Zappia, G.; Sabbioni, C. (1998) Sulphite quantification on damaged stones and mortars. Atmos. Environ.
9 32:783-789.
10 Goodwin, J. E.; Sage, W.; Tilly, G. P. (1969) Study of erosion by solid particles. Proc. Inst. Mech. Eng.
11 184: 279-292.
12 Graedel, T. E.; Nassau, K.; Franey, J. P. (1987) Copper patinas formed in the atmosphere—I. introduction.
13 Corros. Sci. 27: 639-657.
14 Grand Canyon Visibility Transport Commission. (1996) Report of the Grand Canyon Visibility Transport
15 Commission to the United States Environmental Protection Agency.
16 Grant, L. D. (1989) Health effects issues associated with regional and global air pollution problems. In: Proceedings
17 of the changing atmosphere: implications for global security; June, 1988; Toronto, Canada. Geneva,
18 Switzerland: World Meteorological Organization; report no. WMO-No. 710.
19 Gray, H. A.; Cass, G. R. (1998) Source contributions to atmospheric fine carbon particle concentrations.
20 Atmos. Environ. 32: 3805-3825.
21 Greenberg, B. M.; Wilson, M. L; Huang, X.-D.; Duxbury, C. L.; Gerhardt, K. E.; Gensemer, R. W. (1997)
22 The effects of ultraviolet-B radiation on higher plants. In: Wang, W.; Gorsuch, J. W.; Hughes, J. S., eds.
23 Plants for Environmental Studies. New York, NY: Lewis Publishers; pp. 1-35.
24 Guderian, R. (1977) Accumulation of pollutants in plant organs. In: Air pollution: phytotoxicity of acidic gases and
25 its significance in air pollution control. Berlin, Germany: Springer-Verlag; pp. 66-74. (Ecological studies:
26 v. 22).
27 Guderian, R., ed. (1985) Air pollution by photochemical oxidants: formation, transport, control, and effects on
28 plants. New York, NY: Springer-Verlag. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
29 Remmert, H., eds. Ecological studies: analysis and synthesis: v. 52).
30 Guderian, R. (1986) Terrestrial ecosystems: paniculate deposition. In: Legge, A. H.; Krupa, S. V., eds.
31 Air pollutants and their effects on the terrestrial ecosystem. New York, NY: John Wiley & Sons; pp.
32 339-363. (Advances in environmental science and technology: v. 18).
33 Gundersen, P.; Callesen, L; De Vries, W. (1998a) Nitrate leaching in forest ecosystems is related to forest floor C/N
34 ratios. Environ. Pollut. 102(suppl. 1): 403-407.
35 Gundersen, P.; Emmett, B. A.; Kjonaas, O. J.; Koopmans, C. J.; Tietema, A. (1998b) Impact of nitrogen deposition
36 on nitrogen cycling in forests: a synthesis of NITREX data. For. Ecol. Manage. 101: 37-55.
37 Hamilton, R. S.; Revitt, D. M.; Vincent, K. J.; Butlin, R. N. (1995) Sulphur and nitrogen paniculate pollutant
3 8 deposition on to building surfaces. In: Saiz-Jimenez, C., ed. The deterioration of monuments: proceedings of
39 the 2nd international symposium onbiodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci.
40 Total Environ. 167: 57-66.
41 Han, Q.; Rossow, W. B.; Lacis, A. A. (1994) Near-global survey of effective droplet radii in liquid water clouds
42 using ISCCP data. J. Clim. 7: 465-497.
43 Haneef, S. J.; Johnson, J. B.; Thompson, G. E.; Wood, G. C. (1993) Simulation of the degradation of limestones
44 and dolomitic sandstone under dry deposition conditions. In: Corrosion control for low-cost reliability:
45 preceedings [of the] 12th international corrosion congress. Volume 2: process industries plant operation.
46 Houston, TX: National Association of Corrosion Engineers; pp. 700-710.
47 Hanley, N.; Spash, C. L. (1993) Cost-benefit analysis and the environment. Hants, United Kingdom: Edward Elgar.
48 Harwell, M. A.; Myers, V; Young, T.; Bartuska, A.; Gassman, N.; Gentile, J. H.; Harwell, C. C.; Appelbaum, S.;
49 Barko, J.; Causey, B.; Johnson, C.; McLean, A.; Smola, R.; Templet, P.; Tosini, S. (1999) A framework for
50 an ecosystem integrity report card. BioScience 49: 543-556.
51 Hauhs, M. (1989) Lange Bramke: an ecosystem study of a forested catchment. In: Adriano, D. C.; Havas, M., eds.
52 Acidic precipitation: volume 1, case studies. New York, NY: Springer-Verlag; pp. 275-305. (Adriano, D. C.;
53 Salomons, W., eds. Advances in environmental science).
March 2001 4-171 DRAFT-DO NOT QUOTE OR CITE
-------
1 Haynie, F. H. (1976) Air pollution effects on stress induced intergranular corrosion of 7005-T53 aluminum alloy.
2 In: Stress corrosion-new approaches. Philadelphia, PA: American Society for Testing and Materials;
3 pp. 32-43. (ASTM special technical publication 610).
4 Haynie, F. H. (1980) Theoretical air pollution and climate effects on materials confirmed by zinc corrosion data.
5 In: Sereda, P. J.; Litvan, G. G., eds. Durability of building materials and components: proceedings of the
6 1st international conference; August 1978; Ottawa, ON, Canada. Philadelphia, PA: American Society for
7 Testing and Materials; pp. 157-175. (ASTM special technical publication 691).
8 Haynie, F. H. (1986) Environmental factors affecting corrosion of weathering steel. In: Baboian, R., ed. Materials
9 degradation caused by acid rain: developed from the 20th state-of-the-art symposium of the American
10 Chemical Society; June 1985; Arlington, VA. Washington, DC: American Chemical Society; pp. 163-171.
11 (ACS symposium series 318).
12 Haynie, F. H.; Lemmons, T. J. (1990) Paniculate matter soiling of exterior paints at a rural site. Aerosol Sci.
13 Technol. 13: 356-367.
14 Haynie, F. H.; Upham, J. B. (1974) Correlation between corrosion behavior of steel and atmospheric pollution data.
15 In: Coburn, S. K., ed. Corrosion in natural environments: presented at the 76th annual meeting American
16 Society for Testing and Materials; June 1973; Philadelphia, PA. Philadelphia, PA: American Society for
17 Testing and Materials; pp. 33-51. (ASTM special technical publication 558).
18 Haywood, J. M.; Shine, K. P. (1995) The effect of anthropogenic sulfate and soot aerosol on the clear sky planetary
19 radiation budget. Geophys. Res. Lett. 22: 603-606.
20 Haywood, J. M.; Ramaswamy, V.; Soden, B. J. (1999) Tropospheric aerosol climate forcing in clear-sky satellite
21 observations over the oceans. Science (Washington, DC) 283: 1299-1303.
22 Heal, G. (2000) Valuing ecosystem services. Ecosystems 3: 24-30.
23 Hedin, L. O.; Granat, L.; Likens, G. E.; Buishand, T. A.; Galloway, J. N.; Butler, T. J.; Rodhe, H. (1994)
24 Steep declines in atmospheric base cations in regions of Europe and North America. Nature (London)
25 367: 351-354.
26 Heinsdorf, D. (1993) The role of nitrogen in declining Scots pine forests (Pinus sylvestris) in the lowland of
27 East Germany. Water Air Soil Pollut. 69:21-35.
28 Herbert, D. A.; Rastetter, E. B.; Shaver, G. R.; Agren, G. I. (1999) Effects of plant growth characteristics on
29 biogeochemistry and community composition in a changing climate. Ecosystems 2: 367-382.
30 Hicks, B. B. (1986) Differences in wet and dry particle deposition parameters between North America and Europe.
31 In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk assessment and
32 control strategies; proceedings of the second U.S.-Dutch international symposium; May 1985; Williamsburg,
33 VA. Chelsea, MI: Lewis Publishers; pp. 973-982.
34 Hicks, B. B.; Baldocchi, D. D.; Meyers, T. P.; Hosker, R. P., Jr.; Matt, D. R. (1987) A preliminary multiple
3 5 resistance routine for deriving dry deposition velocities from measured quantities. Water Air Soil Pollut.
36 36:311-330.
37 Hinckley, T.; Ford, D.; Segura, G.; Sprugel, D. (1992) Key processes from tree to stand level. In: Wall, G., ed.
3 8 Implications of climate change for Pacific Northwest forest management. Waterloo, ON, Canada: University
39 of Waterloo, Department of Geography; pp. 33-43. (Department of Geography publication series: occasional
40 paper no. 15).
41 Hogan, G. D.; Rennenberg, H.; Fink, S. (1998) Role and effect of sulfur in tree biology. In: Maynard, D. G., ed.
42 Sulfur in the environment. New York, NY: Marcel Dekker, Inc.; pp. 173-217. (Books in soils, plants, and the
43 environment: v. 67).
44 Holbrow, G. L. (1962) Atmospheric pollution: its measurement and some effects on paint. J. Oil Colour Chem.
45 Assoc. 45: 701-718.
46 Holling, C. S. (1986) The resilience of terrestrial ecosystems: local surprise and global change. In: Clark, W. C.;
47 Munn, R. E., eds. Sustainable development of the biosphere. Cambridge, UK: Cambridge University;
48 pp. 292-317.
49 Hooper, D. U.; Vitousek, P. M. (1997) The effects of plant composition and diversity on ecosystem processes.
50 Science 277: 1302-1305.
51 Hornung, M.; Langan, S. J. (1999) Nitrogen deposition, sources, impacts and responses in natural and semi-natural
52 ecosystems. In: Bangan, S. J., ed. Impact of nitrogen deposition on natural ecosystems and semi-natural
53 ecosystems. Dordrect, Netherlands: Kluwer Academic Publishers; pp. 1-14. [Environmental Pollution, no.3].
54 Horvath, H. (1993) Atmospheric light absorption-a review. Atmos. Environ. Part A 27: 293-317.
March 2001 4-172 DRAFT-DO NOT QUOTE OR CITE
-------
1 Houghton, J. T.; Jenkins, G. J.; Ephraums, J. J., eds. (1990) Climate change: the IPCC scientific assessment.
2 Cambridge, MA: Cambridge University Press; p. 55.
3 Hoyt, D. V. (1978) A model for the calculation of solar global insolation. Sol. Energy 21: 27-35.
4 Hughes, M. K. (1981) Cycling of trace metals in ecosystems. In: Lepp, N. W., ed. Effect of heavy metal pollution
5 on plants. Volume 2: metals in the environment. London, United Kingdom: Applied Science Publishers;
6 pp. 95-118.
7 Hughes, J. B.; Daily, G. C.; Ehrlich, P. R. (1997) The loss of population diversity and why it matters. In: Raven,
8 P. H., ed. Nature and Human Society: the quest for a sustainable world. Washington, DC: National Academy
9 Press; pp. 71-83.
10 Humphreys, F. R.; Lambert, M. J.; Kelly, J. (1975) The occurrence of sulphur deficiency in forests. In: McLachlan,
11 K. D., ed. Sulphur in Australasian agriculture. Sydney, Australia: Sydney University Press; pp. 154-162.
12 Huntington, T.G. (2000) The potential for calcium depletion in forest ecosystems of southeastern United States:
13 review and analysis. Global Biogeochem. Cycles 14: 623-638.
14 Husar, R. B.; Wilson, W. E. (1993) Haze and sulfur emission trends in the eastern United States. Environ. Sci.
15 Technol. 27: 12-16.
16 Husar, R. B.; Holloway, J. M.; Patterson, D. E.; Wilson, W. E. (1981) Spatial and temporal pattern of eastern U.S.
17 haziness: a summary. Atmos. Environ. 15: 1919-1928.
18 Husar, R. B.; Elkins, J. B.; Wilson, W. E. (1994) U.S. visibility trends, 1960-1992, regional and national. Presented
19 at: 87th annual meeting of the Air & Waste Management Association; June; Cincinnati, OH. Pittsburgh, PA:
20 Air & Waste Management Association; paper no. 94-MP3.05.
21 Husar, R. B.; Schichtel, B. A.; Falke, S. R.; Wilson, W. E. (2000) Haze trends over the United States, 1980-1995.
22 J. Geophys. Res. Lett.: In Press.
23 Hutchinson, A. J.; Johnson, J. B.; Thompson, G. E.; Wood, G. C.; Sage, P. W.; Cooke, M. J. (1992) The role of
24 fly-ash paniculate material and oxide catalysts in stone degradation. Atmos. Environ. 26A: 2795-2803.
25 Huttunen, S. (1994) Effects of air pollutants on epicuticular wax structure. In: Percy, K. E.; Cape, J. N.; Jagels, R.;
26 Simpson, C. J., eds. Air pollutants and the leaf cuticle. New York, NY: Springer-Verlag; pp. 81-96.
27 (NATO ASI series).
28 Intergovernmental Panel on Climate Change (IPCC). (1995) Climate change 1994: radiative forcing of climate
29 change and an evaluation of the IPCC IS92 emission scenarios. Cambridge, United Kingdom: University
30 Press.
31 Intergovernmental Panel on Climate Change (IPCC). (1996) Climate change 1996: contribution of working group
32 I to the second assessment of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom:
33 Cambridge University Press; p. 572. [IPCC second assessment report].
34 Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
35 of vulnerability. Cambridge, United Kingdom: Cambridge University Press.
36 Irving, P. M., ed. (1991) Acidic deposition: state of science and technology: volumes I-IV. Washington, DC:
37 The U.S. National Acid Precipitation Assessment Program.
3 8 Iyer, H.; Patterson, P.; Malm, W. C. (2000) Trends in the extremes of sulfur concentration distributions. J. Air
39 Waste Manage. Assoc. 50: 802-808.
40 Jacobson, M. Z. (1998) Studying the effects of aerosols on vertical photolysis rate coefficient and temperature
41 profiles over an urban airshed. J. Geophys. Res. [Atmos.] 103: 10,593-10,604.
42 Jacobson, M. Z. (1999) Isolating nitrated and aromatic aerosols and nitrated aromatic gases as sources of ultraviolet
43 light absorption. J. Geophys. Res. 104 (D3): 3527-3542.
44 Janzen, J. (1980) Extinction of light by highly nonspherical strongly absorbing colloidal particles:
45 spectrophotometric determination of volume distributions for carbon blacks. Appl. Opt. 19: 2977-2985.
46 Japar, S. M.; Szkarlat, A. C.; Pierson, W. R. (1984) The determination of the optical properties of airborne particle
47 emissions from diesel vehicles. Sci. Total Environ. 36: 121-130.
48 Japar, S. M.; Brachaczek, W. W.; Gorse, R. A., Jr.; Norbeck, J. M.; Pierson, W. R. (1986) The contribution of
49 elemental carbon to the optical properties of rural atmospheric aerosols. Atmos. Environ. 20: 1281-1289.
50 Jensen, V. (1974) Decomposition of angiosperm tree leaf litter. In: Dickinson, C. H.; Pugh, G. J. F., eds. Biology of
51 plant litter decomposition: v. I. London, United Kingdom: Academic Press; pp. 69-104.
52 Johnson, D. W. (1984) Sulfur cycling in forests. Biogeochemistry 1: 29-43.
53
54
March 2001 4-173 DRAFT-DO NOT QUOTE OR CITE
-------
1 Johnson, D. W. (1992) Base cation distribution and cycling. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric
2 deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
3 Springer-Verlag, Inc.; pp. 275-333. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
4 Ecological studies, analysis and synthesis: v. 91).
5 Johnson, F. R.; Desvousges, W. H. (1997) Estimating stated preferences with rated-pair data: environmental, health,
6 and employment effects of energy programs. J. Environ. Econ. Manage. 34: 79-99.
7 Johnson, D. W.; Lindberg, S. E., eds. (1992a) Atmospheric deposition and forest nutrient cycling: a synthesis of the
8 integrated forest study. New York, NY: Springer-Verlag, Inc. (Billings, W. D.; Golley, F.; Lange, O. L.;
9 Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
10 Johnson, D. W.; Lindberg, S. E., eds. (1992b) Nitrogen chemistry, deposition, and cycling in forests. In: Johnson,
11 D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated
12 forest study. New York, NY: Springer-Verlag, Inc.; pp. 150-213. (Billings, W. D.; Golley, F.; Lange, O. L.;
13 Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
14 Johnson, D. W.; Mitchell, M. J. (1988) Responses of forest ecosystems to changing sulfur inputs. In: Maynard,
15 D. G., ed. Sulfur in the environment. New York, NY: Marcel Dekker, Inc.; pp. 219-262. [Books in Soils,
16 Plants, and the Environment, v. 67].
17 Johnson, D. W.; Taylor, G. E. (1989) Role of air pollution in forest decline in eastern North America. Water Air
18 Soil Pollut. 48:21-43.
19 Johnson, D. W.; Todd, D. E., Jr. (1998) Harvesting effects on long-term changes in nutrient pools of mixed oak
20 forest. Soil Sci. Soc. Amer. J. 62: 1725-1735.
21 Johnson, C. D.; Latimer, D. A.; Bergstrom, R. W.; Hogo, H. (1980) User's manual for the plume visibility model
22 (PLUVUE). Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
23 Planning and Standards; report no. EPA-450/4-80-032. Available from: NTIS, Springfield, VA;
24 PB81-163297.
25 Johnson, D. W.; Henderson, G. S.; Huff, D. D.; Lindberg, S. E.; Richter, D. D.; Shriner, D. S.; Todd, D. E.;
26 Turner, J. (1982a) Cycling of organic and inorganic sulphur in a chestnut oak forest. Oecologia 54: 141-148.
27 Johnson, D. W.; Cole, D. W.; Bledsoe, C. S.; Cromack, K.; Edmonds, R. L.; Gessel, S. P.; Grier, C. C.; Richards,
28 B. N.; Vogt, K. A. (1982b) Nutrient cycling in forests of the Pacific Northwest. In: Edmonds, R. L., ed.
29 Analysis of coniferous forest ecosystems in the Western United States. Stroudsburg, PA: Hutchinson Ross
30 Publishing Company; pp. 186-232. (US/IBP synthesis series: v. 14).
31 Johnson, D. W.; Van Miegroet, H.; Lindberg, S. E.; Todd, D. E.; Harrison, R. B. (1991a) Nutrient cycling in red
32 spruce forests of the Great Smoky Mountains. Can. J. For. Res. 21: 769-787.
33 Johnson, D. W.; Cresser, M. S.; Nilsson, S. I.; Turner, J.; Ulrich, B.; Binkley, D.; Cole, D. W. (1991b) Soil changes
34 in forest ecosystems: evidence for and probable causes. Proc. R. Soc. Edinburgh Sect. B: Biol. Sci.
35 97B: 81-116.
36 Johnson, D. W.; Swank, W. T.; Vose, J. M. (1993) Simulated effects of atmospheric sulfur deposition on nutrient
37 cycling in a mixed deciduous forest. Biogeochemistry 23: 169-196.
38 Johnson, A. H.; Friedland, A. J.; Miller, E. K.; Siccama, T. G. (1994a) Acid rain and soils of the Adirondacks. III.
39 Rates of soil acidification in a montane spruce-fir forest at Whiteface Mountain, New York. Can. J. For. Res.
40 24:663-669.
41 Johnson, A. H.; Andersen, S. B.; Siccama, T. G. (1994b) Acid rain and soils of the Adirondacks. I. Changes inpH
42 and available calcium, 1930-1984. Can. J. For. Res. 24: 39-45.
43 Johnson, D. W.; Susfalk, R. B.; Brewer, P. F. (1996) Simulated responses of red spruce forest soils to reduced
44 sulfur and nitrogen deposition. J. Environ. Qual. 25: 1300-1309.
45 Johnson, D.W.; Susfalk, R.B.; Brewer, P.,F.; Swank, W.T. (1999) Simulated effects of reduced sulfur, nitrogen, and
46 base cation deposition on soils and solutions in southern Appalachian forests. J. Environ. Qual
47 28: 1336-1346.
48 Joslin, J. D.; Wolfe, M. H. (1992) Red spruce soil chemistry and root distribution across a cloud water deposition
49 gradient. Can. J. For. Res. 22: 893-904.
50 Joslin, J. D.; Wolfe, M. H. (1994) Foliar deficiencies of mature Southern Appalachian red spruce determined from
51 fertilizer trials. Soil Sci. Soc. Am. J. 58: 1572-1579.
52 Joslin, J. D.; Kelly, J. M.; Van Miegroet, H. (1992) Soil chemistry and nutrition of North American spruce-fir
53 stands: evidence for recent change. J. Environ. Qual. 21: 12-30.
54 Justus, C. G.; Murphey, B. B. (1994) Temporal trends in surface irradiance at ultraviolet wavelengths. J. Geophys.
55 Res. 99 (Dl): 1389-1394.
March 2001 4-174 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kalkstein, L. S.; Greene, J. S. (1997) An evaluation of climate/mortality relationships in large U.S. cities and the
2 possible impacts of a climate change. Environ. Health Perspect. 105: 84-93.
3 Karmoker, J. L.; Clarkson, D. T.; Saker, L. R.; Rooney, J. M; Purves, J. V. (1991) Sulphate deprivation depresses
4 the transport of nitrogen to the xylem and the hydraulic conductivity of barley (Hordeum vulgare L.) roots.
5 Planta 185: 269-278.
6 Katsev, I. L.; Zege, E. P. (1994) The modern theory of black object visibility and meteorological visibility range.
7 Atmos. Environ. 28: 763-767.
8 Kelly, J.; Lambert, M. J. (1972) The relationship between sulphur and nitrogen in the foliage of Pinus radiata.
9 Plant Soil 37: 395-407.
10 Kenk, G.; Fischer, H. (1988) Evidence from nitrogen fertilisation in the forests of Germany. Environ. Pollut.
11 54:199-218.
12 Kerker, M. (1969) The scattering of light and other electromagnetic radiation. New York, NY: Academic Press, Inc.
13 (Loebl, E. M., ed. Physical chemistry: a series of monographs, no. 16).
14 Kiehl, J. T.; Briegleb, B. P. (1993) The relative roles of sulfate aerosols and greenhouse gases in climate forcing.
15 Science (Washington, DC) 260: 311-314.
16 Kodama, Y.; Lee, S. D., eds. (1994) The 13th UOEH international symposium and the 2nd pan Pacific cooperative
17 symposium on impact of increased UV-B exposure on human health and ecosystem; October 1993;
18 Kitakyushu, Japan. Kitakyushu, Japan: University of Occupational and Environmental Health.
19 Koschmieder, H. (1924) Theorie der horizontalen Sichtweite II: Kontrast und Sichtweite [Theory of horizontal
20 visibility]. Beitr. Phys. Atmos. 12: 171-181.
21 Koslow, E. E.; Smith, W. H.; Staskawicz, B. J. (1977) Lead-containing particles on urban leaf surfaces.
22 Environ. Sci. Technol. 11: 1019-1021.
23 Krajickova, A.; Mejstfik, V. (1984) The effect of fly-ash particles on the plugging of stomata. Environ. Pollut.
24 36:83-93.
25 Kripke, M. L. (1989) Health effects of stratospheric ozone depletion: an overview. In: Schneider, T.; Lee, S. D.;
26 Wolters, G. J. R.; Grant, L. D. eds. Atmospheric ozone research and its policy implications: proceedings of
27 the Third US-Dutch International Symposium; May, 1988; Nijmegen, The Netherlands. Amsterdam, The
28 Netherlands: Elsevier, pp. 795-802. [Studies in Environmental Science 35].
29 Krupa, S. V.; Legge, A. H. (1986) Single or joint effects of coarse and fine paniculate sulfur aerosols and ozone on
30 vegetation. In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk assessment
31 and control strategies, proceedings of the second U.S.-Dutch international symposium; May 1985;
32 Williamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 879-887.
33 Krupa, S. V.; Legge, A. H. (1998) Sulphur dioxide, paniculate sulphur and its impacts on a boreal forest ecosystem.
34 In: Ambasht, R. S., ed. Modern trends in ecology and environment. Leiden, The Netherlands: Backhuys
35 Publishers; pp. 285-306.
3 6 Kylin, H.; Grimvall, E.; Ostman, C. (1994) Environmental monitoring of polychlorinated biphenyls using pine
37 needles as passive samplers. Environ. Sci. Technol. 28: 1320-1324.
3 8 Lamersdorf, N. P.; Meyer, M. (1993) Nutrient cycling and acidification of a northwest German forest site with high
39 atmospheric nitrogen deposition. For. Ecol. Manage. 62: 323-354.
40 Latimer, D. A. (1988) Plume perceptibility thresholds. Presented at: 81st annual meeting of the Air Pollution
41 Control Association; June; Dallas, TX. Pittsburgh, PA: Air Pollution Control Association; paper no. 88-56.6.
42 Latimer, D. A.; Bergstrom, R. W.; Hayes, S. R.; Liu, M.-K.; Seinfeld, J. H.; Whitten, G. Z.; Wojcik, M. A.; Hillyer,
43 M. J. (1978) The development of mathematical models for the prediction of anthropogenic visibility
44 impairment: volumes I-III. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
45 Air Quality Planning and Standards; EPA report no. EPA-450/3-78-1 lOa-c. Available from: NTIS,
46 Springfield, VA; PB-293118-SET.
47 Laurenroth, W. K.; Milchunas, D. G. (1985) SO2 effects on plant community function. In: Winner, W. E.; Mooney,
48 H. A.; Goldstein, R. A., eds. Sulfur dioxide and vegetation: physiology, ecology, and policy issues. Stanford,
49 CA: Stanford University Press; pp. 454-477.
50 Lawrence, G. B.; Huntington, T. G. (1999) Soil-Calcium depletion linked to acid rain and forest growth in the
51 eastern United States. Reston, VA: U.S. Geological Survey. Available: http://btdqs.usgs.gov/acidrain [1999,
52 November 23].
53 Lawrence, G. B.; David, M. B.; Shortle, W. C. (1995) A new mechanism for calcium loss in forest-floor soils.
54 Nature 378: 162-164.
March 2001 4-175 DRAFT-DO NOT QUOTE OR CITE
-------
1 Legge, A. H. (1980) Primary productivity, sulfur dioxide, and the forest ecosystem: an overview of a case study.
2 In: Miller, P. R., ed. Proceedings of symposium on effects of air pollutants on Mediterranean and temperate
3 forest ecosystems; June; Riverside, CA. Berkeley, CA: U.S. Department of Agriculture, Pacific Southwest
4 Forest and Range Experiment Station; pp. 51-62; Forest Service general technical report no. PSW-43.
5 Available from: NTIS, Springfield, VA; PB81-133720.
6 Lerman, S. L.; Darley, E. F. (1975) Particulates. In: Mudd, J. B.; Kozlowski, T. T., eds. Responses of plants to air
7 pollution. New York, NY: Academic Press; pp. 141-158.
8 Levin, S. A. (1998) Ecosystems and the biosphere as complex adaptive systems. Ecosystems 1: 431-436.
9 Li-Jones, X.; Prospero, J. M. (1998) Variations in the size distribution of non-sea-salt sulfate aerosol in the marine
10 boundary layer at Barbados: impact of African dust. J. Geophys. Res. 103: 16,073-16,084.
11 Likens, G. E.; Driscoll, C. T.; Buso, D. C. (1996) Long-term effects of acid rain: response and recovery of a forest
12 ecosystem. Science (Washington, DC) 272: 244-246.
13 Lindberg, S. E.; Harriss, R. C. (1981) The role of atmospheric deposition in an eastern U.S. deciduous forest.
14 Water Air Soil Pollut. 16: 13-31.
15 Lindberg, S. E.; Lovett, G. M.; Meiwes, K.-J. (1987) Deposition and forest canopy interactions of airborne nitrate.
16 In: Hutchinson, T. C.; Meema, K. M., eds. Effects of atmospheric pollutants on forests, wetlands and
17 agricultural ecosystems. Berlin, Federal Republic of Germany: Springer-Verlag; pp. 117-130. (NATO as
18 1 series: v. G16).
19 Liu, S. C.; McKeen, S. A.; Madronich, S. (1991) Effect of anthropogenic aerosols on biologically active ultraviolet
20 radiation. Geophys. Res. Lett. 18: 2265-2268.
21 Lokke, H.; Bak, J.; Falkengren-Grerup, U.; Finlay, R. D.; Ilvesniemi, H.; Nygaard, P. H.; Starr, M. (1996)
22 Critical loads of acidic deposition for forest soils: is the current approach adequate. Ambio 25: 510-516.
23 Longstreth, J. (1999) Public health consequences of global climate change in the United States—some regions may
24 suffer disproportionately. Environ. Health Perspect. 107(suppl. 1): 169-179.
25 Lorente, J.; Redano, A.; de Cabo, X. (1994) Influence of urban aerosol on spectral solar irradiance. J. Appl.
26 Meteorol. 33:406-415.
27 Lorusso, S.; Marabelli, M.; Troili, M. (1997) Air pollution and the deterioration of historic monuments. J. Environ.
28 Pathol. Toxicol. Oncol. 16: 171-173.
29 Lovett, G. M. (1984) Atmospheric deposition to forests. In: Breece, L.; Hasbrouck, S., eds. Forest responses to
30 acidic deposition: proceedings of the U. S.-Canadian conference; August 1983; Orono, ME. Orono, ME:
31 University of Maine, Orono, Land and Water Resources Center; pp. 7-18.
32 Lovett, G. M.; Lindberg, S. E. (1984) Dry deposition and canopy exchange in a mixed oak forest as determined by
33 analysis of throughfall. J. Appl. Ecol. 21: 1013-1027.
34 Lovett, G. M.; Lindberg, S. E. (1986) Dry deposition of nitrate to a deciduous forest. Biogeochemistry 2: 137-148.
35 Lovett, G. M.; Lindberg, S. E. (1993) Atmospheric deposition and canopy interactions of nitrogen in forests. Can. J.
36 For. Res. 23: 1603-1616.
37 Lowenthal, D. H.; Rogers, C. F.; Saxena, P.; Watson, J. G.; Chow, J. C. (1995) Sensitivity of estimated light
38 extinction coefficients to model assumptions and measurement errors. Atmos. Environ. 29: 751-766.
39 Madronich, S. (1987) Photodissociation in the Atmosphere. 1. Actinic flux and the effects of ground reflections and
40 clouds. J. Geophys. Res. [Atmos.] 92: 9740-9752.
41 Mahadev, S.; Henry, R. C. (1999) Application of a color-appearance model to vision through atmospheric haze.
42 Color Res. Appl. 24: 112-120.
43 Malm, W. C.; Pitchford, M. L. (1997) Comparison of calculated sulfate scattering efficiencies as estimated from
44 size-resolved particle measurements at three national locations. Atmos. Environ. 31: 1315-1325.
45 Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
46 concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
47 Malm, W. C.; Molenar, J. V.; Eldred, R. A.; Sisler, J. F. (1996) Examining the relationship among atmospheric
48 aerosols and light scattering and extinction in the Grand Canyon area. J. Geophys. Res. [Atmos.]
49 101: 19,251-19,265.
50 Malm, W. C.; Day, D.; Kreidenweis, S. M. (1997) Comparison of measured and reconstructed scattering during an
51 intensive field study at Great Smoky Mountains National Park. Presented at: 90th annual meeting and
52 exhibition of the Air & Waste Management Association; June; Toronto, ON, CA. Pittsburgh, PA: Air &
53 Waste Management Association; 97-WA70.02.
March 2001 4-176 DRAFT-DO NOT QUOTE OR CITE
-------
1 Mansfeld, F. (1980) Regional air pollution study: effects of airborne sulfur pollutants on materials. Research
2 Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory;
3 EPA report no. EPA-600/4-80-007. Available from: NTIS, Springfield, VA; PB81-126351.
4 Marchwinska, E.; Kucharski, R. (1987) The combined influence of SO2 and heavy metal-containing particulates on
5 beans, carrots and parsley. Environ. Monit. Assess. 8: 11-25.
6 Markewitz, D.; Richter, D.D.; Allen, L. H.; Urrego, J. B. (1998) Three decades of observed soil acidification in the
7 Calhoun Experimental Forest: Has acid rain made a difference? Soil Sci. Soc. Amer. J. 62: 1428-1439.
8 Marschner, H. (1995) Sulfur. In: Mineral nutrition of higher plants. 2nd ed. London, United Kingdom: Academic
9 Press; pp. 255-265.
10 Martin, M. H.; Coughtrey, P. J. (1981) Impact of metals on ecosystem function and productivity. In: Lepp, N. W.,
11 ed. Effect of heavy metal pollution on plants: volume 2, metals in the environment. Barking, United
12 Kingdom: Applied Science Publishers; pp. 119-158. (Mellanby, K., ed. Pollution monitoring series).
13 Martin, C. E.; Gravatt, D. A.; Loeschen, V. S. (1992) Photo synthetic responses of three species to acute exposures
14 of nitrate- and sulphate-containing aerosols. Atmos. Environ. Part A 26: 381-391.
15 Materna, J. (1984) Impact of atmospheric pollution on natural ecosystems. In: Treshow, M., ed. Air pollution and
16 plant life. New York, NY: John Wiley & Sons; pp. 397-416.
17 Mathai, C. V. (1995) The Grand Canyon visibility. EM (December): 20-31.
18 Matson, P. A.; Parton, W. J.; Power, A. G.; Swift, M. J. (1997) Agricultural intensification and ecosystem
19 properties. Science (Washington, DC) 277: 504-509.
20 May, P. F.; Till, A. R.; Gumming, M. J. (1972) Systems analysis of sulphur kinetics in pastures grazed by sheep.
21 J. Appl. Ecol. 9: 25-49.
22 McGee, E. S.; Mossotti, V. G. (1992) Gypsum accumulation on carbonate stone. Atmos. Environ. Part B
23 26:249-253.
24 McGowan, T. F.; Lipinski, G. E.; Santoleri, J. J. (1993) New rules affect the handling of waste fuels. Chem. Eng.
25 (March): 122-128.
26 McGrattan, K. B.; Walton, W. D.; Putori, A. D., Jr.; Twilley, W. H.; McElroy, J. A.; Evans, D. D. (1995) Smoke
27 plume trajectory from in situ buring of crude oil in Alaska: field experiments. Gaithersburg, MD: U.S.
28 National Institute of Standards and Technology; report no. NISTIR 5764. Available from: NTIS, Springfield,
29 VA;PB96-131560.
30 McGrattan, K. B.; Trelles, J.; Baum, H. R.; Rehm, R. G. (1996) Smoke plume trajectory over two-dimensional
31 terrain. In: Beall, K., ed. Annual conference on fire research: book of abstracts; October; Gaithersburg, MD.
32 Gaithersburg, MD: United States Department of Commerce, National Institute of Standards and Technology;
33 report no. NISTIR 5904; pp. 65-66. Available from: NTIS, Springfield, VA; PB97-153514.
34 McLachlan, M. S. (1996a) Bioaccumulation of hydrophobic chemicals in agricultural food chains. Environ. Sci.
35 Tech. 30: 252-259.
36 McLachlan, M. S. (1996b) Biological uptake and transfer of polychlorinated dibenzo-p-dioxins and dibenzofurans.
37 In: Hester, R. E.; Harrison, R. M., eds. Chlorinated organic micropollutants. Cambridge, United Kingdom:
3 8 Royal Society of Chemistry; pp. 31-52. (Issues in environmental science and technology).
39 McLachlan, M. S. (1999) Framework for the interpretation of measurements of SOCs in plants. Environ. Sci.
40 Technol. 33: 1799-1804.
41 McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
42 aerosols. Atmos. Environ. 23: 497-507.
43 McMurry, P. H.; Zhang, X.; Lee, C.-T. (1996) Issues in aerosol measurement for optics assessments. J. Geophys.
44 Res. [Atmos.] 101: 19,189-19,197.
45 Meiwes, K. J.; Khanna, P. K. (1981) Distribution and cycling of sulphur in the vegetation of two forest ecosystems
46 in an acid rain environment. Plant Soil 60: 369-375.
47 Meng, R. Z.; Karamchandani, P.; Seigneur, C. (2000) Simulation of stack plume opacity. J. Air Waste Manage.
48 Assoc. 50: 869-874.
49 Mercer, G. S. (1994) Transmissometer data reporting (IMPROVE protocol). Fort Collins, CO: Air Resource
50 Specialists, Inc.; technical instruction no. 4500-5100.
51 Mid-Atlantic Regional Assessment (MARA) Team. (2000) Preparing for a changing climate: the potential
52 conswquences of climate variability and change (Mid-Atlantic overview). Washington, DC: U.S. Global
53 Change Research Program (USGCRP).
54
55
March 2001 4-177 DRAFT-DO NOT QUOTE OR CITE
-------
1 Middleton, P. (1997) DAQM-simulated spatial and temporal differences among visibility, PM, and other air quality
2 concerns under realistic emission change scenarios. In: Special issue on aerosols and atmospheric optics-part
3 II, [published papers from the international specialty conference; September 1994; Snowbird, UT]. J. Air
4 Waste Manage. Assoc. 47: 302-316.
5 Mie, G. (1908) Beitrage zur Optik triiber Medien, speziell kolloidaler Metallosungen [Optics of cloudy media,
6 especially colloidal metal solutions]. Ann. Phys. (Leipzig) 25: 377-445.
7 Millar, C. S. (1974) Decomposition of coniferous leaf litter. In: Dickinson, C. H.; Pugh, G. J. F., eds. Biology of
8 plant litter decomposition: volume I. New York, NY: Academic Press; pp. 105-128.
9 Miller, P. R.; McBride, J. R., eds. (1999) Oxidant air pollution impacts in the Montane forests of southern
10 California: a case study of the San Bernardino Mountains. New York, NY: Springer-Verlag.
11 (Ecological Studies: v. 134.)
12 Miller, W. C.; Fornes, R. E.; Gilbert, R. D.; Speer, A.; Spence, J. (1992) Removal of CaCO3 extender in residential
13 coatings by atmospheric acidic deposition. In: Measurement of toxic and related air pollutants: proceedings
14 of the 1992 U.S. EPA/A&WMA international symposium. Pittsburgh, PA: Air & Waste Management
15 Association; pp. 129-134. (A&WMA publication VIP-25).
16 Miller, E. K.; Panek, J. A.; Friedland, A. J.; Kadlecek, J.; Mohnen, V. A. (1993) Atmospheric deposition to a
17 high-elevation forest at Whiteface Mountain, New York, USA. Tellus Ser. B 45: 209-227.
18 Mims, F. M., III. (1996) Significant reduction of UVB caused by smoke from biomass burning in Brazil.
19 Photochem. Photobiol. 64: 814-816.
20 Mitchell, M. J.; Harrison, R. B.; Fitzgerald, J. W.; Johnson, D. W.; Lindberg, S. E.; Zhang, Y.; Autry, A. (1992a)
21 Sulfur distribution and cycling in forest ecosystems. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric
22 deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
23 Springer-Verlag; pp. 90-129. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
24 Ecological studies: analysis and synthesis: v. 91).
25 Mitchell, M. J.; Foster, N. W.; Shepard, J. P.; Morrison, I. K. (1992b) Nutrient cycling in Huntington Forest and
26 Turkey Lakes deciduous stands: nitrogen and sulfur. Can. J. For. Res. 22: 457-464.
27 Molenar, J. V.; Persha, G.; Malm, W. C. (1990) Long path transmissometer for measuring ambient atmospheric
28 extinction. In: Mathai, C. V., ed. Visibility and fine particles: an A&WMA/EPA international specialty
29 conference; October 1989; Estes Park, CO. Pittsburgh PA: Air & Waste Management Association;
30 pp. 293-304. (AWMA transactions series no. TR-17).
31 Molenar, J. V.; Cismoski, D. S.; Tree, R. M. (1992) Intercomparison of ambient optical monitoring techniques.
32 Presented at: 85th annual meeting & exhibition of the Air & Waste Management Association; June;
33 Kansas City, MO. Pittsburgh, PA: Air & Waste Management Association; paper no. 92-60.09.
34 Molenar, J. V.; Malm, W. C.; Johnson, C. E. (1994) Visual air quality simulation techniques. Atmos. Environ.
35 28: 1055-1063.
36 Moran, M. D.; Pielke, R. A. (1994) Delayed shear enhancement in mesoscale atmospheric dispersion. Presented at:
37 8th joint conference on applications of air pollution meteorology with A&WMA; January; Nashville, TN.
38 Boston, MA: Meteorological Society; pp. 96-103.
39 Moropoulou, A.; Bisbikou, K.; Torts, K.; Van Grieken, R.; Zezza, F.; Macri, F. (1998) Origin and growth of
40 weathering crusts on ancient marbles in industrial atmosphere. Atmos. Environ. 32: 967-982.
41 Naeem, S.; Thompson, L. J.; Lawler, S. P.; Lawton, J. H.; Woodfin, R. M. (1994) Declining biodiversity can alter
42 the performance of ecosystems. Nature 368: 734-737.
43 Nash, T. H., III. (1975) Influence of effluents from a zinc factory on lichens. Ecol. Monogr. 45: 183-198.
44 National Acid Precipitation Assessment Program. (1991) Acidic deposition: state of science and technology.
45 Summary report of the National Acid Precipitation Assessment Program. Washington, DC: U. S. National
46 Acid Precipitation Assessment Program.
47 National Oceanic and Atmospheric Administration. (1992) ASOS (Automated Surface Observing System) user's
48 guide. Washington, DC: National Oceanic and Atmospheric Administration.
49 National Park Service. (1998) IMPROVE program goals and objectives. Washington, DC: National Park Service.
50 Available: www2.nature.nps.gov/ard/impr/index.htm [1999, November 24].
51 National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
52 National Academy Press. 3v.
53
March 2001 4-178 DRAFT-DO NOT QUOTE OR CITE
-------
1 National Science and Technology Council. (1998) National acid precipitation assessment program biennial report to
2 Congress: an integrated assessment; executive summary. Silver Spring, MD: U.S. Department of Commerce,
3 National Oceanic and Atmospheric Administration. Available:
4 www.nnic.noaa.gov/CENR/NAPAP/NAPAP_96.htm [1999, November 24].
5 National Weather Service. (1998) Visibility. Washington, DC: National Weather Service, National Oceanic and
6 Atmospheric Administration. Available: www.nws.noaa.gov/asos/vsby.htm [1999, November 24].
7 Nazaroff, W. W.; Cass, G. R. (1991) Protecting museum collections from soiling due to the deposition of airborne
8 particles. Atmos. Environ. Part A 25: 841-852.
9 Neff, W. D. (1997) The Denver brown cloud studies from the perspective of model assessment needs and the role of
10 meteorology. J. Air Waste Manage. Assoc.
11 Nielsen, K. E.; Ladekarl, U. L.; Nornberg, P. (1999) Dynamic soil processes on heathland due to changes in
12 vegetation to oak and Sitka spruce. For. Ecol. Manage. 114: 107-116.
13 Nihlgard, B. (1985) The ammonium hypothesis~an additional explanation to the forest dieback in Europe. Ambio
14 14:2-8.
15 Nikliriska, M.; Laskowski, R.; Maryariski, M. (1998) Effect of heavy metals and storage time on two types of forest
16 litter: basal respiration rate and exchangeable metals. Ecotoxicol. Environ. Safety 41: 8-18.
17 Noble, I. R.; Dirzo, R. (1997) Forests as human-dominated ecosystems. Science (Washington, DC) 277: 522-525.
18 Nord, A. G.; Ericsson, T. (1993) Chemical analysis of thin black layers on building stone. Stud. Conserv. 38: 25-35.
19 Norton, R. A.; Childers, N. F. (1954) Experiments with urea sprays on the peach. Proc. Am. Soc. Hortic. Sci.
20 63:23-31.
21 Novakov, T.; Penner, J. E. (1993) Large contribution of organic aerosols to cloud-condensation-nuclei
22 concentrations. Nature (London) 365: 823-826.
23 Ockenden, W. A.; Stinnes, E.; Parker, C.; Jones, K. C. (1998) Observations on persistent organic pollutants in
24 plants: implications for their use as passive air samplers and for POP cycling. Environ. Sci. Technol.
25 32:2721-2726.
26 Odum, E. P. (1969) The strategy of ecosystem development: an understanding of ecological succession provides a
27 basis for resolving man's conflict with nature. Science (Washington, DC) 164: 262-270.
28 Odum, E. P. (1985) Trends expected in stressed ecosystems. BioScience 35: 419-422.
29 Odum, E. P. (1993) The ecosystem. In: Ecology and our endangered life-support systems. 2nd ed. Sunderland, MA:
30 Sinauer Associates, Inc.; pp. 38-67.
31 Oehme, M.; Biseth, A.; Schlabach, M.; Wiig, O. (1995) Concentrations of polychlorinated dibenzo-p-dioxins,
32 dibenzofurans and non-ortho substituted biphenyls in polar bear milk from Svalbard (Norway). Environ.
33 Pollut. 90: 401-407.
34 Olszyk, D. M.; Bytnerowicz, A.; Takemoto, B. K. (1989) Photochemical oxidant pollution and vegetation: effects of
35 gases, fog, and particles. Environ. Pollut. 61: 11-29.
36 Omar, A. H.; Biegalski, S.; Larson, S. M.; Landsberger, S. (1999) Paniculate contributions to light extinction and
37 local forcing at a rural Illinois site. Atmos. Environ. 33: 2637-2646.
3 8 Ormrod, D. P. (1984) Impact of trace element pollution on plants. In: Treshow, M., ed. Air pollution and plant life.
39 Chichester, United Kingdom: John Wiley & Sons; pp. 291-319.
40 Paerl, H. W.; Bales, J. D.; Ausley, L. W.; Buzzelli, C. P.; Crowder, L. B.; Eby, L. A.; Go, M.; Peierls, B. L.;
41 Richardson, T. L.; Ramus, J. S. (2000) Recent hurricanes result in continuing ecosystem impacts on USA's
42 largest lagoonal estuary: Pamlico Sound, North Carolina. EOS Trans. Am. Geophys. U.: In Press.
43 Pan, W.; Tatang, M. A.; McRae, G. J.; Prinn, R. G. (1997) Uncertainty analysis of direct radiative forcing by
44 anthropogenic sulfate aerosols. J. Geophys. Res. 102 (D18): 21,915-21,924.
45 Pan, W.; Tatang, M. A.; McRae, G. J.; Prinn, R. G. (1998) Uncertainty analysis of indirect radiative forcing by
46 anthropogenic sulfate aerosols. J. Geophys. Res. 103 (D4): 3815-3823.
47 Parish, S. B. (1910) The effect of cement dust on citrus trees. Plant World 13: 288-291.
48 Parker, A. (1955) The destructive effects of air pollution on materials. Colliery Guardian 191: 447-450.
49 Parker, G. G. (1983) Throughfall and stemflow in the forest nutrient cycle. In: Macfadyen, A.; Ford, E. D., eds.
50 Advances in ecological research, v. 13; pp. 57-120.
51 Patz, J. A.; Engelberg, D.; Last, J. (2000a) The effects of changing weather on public health. Ann. Rev. Public
52 Health 21: 271-307.
53
54
March 2001 4-179 DRAFT-DO NOT QUOTE OR CITE
-------
1 Patz, J. A.; McGeehin, M. A.; Bernard, S. M; Ebi, K. L.; Epstein, P. R.; Grambsch, A.; Gubler, D. I; Reiter, P.;
2 Romieu, L; Rose, J. B.; Samet, J. M.; Trtanj, J. (2000b) The potential health impacts of climate variability
3 and change for the United States: executive summary of the report of the health sector of the U.S. national
4 assessment. Environ. Health Perspect. 108: 367-376.
5 Pearcy, R. W.; Bjorkman, O.; Caldwell, M. M.; Keeley, J. E.; Monson, R. K.; Strain, B. R. (1987) Carbon gain by
6 plants in natural environments. BioScience 37: 21-29.
7 Percy, K. E. (1991) Effects of acid rain on forest vegetation: morphological and non-mensurational growth effects.
8 In: Rennie, P. J.; Robitaille, G., eds. Effects of Acid Rain on Forest Resources: Proceedings of the
9 Conference ; June, 1983; Sainte-Foy, Quebec. Ottawa, Canada: Canadian Forestry Service; pp. 97-111.
10 Percy, W. H.; Sutherland, J.; Christensen, J. (1991) Paradoxical relationship between substrates and agonist-induced
11 contractions of opossum esophageal body and sphincter invitro. Dig. Dis. Sci. 36: 1057-1065.
12 Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Dutcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
13 dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
14 Peterjohn, W. T.; Adams, M. B.; Gilliam, F. S. (1996) Symptoms of nitrogen saturation in two central Appalachian
15 hardwood forest ecosystems. Biogeochemistry 35: 507-522.
16 Peterson, J. T.; Flowers, E. C.; Rudisill, J. H. (1978) Urban-rural solar radiation and atmospheric turbidity
17 measurements in the Los Angeles basin. J. Appl. Meteorol. 17: 1595-1609.
18 Peterson, G.; Allen, C. R.; Holling, C. S. (1998) Ecological resilience, biodiversity, and scale. Ecosystems 1: 6-18.
19 Petuskey, W. T.; Richardson, D. A.; Dolske, D. A. (1995) Aspects of the deterioration of sandstone masonry in
20 Anasazi dwelling ruins at Mesa Verde National Park, Colorado, USA. Sci. Total Environ. 167: 145-159.
21 Pimentel, D.; Wilson, C.; McCullum, C.; Huang, R.; Dwen, P.; Flack, J.; Tran, Q.; Saltman, T.; Cliff, B. (1997)
22 Economic and environmental benefits of biodiversity. BioScience 47: 747-757.
23 Pimm, S. L. (1984) The complexity and stability of ecosystems. Nature (London) 307: 321-326.
24 Pitchford, M. L.; Malm, W. C. (1994) Development and applications of a standard visual index. Atmos. Environ.
25 28: 1049-1054.
26 Pitchford, M. L.; McMurry, P. H. (1994) Relationship between measured water vapor growth and chemistry of
27 atmospheric aerosol for Grand Canyon, Arizona, in winter 1990. Atmos. Environ. 28: 827-839.
28 Prescott, C. E.; Parkinson, D. (1985) Effects of sulphur pollution on rates of litter decomposition in a pine forest.
29 Can. J. Bot. 63: 1436-1443.
30 Pryor, S.; Steyn, D. (1994) Visibility and ambient aerosols in southwestern British Columbia during REVEAL
31 (REgional Visibility Experimental Assessment in the Lower Fraser Valley): application of analytical
32 techniques for source apportionment and related visibility impacts of the REVEAL fine paniculate optical
33 and scene data. Victoria, BC, Canada: British Columbia Ministry of Environment, Lands and Parks.
34 Available from: NTIS, Springfield, VA; MIC-96-00314/INS.
35 Pueschel, R. F. (1993) Potential climatic effects of anthropogenic aerosols. In: Jennings, S. G., ed. Aerosol effects
36 on climate. Tucson, AZ: University of Arizona Press; pp. 110-132.
37 Rapport, D. J.; Whitford, W. G. (1999) How ecosystems respond to stress: common properties of arid and aquatic
38 systems. BioScience 49: 193-203.
39 Raynal, D. J.; Joslin, J. D.; Thornton, F. C.; Schaedle, M.; Henderson, G. S. (1990) Sensitivity of tree seedlings to
40 aluminum: III. Red spruce and loblolly pine. J. Environ. Qual. 19: 180-187.
41 Realini, M.; Negrotti, R.; Appollonia, L.; Vaudan, D. (1995) Deposition of paniculate matter on stone surfaces: an
42 experimental verification of its effects on Carrara marble. In: Saiz-Jimenez, C., ed. The deterioration of
43 monuments: proceedings of the 2nd international symposium on biodeterioration and biodegradation;
44 January 1994; Sevilla, Spain. Sci. Total Environ. 167: 67-72.
45 Rennie, P. J. (1955) The uptake of nutrients by mature forest growth. Plant Soil 7: 49-95.
46 Reuss, J. O. (1983) Implications of the calcium-aluminum exchange system for the effect of acid precipitation on
47 soils. J. Environ. Qual. 12: 591-595.
48 Reuss, J. O.; Johnson, D. W. (1986) Acid deposition and the acidification of soils and waters. New York, NY:
49 Springer-Verlag. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological
50 studies: analysis and synthesis: v. 59).
51 Richards, L. W. (1973) Size distribution of pigment particles: measurement and characterization by light-scattering
52 techniques. In: Patton, T. C., ed. Pigment handbook. V. III. Characterization and physical relationships.
53 New York, NY: John Wiley & Sons; pp. 89-100.
54 Richards, L. W. (1999) Use of the deciview haze index as an indicator for regional haze. J. Air Waste Manage.
55 Assoc. 49: 1230-1237.
March 2001 4-180 DRAFT-DO NOT QUOTE OR CITE
-------
1 Richards, L. W.; Blanchard, C. L.; Blumenthal, D. L., eds. (1991) Navajo generating station visibility study.
2 Final Report. Petaluma, CA: Sonoma Technology, Inc.; report no. STI-90200-1124-FR.
3 Riggan, P. J.; Lockwood, R. N.; Lopez, E. N. (1985) Deposition and processing of airborne nitrogen pollutants in
4 Mediterranean-type ecosystems of southern California. Environ. Sci. Technol. 19: 781-789.
5 Rodney, D. R. (1952) The entrance of nitrogen compounds through the epidermis of apple leaves. Proc. Am. Soc.
6 Hortic. Sci. 59: 99-102.
7 Roelofs, J. G. M; Kempers, A. J.; Houdijk, A. L. F. M; Jansen, J. (1985) The effect of air-borne ammonium
8 sulphate onPinus nigra var. maritima in the Netherlands. Plant Soil 84: 45-56.
9 Roelofs, J. G. M.; Boxman, A. W.; Van Dijk, H. F. G. (1987) Effects of airborne ammonium on natural vegetation
10 and forests. In: Asman, W. A. H.; Diederen, H. S. M. A., eds. Ammonia and acidification: proceedings [of a]
11 symposium of the European Association for the Science of Air Pollution (EURAS AP); April; Bilthoven,
12 The Netherlands. European Association for the Science of Air Pollution; pp. 266-276.
13 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993) Sources of fine organic
14 aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks.
15 Environ. Sci. Technol. 27: 1892-1904.
16 Rosen, H.; Hansen, A. D. A.; Gundel, L.; Novakov, T. (1978) Identification of the optically absorbing component in
17 urban aerosols. Appl. Opt. 17: 3859-3861.
18 Rosenberg, N. J.; Blad, B. L.; Verma, S. B. (1983) Microclimate: the biological environment. 2nd ed. New York,
19 NY: John Wiley & Sons.
20 Ross, D. M.; Malm, W. C.; Iyer, H. K. (1997) Human visual sensitivity to plumes with a Gaussian luminance
21 distribution: experiments to develop an empirical probability of detection model. J. Air Waste Manage.
22 Assoc. 47: 370-382.
23 Rovira, A. D.; Davey, C. B. (1974) Biology of the rhizosphere. In: Carson, E. W., ed. The plant root and its
24 environment: proceedings of an institute; July, 1971; Blacksburg, VA. Charlottesville, VA: University Press
25 of Virginia; pp. 153-204.
26 Sabbioni, C. (1995) Contribution of atmospheric deposition to the formation of damage layers. In: Jimenez-Saiz, C.,
27 ed. The deterioration of monuments: proceedings of the 2nd international symposium on biodeterioration and
28 biodegradation; January 1994; Sevilla, Spain. Sci. Total Environ. 167: 49-56.
29 Sabbioni, C.; Zappia, G.; Gobbi, G. (1992) Carbonaceous particles on carbonate building stones in a simulated
30 system. In: Proceedings of the 1992 European aerosol conference; September; Oxford, United Kingdom.
31 J. Aerosol Sci. 23(suppl. 1): S921-S924.
32 Sabbioni, C.; Zappia, G.; Gobbi, G. (1996) Carbonaceous particles and stone damage in a laboratory exposure
33 system. J. Geophys. Res. [Atmos.] 101: 19,621-19,627.
34 Sabbioni, C.; Zappia, G.; Ghedini, N.; Gobbi, G.; Favoni, O. (1998) Black crusts on ancient mortars.
35 Atmos. Environ. 32: 215-223.
36 Saiz-Jimenez, C. (1993) Deposition of airborne organic pollutants on historic buildings. Atmos. Environ. Part B
37 27:77-85.
3 8 Samecka-Cymerman, A.; Kempers, A. J. (1999) Bioindication of heavy metals in the town Wro/caw (Poland) with
39 evergreen plants. Atmos. Environ. 33: 419-430.
40 Sanyal, B.; Singhania, G. K. (1956) Atmospheric corrosion of metals: part I. J. Sci. Ind. Res. Sect. B 15: 448-455.
41 Saunders, P. J. W.; Godzik, S. (1986) Terrestrial vegetation-air pollutant interactions: nongaseous air pollutants.
42 In: Legge, A. H.; Krupa, S. V., eds. Air pollutants and their effects on the terrestrial ecosystem. New York,
43 NY: John Wiley & Sons; pp. 389-394. (Advances in environmental science and technology: v. 18).
44 Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. H. (1995) Organics alter hygroscopic behavior of
45 atmospheric particles. J. Geophys. Res. [Atmos.] 100: 18,755-18,770.
46 Schiavon, N.; Chiavari, G.; Schiavon, G.; Fabbri, D. (1995) Nature and decay effects of urban soiling on granitic
47 building stones. In: Jimenez-Saiz, C., ed. The deterioration of monuments: proceedings of the 2nd
48 international symposium on biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci. Total
49 Environ. 167: 87-101.
50 Schier, G. A.; Jensen, K. (1992) Atmospheric deposition effects on foliar injury and foliar leaching in red spruce.
51 In: Eager, C.; Adams, M. B., eds. Ecology and decline of red spruce in the eastern United States. New York,
52 NY: Springer-Verlag; pp. 271-294. [Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. D.; Remmert, H.,
53 eds. Ecological Studies, Analysis and Synthesis Ecological Studies, v. 96].
March 2001 4-181 DRAFT-DO NOT QUOTE OR CITE
-------
1 Schnug, E. (1997) Significance of sulphur for the quality of domesticated plants. In: Cram, W. J.; De Kok, L. J.;
2 Stolen, L; Brunold, C.; Rennenberg, H., eds. Sulphur metabolism in higher plants: molecular,
3 ecophysiological and nutritional aspects. Leiden, The Netherlands: Backhuys Publishers; pp. 109-130.
4 Schulze, E.-D. (1989) Air pollution and forest decline in a spruce (Picea abies) forest. Science (Washington, DC)
5 244:776-783.
6 Schuster, P. F.; Reddy, M. M; Sherwood, S. I. (1994) Effects of acid rain and sulfur dioxide on marble dissolution.
7 Mater. Perform. 33:76-80.
8 Schwar, M. J. R. (1998) Nuisance dust deposition and soiling rate measurements. Environ. Technol. 19: 223-229.
9 Seckmeyer, G.; McKenzie, R. L. (1992) Increased ultraviolet radiation in New Zealand (45 degrees S) relative to
10 Germany (48 degrees N). Nature 359: 135ff.
11 Seigneur, C.; Wu, X. A.; Constantinou, E.; Gillespie, P.; Bergstrom, R. W.; Sykes, L; Venkatram, A.;
12 Karamchandani, P. (1997) Formulation of a second-generation reactive plume and visibility model. J. Air
13 Waste Manage. Assoc. 47: 176-184.
14 Seinfeld, J. H. (1989) Urban air pollution: state of the science. Science 243: 745-752.
15 Sereda, P. J. (1974) Weather factors affecting corrosion of metals. In: Corrosion in natural environments: three
16 symposia presented at the seventy-sixth annual meeting [of the] American Society for Testing and Materials;
17 June 1973; Philadelphia, PA. Philadelphia, PA: American Society for Testing and Materials; pp. 7-22.
18 (ASTM special technical publication 558).
19 Shannon, J. D.; Trexler, E. C., Jr.; Sonnenblick, R. (1997) Modeling visibility for assessment. Atmos. Environ.
20 31:3719-3727.
21 Sievering, H.; Rusch, D.; Marquez, L. (1996) Nitric acid, paniculate nitrate and ammonium in the continental free
22 troposphere: nitrogen deposition to an alpine tundra ecosystem. Atmos. Environ. 30: 2527-2537.
23 Shodor Education Foundation, Inc.; Industrial Extension Service of North Carolina State University; North Carolina
24 Supercomputing Center. (1996) Air quality meteorology. Session 8: physical meteorology and visibility.
25 Part 3: visibility. Research Triangle Park, NC: U.S. Environmental Protection Agency, National Exposure
26 Research Laboratory and U.S. National Oceanic and Atmospheric Administration. Available:
27 http://www.shodor.org/metweb/outline.html [1999, November 24].
28 Shortle, W. C.; Smith, K. T. (1988) Aluminum-induced calcium deficiency syndrome in declining red spruce.
29 Science (Washington, DC) 240: 1017-1018.
30 Shortle, W. C.; Bondietti, E. A. (1992) Timing, magnitude, and impact of acidic deposition on sensitive forest sites.
31 Water Air Soil Pollut. 61: 253-267.
32 Showak, W.; Dunbar, S. R. (1982) Effect of 1 percent copper addition on atmospheric corrosion of rolled zinc after
33 20 years' exposure. In: Dean, S. W.; Rhea, E. C., eds. Atmospheric corrosion of metals: a symposium;
34 May 1980; Denver, CO. Philadelphia, PA: American Society for Testing and Materials; pp. 135-162.
3 5 (ASTM special technical publication 767).
36 Shriner, D. S.; Henderson, G. S. (1978) Sulfur distribution and cycling in a deciduous forest watershed. J. Environ.
37 Qual. 7: 392-397.
38 Simonich; S. L.; Hites, R. A. (1995) Organic pollutant accumulation in vegetation. Environ. Sci. Technol.
39 29:2905-2914.
40 Simpson, J. W.; Horrobin, P. J., eds. (1970) The weathering and performance of building materials. New York, NY:
41 Wiley-Interscience.
42 Sisler, J. F.; Cahill, T. A. (1993) Spacial and temporal patterns and the chemical composition of the haze and its
43 impact on visibility in Alaska. In: Proceedings [of the] 86th annual meeting & exhibition of the Air & Waste
44 Management Association. Volume 8: general environmental topics; June; Denver, CO. Pittsburgh, PA: Air &
45 Waste Management Association; 93-MP-4.03.
46 Sisler, J. F.; Malm, W. C. (1994) The relative importance of soluble aerosols to spatial and seasonal trends of
47 impaired visibility in the United States. Atmos. Environ. 28: 851-862.
48 Sisler, J. F.; Malm, W. C. (2000) Interpretation of trends of PM2 5 and reconstructed visibility from the IMPROVE
49 network. J. Air Waste Manage. Assoc.: submitted.
50 Sisler, J. F.; Huffman, D.; Lattimer, D. A.; Malm, W. C.; Pitchford, M. L. (1993) Spatial and temporal patterns and
51 the chemical composition of the haze in the United States: an anlysis of data from the IMPROVE network,
52 1988-1991. Fort Collins, CO: Colorado State University, Cooperative Institute for Research in the
53 Atmosphere (CIRA).
54 Skeffington, R. A.; Wilson, E. J. (1988) Excess nitrogen deposition: issues for consideration. Environ. Pollut.
55 54: 159-184.
March 2001 4-182 DRAFT-DO NOT QUOTE OR CITE
-------
1 Skerry, B. S.; Alavi, A.; Lindgren, K. I. (1988) Environmental and electrochemical test methods for the evaluation
2 of protective organic coatings. J. Coat. Technol. 60: 97-106.
3 Sloane, C. S.; Watson, J.; Chow, J.; Pritchett, L.; Richards, L. W. (1991) Size-segregated fine particle
4 measurements by chemical species and their impact on visibility impairment in Denver. Atmos. Environ.
5 Part A 25: 1013-1024.
6 Smith, W. H. (1974) Air pollution - effects on the structure and function of the temperate forest ecosystem.
7 Environ. Pollut. 6: 111-129.
8 Smith, W. H. (1990a) Forests as sinks for air contaminants: soil compartment. In: Air pollution and forests:
9 interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag;
10 pp. 113-146. (Springer series on environmental management).
11 Smith, W. H. (1990b) Forest biotic agent stress: air pollutants and disease caused by microbial pathogens. In: Air
12 pollution and forests: interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY:
13 Springer-Verlag; pp. 366-397. (Springer series on environmental management).
14 Smith, W. H. (1990c) Forests as sinks for air contaminants: vegetative compartment. In: Air pollution and forests:
15 interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag;
16 pp. 147-180. (Springer series on environmental management).
17 Smith, W. H. (1990d) Forests as sources of hydrocarbons, particulates, and other contaminants. In: Air pollution
18 and forests: interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY:
19 Springer-Verlag; pp. 83-112. (Springer series on environmental management).
20 Smith, W. H. (1990e) Forest nutrient cycling: toxic ions. In: Air pollution and forests: interactions between air
21 contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag; pp. 225-268. (Springer series
22 on environmental management).
23 Smith, W. H. (1991) Air pollution and forest damage. Chem. Eng. News 69(45): 30-43.
24 Smith, W. H.; Staskawicz, B. J. (1977) Removal of atmospheric particles by leaves and twigs of urban trees:
25 some preliminary observations and assessment of research needs. Environ. Manage. (N. Y.) 1: 317-330.
26 Soulen, P. E.; Frederick, J. E. (1999) Estimating biologically active UV irradiance from satellite radiance
27 measurements: a sensitivity study. J. Geophys. Res. 104: 4117-4126.
28 Spence, J. W.; Haynie, F. H. (1972) Paint technology and air pollution: a survey and economic assessment.
29 Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Programs; publication no.
30 AP-103. Available from: NTIS, Springfield, VA; PB-210736.
31 Spence, J. W.; Haynie, F.; Upham, J. B. (1975) Effects of gaseous pollutants on paints: a chamber study. J. Paint
32 Technol. 47: 57-63.
33 Spence, J. W.; Haynie, F. H.; Lipfert, F. W.; Cramer, S. D.; McDonald, L. G. (1992) Atmospheric corrosion model
34 for galvanized steel structures. Corrosion 48: 1009-1019.
35 Spiker, E. C.; Comer, V. J.; Hosker, R. P.; Sherwood, S. I. (1992) Dry deposition of SO2 on limestone and marble:
36 the role of humidity. In: Rodrigues, J. D.; Henriques, F.; Jeremias, F. T., eds. Proceedings of the 7th
37 international congress on deterioration and conservation of stone; June; Lisbon, Portugal. Lisbon, Portugal:
38 Laboratorio Nacional de Engenharia Civil; pp. 397-406.
39 Spinka, J. (1971) Vliv znecisteneho ovzdusi na ovocne stromy azeleninu [Effects of polluted air on fruit trees and
40 legumes]. Ziva 19: 13-15.
41 Stevens, R. K.; Dzubay, T. G.; Lewis, C. W.; Shaw, R. W., Jr. (1984) Source apportionment methods applied to the
42 determination of the origin of ambient aerosols that affect visibility in forested areas. Atmos. Environ.
43 18:261-272.
44 Stoddard, J. L. (1994) Long-term changes in watershed retention of nitrogen: its causes and aquatic consequences.
45 In: Baker, L. A., ed. Environmental chemistry of lakes and reservoirs. Washington, DC: American Chemical
46 Society; pp. 223-284. (Advances in chemistry series no. 237).
47 Stoddard, J. L.; Murdoch, P. S. (1991) Catskill Mountains. In: Charles, D. F., ed. Acidic deposition and aquatic
48 ecosystems: regional case studies. New York, NY: Springer-Verlag; pp. 237-271.
49 Storm, G. L.; Fosmire, G. J.; Bellis, E. D. (1994) Persistence of metals in soil and selected vertebrates in the vicinity
50 of the Palmerton zinc smelters. J. Environ. Qual. 23: 508-514.
51 Strandberg, H.; Johansson, L.-G. (1997) The formation of black patina on copper in humid air containing traces of
52 SO2. J. Electrochem. Soc. 144: 81-89.
53 Suzuki, D. T. (1997) Barriers to perception: from a world of interconnection to fragmentation. In: Raven, P. H., ed.
54 Nature and Human Society: the quest for a sustainable world. Washington, DC: National Academy Press;
55 pp. 11-21.
March 2001 4-183 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sydberger, T.; Ericsson, R. (1977) Laboratory testing of the atmospheric corrosion of steel. Werkst. Korros.
2 28: 154-158.
3 Sykes, R. I.; Gabruk, R. S. (1997) A second-order closure model for the effect of averaging time on turbulent plume
4 dispersion. J. Appl. Meteorol. 36: 1038-1045.
5 Systems Applications International, Inc. (1998) User's guide to the regulatory modeling system for aerosols and
6 deposition (REMSAD). San Rafael, CA: Systems Applications International, Inc.; report no.
7 SYSAPP98-96\42r2.
8 Tang, I. N. (1997) Thermodynamic and optical properties of mixed-salt aerosols of atmospheric importance.
9 J. Geophys. Res. [Atmos.] 102: 1883-1893.
10 Taylor, G. E., Jr.; Johnson, D. W.; Andersen, C. P. (1994) Air pollution and forest ecosystems: a regional to global
11 perspective. Ecol. Appl. 4: 662-689.
12 Tegen, I.; Lacis, A. A. (1996) Modelling of particle size distribution and its influence on the radiative properties of
13 mineral dust aerosol. J. Geophys. Res. Atmos. 101: 19,237-19,244.
14 Tegen, I.; Lacis, A. A.; Fung, I. (1996) The influence on climate forcing of mineral aerosols from disturbed soils.
15 Nature (London) 380: 419-422.
16 Thornton, F. C.; Schaedle, M; Raynal, D. J. (1987) Effects of aluminum on red spruce seedlings in solution culture.
17 Environ. Exp. Bot. 27: 489-498.
18 Tilman, D. (1996) Biodiversity: population versus ecosystem stability. Ecology 77: 350-363.
19 Tilman, D. (2000) Causes, consequences and ethics of biodiversity. Nature 405: 208-211.
20 Tilman, D.; Downing, J. A. (1994) Biodiversity and stability in grasslands. Nature (London) 367: 363-365.
21 Tong, S. T. Y. (1991) The retention of copper and lead paniculate matter in plant foliage and forest soil.
22 Environ. Int. 17:31-37.
23 Trettin, C.C.; Johnson, D.W.; Todd, D.E., Jr. (1999) Forest nutrient and carbon pools at Walker Branch watershed:
24 changes during a 21-year-period. Soil Sci. Soc. Amer. J. 63: 1436-1448.
25 Trijonis, J.; Shapland, D. (1979) Existing visibility levels in the United States: isopleth maps of visibility in
26 suburban/nonurban areas during 1974-76. Research Triangle Park, NC: U.S. Environmental Protection
27 Agency, Office of Air Quality Planning and Standards; report no. EPA-450/5-79-010. Available from:
28 NTIS, Springfield, VA; PB80-156177.
29 Trijonis, J. C.; Pitchford, M.; McGown, M. (1987) Preliminary extinction budget results from the RESOLVE
30 program. In: Bhardwaja, P. S., ed. Visibility protection: research and policy aspects, an APCA international
31 specialty conference; September 1986; Grand Teton National Park, WY. Pittsburgh, PA: Air Pollution
32 Control Association; pp. 872-883. (APCA transactions series no. TR-10).
33 Trijonis, J. C.; Malm, W. C.; Pitchford, M.; White, W. H. (1991) Visibility: existing and historical
34 conditions-causes and effects. In: Irving, P. M., ed. Acidic deposition: state of science and technology,
3 5 volume III: terrestrial, materials, health and visibility effects. Washington, DC: The U.S. National Acid
36 Precipitation Assessment Program. (State of science and technology report no. 24).
37 Turner, J.; Kelly, J. (1981) Relationships between soil nutrients and vegetation in a north coast forest, New South
38 Wales. Aust. For. Res. 11: 201-208
39 Turner, J.; Lambert, M. J. (1980) Sulfur nutrition of forests. In: Shriner, D. S.; Richmond, C. R.; Lingberg, S. E.,
40 eds. Atmospheric sulfur deposition: environmental impact and health effects, proceedings of the second life
41 sciences symposium; October 1979; Gatlinburg, TN. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.;
42 pp. 321-333.
43 Turner, J.; Lambert, M. J.; Gessel, S. P. (1977) Use of foliage sulphate concentrations to predict response to urea
44 application by Douglas-fir. Can. J. For. Res. 7: 476-480.
45 Twomey, S. (1974) Pollution and the planetary albedo. Atmos. Environ. 8: 1251-1256.
46 Twomey, S. (1977) Atmospheric aerosols. Amsterdam, The Netherlands: Elsevier Scientific Publishing Company;
47 p. 237.
48 Tyler, G. (1972) Heavy metals pollute nature, may reduce productivity. Ambio 1: 52-59.
49 U.S. Code. (1991) Clean Air Act, §109, national ambient air quality standards. U. S. C. 42: §7409.
50 U.S. Environmental Protection Agency. (1979) Protecting visibility: an EPA report to Congress. Research Triangle
51 Park, NC: Office of Air Quality Planning and Standards; EPA report no. EPA-450/5-79-008. Available from:
52 NTIS, Springfield, VA; PB80-220320.
53 U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
54 Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
55 Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
March 2001 4-184 DRAFT-DO NOT QUOTE OR CITE
-------
1 U.S. Environmental Protection Agency. (1986 ) Air quality criteria for lead. Research Triangle Park, NC: Office of
2 Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
3 EPA-600/8-83/028aF-dF. 4v. Available from: NTIS, Springfield, VA; PB87-142378.
4 U.S. Environmental Protection Agency. (1987) Assessing the risks of trace gases that can modify the stratosphere.
5 Washington, DC: U.S. Environmental Protection Agency, Office of Air Radiation; report no. EPA
6 400/1-87/001A-H.
7 U.S. Environmental Protection Agency. (1988) Workbook for plume visual impact screening and analysis. Research
8 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-450/4-88-015.
9 Available from: NTIS, Springfield, VA; PB89-151286.
10 U.S. Environmental Protection Agency. (1992) Interagency workgroup on air quality modeling (IWAQM) work
11 plan rationale. Research Triangle Park, NC: Office of Air Quality Planning and Standards; report no.
12 EPA-454/R-92-001.
13 U.S. Environmental Protection Agency. (1993) Air quality criteria for oxides of nitrogen. Research Triangle Park,
14 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
15 nos. EPA/600/8-9!/049aF-cF. 3v. Available from: NTIS, Springfield, VA; PB95-124533, PB95-124525, and
16 PB95-124517.
17 U.S. Environmental Protection Agency. (1995a) Inter-agency workgroup on air quality modeling (IWAQM):
18 assessment of phase I recommendations regarding the use of MESOPUFF II. Research Triangle Park, NC:
19 U.S. Environmental Protection Agency, Emissions, Monitoring and Analysis Division; report no.
20 EPA-454/R-95-006. Available from: NTIS, Springfield, VA; PB95-241931. P. 106 (1995a).
21 U.S. Environmental Protection Agency. (1995b) Testing of meteorological and dispersion models foruse in
22 regional air quality modeling. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
23 report no. EPA-454/R-95-005. Available from: NTIS, Springfield, VA; PB95-215638.
24 U.S. Environmental Protection Agency. (1995c) Interim findings on the status of visibility research. Research
25 Triangle Park, NC: Office of Research and Development; report no. EPA/600/R-95/021.
26 U.S. Environmental Protection Agency. (1996a) Air quality criteria for paniculate matter. Research Triangle Park,
27 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
28 U.S. Environmental Protection Agency. (1996b) Air quality criteria for ozone and related photochemical oxidants.
29 Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF.
30 3v. Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available
31 online at: www.epa.gov/ncea/ozone.htm.
32 U.S. Environmental Protection Agency. (1997a) Nitrogen oxides: impacts on public health and the environment.
33 Washington, DC: Office of Air and Radiation; August. Available:
34 www.epa.gov/ttncaaal/tl/reports/noxrept.pdf [1999, November 24].
35 U.S. Environmental Protection Agency. (1997b) Summary of the paniculate monitoring program. Washington, DC:
3 6 Office of Air and Radiation.
37 U.S. Environmental Protection Agency. (1999) Protection of visibility. C. F. R. 40: "Section" 51.300-51.309.
38 U.S. Environmental Protection Agency. (2000a) Deposition of air pollutants to the great waters. Third report to
39 Congress. [Executive Summary]. Research Triangle Park, NC: U.S. Environmental Protection Agency,
40 Office of Air Quality Planning and Standards; report no. EPA-453/R-00-005.
41 U.S. Environmental Protection Agency. (2000b) Latest findings on national air quality: 1999 status and trends.
42 Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
43 Standards; report no. EPA-454/F-00-002.
44 U. S. Global Change Research Program (USGCRP). (2000) Climate Change Impacts on the United States: the
45 Potential Consequences of Climate Variability and Change (Overview), Report of National Assessment
46 Synthesis Team (NAST). NSTC Review Draft (September).
47 United Nations Environment Programme (UNEP). (1986) Report of the international conference on the assessment
48 of the role of carbon dioxide and of other greenhouse gases in climate variations and associated impacts;
49 October 1985; Villach, Austria. Paris, France: International Council of Scientific Unions. [World
50 Meteorological Organization series: WMO no. 661].
51 United Nations Environment Programme (UNEP). (1998) Environmental effects of ozone depletion: 1998
52 assessment. J. Photochem. Photobiol. B 46: 1-4.
53 Unsworth, M. H.; Wilshaw, J. C. (1989) Wet, occult and dry deposition of pollutants onforests. Agric. For.
54 Meteorol. 47: 221-238.
March 2001 4-185 DRAFT-DO NOT QUOTE OR CITE
-------
1 Van Breemen, N.; Van Dijk, H. F. G. (1988) Ecosystem effects of atmospheric deposition of nitrogen in The
2 Netherlands. In: Dempster, J. P.; Manning, W. J., eds. Excess nitrogen deposition. Environ. Pollut.
3 54:249-274.
4 Van Breemen, N.; Burrough, P. A.; Velthorst, E. J.; Van Dobben, H. F.; De Wit, T.; Ridder, T. B.; Reijnders, H. F.
5 R. (1982) Soil acidification from atmospheric ammonium sulphate in forest canopy throughfall. Nature
6 (London) 299: 548-550.
7 Van Breemen, N.; Mulder, J.; Driscoll, C. T. (1983) Acidification and alkalinization of soils. Plant Soil
8 75:283-308.
9 Van der Leun, J. C.; Tang, X.; Tevini, M. (1995) Environmental effects of ozone depletion: 1994 assessment.
10 Ambio24:138.
11 Van der Leun, J. C.; Tang, X.; Tevini, M., eds. (1998) Environmental effects of ozone depletion: 1998 assessment.
12 J. Photochem. Photobiol. B: Biology 46(1-3): 1-108.
13 Van Dijk, H. F. G.; Roelofs, J. G. M. (1988) Effects of excessive ammonium deposition on the nutritional status
14 and condition of pine needles. Physiol. Plant. 73: 494-501.
15 Van Dorland, R.; Dentener, F. J.; Lelieveld, J. (1997) Radiative forcing due to tropospheric ozone and sulfate
16 aerosols. J. Geophys. Res. [Atmos.] 102: 28,079-28,100.
17 Van Miegroet, H.; Cole, D. W. (1984) The impact of nitrification on soil acidification and cation leaching in red
18 alder ecosystem. J. Environ. Qual. 13: 586-590.
19 Van Miegroet, H.; Johnson, D. W.; Todd, D. E. (1993) Foliar response of red spruce saplings to fertilization with
20 Ca and Mg in the Great Smoky Mountains National Park. Can. J. For. Res. 23: 89-95.
21 Vasconcelos, L. A. de P.; Kahl, J. D. W.; Liu, D.; Macias, E. S.; White, W. H. (1996) Patterns of dust transport to
22 the Grand Canyon. Geophys. Res. Lett. 23: 3187-3190.
23 Viles, H. A. (1990) The early stages of building stone decay in an urban environment. Atmos. Environ.
24 24A: 229-232.
25 Vitousek, P. M.; Mooney, H. A.; Lubchenco, J.; Melillo, J. M. (1997) Human domination of Earth's ecosystems.
26 Science (Washington, DC) 277: 494-499.
27 Vogt, K. A.; Dahlgren, R.; Ugolini, F.; Zabowski, D.; Moore, E. E.; Zasoski, R. (1987a) Aluminum, Fe, Ca, Mg, K,
28 Mn, Cu, Zn and P in above- and belowground biomass. I. Abies amabilis and Tsuga mertensiana.
29 Biogeochemistry 4: 277-294.
30 Vogt, K. A.; Dahlgren, R.; Ugolini, F.; Zabowski, D.; Moore, E. E.; Zasoski, R. (1987b) Aluminum, Fe, Ca, Mg, K,
31 Mn, Cu, Zn and P in above- and belowground biomass. II. Pools, and circulation in a subalpine Abies
32 amabilis stand. Biogeochemistry 4: 295-311.
33 Vose, J. M.; Swank, W. T. (1990) Preliminary estimates of foliar absorption of 15N-labeled nitric acid vapor
34 (HNO3) by mature eastern white pine (Pinus strobus). Can. J. For. Res. 20: 857-860.
35 Waggoner, A. P.; Weiss, R. E. (1980) Comparison of fine particle mass concentration and light scattering extinction
36 in ambient aerosol. Atmos. Environ. 14: 623-626.
37 Waggoner, A. P.; Weiss, R. E.; Ahlquist, N. C.; Covert, D. S.; Will, S.; Charlson, R. J. (1981) Optical
38 characteristics of atmospheric aerosols. In: White, W. H.; Moore, D. J.; Lodge, J. P., Jr., eds. Plumes and
39 visibility: measurements and model components: proceedings of the symposium; November 1980;
40 Grand Canyon National Park, AZ. Atmos. Environ. 15: 1891-1909.
41 Wagrowski, D. M.; Hites, R. A. (1997) Polycyclic aromatic hydrocarbon accumulation in urban, suburban, and rural
42 vegetation. Environ. Sci. Technol. 31: 279-282.
43 Wall, D. H. (1999) Biodiversity and ecosystem functioning. BioScience 49: 107-108.
44 Wall, D. H.; Moore, J. C. (1999) Interactions underground: soil biodiversity, mutualism, and ecosystem processes.
45 Bioscience 49: 109-117.
46 Walton, J. R.; Johnson, J. B.; Wood, G. C. (1982) Atmospheric corrosion initiation by sulphur dioxide and
47 paniculate matter: II. characterisation and corrosivity of individual paniculate atmospheric pollutants.
48 Br. Corros. J. 17: 65-70.
49 Wang, W. C.; Yung, Y. L.; Lacis, A. A.; Mo, T.; Hansen, J. E. (1976) Greenhouse effects due to man-made
50 perturbations of trace gases. Science (Washington, DC) 194: 685-690.
51 Waring, R. H. (1987) Nitrate pollution: a particular danger to boreal and subalpine coniferous forests. In:
52 Fujimori, T.; Kimura, M., eds. Human impacts and management of mountain forests: [proceedings of a
53 symposium]. Ibaraki, Japan: Forestry and Forest Products Research Institute; pp. 93-105.
54 Waring, R. H.; Schlesinger, W. H. (1985) The carbon balance of trees. In: Forest ecosystems: concepts and
55 management. Orlando, FL: Academic Press, Inc.; pp. 7-37.
March 2001 4-186 DRAFT-DO NOT QUOTE OR CITE
-------
1 Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ.
2 19:241-266.
3 Watson, J. G.; Chow, J. C.; Richards, L. W.; Haase, D. L.; McDade, C.; Dietrich, D. L.; Moon, D.; Sloane, C.
4 (1991) The 1989-90 Phoenix urban haze study, volume II: the apportionment of light extinction to sources
5 [final report]. Reno, NV: University and Community College System of Nevada, Desert Research Institute;
6 DRI document no. 8931.5F1.
7 Webb, A. H.; Bawden, R. J.; Busby, A. K.; Hopkins, J. N. (1992) Studies on the effects of air pollution on
8 limestone degradation in Great Britain. Atmos. Environ. Part B 26: 165-181.
9 Wedin, D. A.; Tilman, D. (1996) Influence of nitrogen loading and species composition on the carbon balance of
10 grasslands. Science 274: 1720-1723.
11 Weinbaum, S. A.; Neumann, P. M. (1977) Uptake and metabolism of 15N-labeled potassium nitrate by French
12 prune (Primus domestica L.) leaves and the effects of two surfactants. J. Am. Soc. Hortic. Sci. 102: 601-604.
13 Wellburn, A. R. (1990) Tansley review no. 24: Why are atmospheric oxides of nitrogen usually phytotoxic and not
14 alternative fertilizers? New Phytol. 115: 395-429.
15 Wenny, B. N.; Schafer, J. S.; DeLuisi, J. J.; Saxena, V. K.; Barnard, W. F.; Petropavlovskikh, I. V.; Vergamini,
16 A. J. (1998) A study of regional aerosol radiative properties and effects on ultraviolet-B radiation.
17 J. Geophys. Res. [Atmos.] 103: 17,083-17,097.
18 Wesselink, L. G.; Meiwes, K.-J.; Matzner, E.; Stein, A. (1995) Long-term changes in water and soil chemistry in
19 spruce and beech forests, Soiling, Germany. Environ. Sci. Technol. 29: 51-58.
20 Westman, W. E. (1977) How much are nature's services worth? Measuring the social benefits of ecosystem
21 functioning is both controversial and illuminating. Science (Washington, DC) 197: 960-964.
22 White, W. H. (1976) Reduction of visibility by sulphates in photochemical smog. Nature (London) 264: 735-736.
23 White, W. H. (1997) Deteriorating air or improving measurements? On interpreting concatenate time series.
24 J. Geophys. Res. [Atmos.] 102: 6813-6821.
25 White, W. H.; Macias, E. S. (1990) Light scattering by haze and dust at Spirit Mountain, Nevada. In: Mathai, C. V.,
26 ed. Visibility and fine particles: an A&WMA/EPA international specialty conference; October 1989; Estes
27 Park, CO. Pittsburgh, PA; Air & Waste Management Association; pp. 914-922. (A&WMA transactions
28 series no. TR-17).
29 White, W. H.; Seigneur, C.; Heinold, D. W.; Eltgroth, M. W.; Richards, L. W.; Roberts, P. T.; Bhardwaja, P. S.;
30 Conner, W. D.; Wilson, W. E., Jr. (1985) Predicting the visibility of chimney plumes: an intercomparison of
31 four models with observations at a well-controlled power plant. Atmos. Environ. 19: 515-528.
32 White, W. H.; Macias, E. S.; Nininger, R. C.; Schorran, D. (1994) Size-resolved measurements of light scattering by
33 ambient particles in the southwestern U.S. A. Atmos. Environ. 28: 909-921.
34 White, W. H.; Macias, E. S.; Vasconcelos, L. A. de P.; Farber, R. J.; Mirabella, V. A.; Green, M. C.; Pitchford,
35 M. L. (1999) Tracking regional background in a haze attribution experiment. J. Air Waste Manage. Assoc.
36 49:599-602.
37 Williams, M. W.; Baron, J. S.; Caine, N.; Sommerfeld, R.; Sanford, R. (1996) Nitrogen saturation in the
38 Rocky Mountains. Environ. Sci. Technol. 30: 640-646.
39 Wilson, E. O. (1997) The creation of biodiversity. In: Raven, P. H., ed. Nature and human society: the quest for a
40 sustainable world: proceedings of the 1997 forum on biodiversity; October; sponsored by the National
41 Research Council. Washington, DC: National Academy Press.
42 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
43 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
44 Winner, W. E.; Atkinson, C. J. (1986) Absorption of air pollution by plants, and consequences for growth.
45 Trends Ecol. Evol. 1: 15-18.
46 Winner, W. E.; Bewley, J. D. (1978a) Terrestrial mosses as bioindicators of SO2 pollution stress: synecological
47 analysis and the index of atmospheric purity. Oecologia 35: 221-230.
48 Winner, W. E.; Bewley, J. D. (1978b) Contrasts between bryophyte and vascular plant synecological responses in
49 an SO2-stressed white spruce association in central Alberta. Oecologia 33:311-325.
50 Wittenburg, C.; Dannecker, W. (1992) Dry deposition and deposition velocity of airborne acidic species upon
51 different sandstones. In: Proceedings of the 1992 European aerosol conference; September; Oxford,
52 United Kingdom. J. Aerosol Sci. 23(suppl. 1): S869-S872.
53 Wolff, G. T.; Collins, D. C.; Rodgers, W. R.; Verma, M. H.; Wong, C. A. (1990) Spotting of automotive finishes
54 from the interactions between dry deposition of crustal material and wet deposition of sulfate. J. Air Waste
55 Manage. Assoc. 40: 1638-1648.
March 2001 4-187 DRAFT-DO NOT QUOTE OR CITE
-------
1 World Health Organization. (1997) Nitrogen oxides. 2nd ed. Geneva, Switzerland: World Health Organization.
2 (Environmental health criteria 188).
3 World Meteorological Organization. (1988) Developing policies for responding to climatic change: a summary of
4 the discussions and recommendations of workshops; September-October 1987; Villach, Austria; and
5 November 1987; Bellagio, Austria. Geneva, Switzerland: World Meteorological Organization; report no.
6 WMO/TD; no. 225. [World Climate Impact Programme series report no. WCIP-1].
7 World Meteorological Organization. (1999) Scientific assessment of ozone depletion: 1998. Geneva, Switzerland:
8 World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
9 World Resources Institute. (2000) World resources 2000-2001: people and ecosystems: the fraying web of life.
10 Washington, DC: World Resources Institute.
11 Wu, P.-M.; Okada, K. (1994) Nature of coarse nitrate particles in the atmosphere~a single particle approach.
12 Atmos. Environ. 28: 2053-2060.
13 Wu, X. A.; Seigneur, C.; Bergstrom, R. W. (1996) Evaluation of a sectional representation of size distributions for
14 calculating aerosol optical properties. J. Geophys. Res. [Atmos.] 101: 19,277-19,283.
15 Xu, J. R.; Balik, C. M. (1989) Infrared attenuated total reflectance study of latex paint films exposed to aqueous
16 sulfur dioxide. J. Appl. Polym. Sci. 38: 173-183.
17 Yanai, R. D.; Siccama, T. G.; Arthur, M.. A.; Federer, C. A.; Friedland, A. J. (1999) Accumulation and depletion
18 of base cations in forest floors in the northeastern United States. Ecology 80: 2774-2787.
19 Yerrapragada, S. S.; Jaynes, J. H.; Chirra, S. R.; Gauri, K. L. (1994) Rate of weathering of marble due to dry
20 deposition of ambient sulfur and nitrogen dioxides. Anal. Chem. 66: 655-659.
21 Yocom, J. E.; Grappone, N. (1976) Effects of power plant emissions on materials. Palo Alto, CA: Electric Power
22 Research Institute; report no. EPRI/EC-139. Available from: NTIS, Springfield, VA; PB-257 539.
23 Yocom, J. E.; Upham, J. B. (1977) Effects of economic materials and structures. In: Stern, A. C., ed. Air pollution.
24 3rd ed. Volume II. New York, NY: Academic Press; pp. 93-94.
25 Young, P. (1996) Pollution-fueled "biodeterioration" threatens historic stone: researchers are probing the role
26 microorganisms play in the decay of buildings and artwork. Environ. Sci. Technol. 30: 206A-208A.
27 Young, J. R.; Ellis, E. C.; Hidy, G. M. (1988) Deposition of air-borne acidifiers in the western environment.
28 J. Environ. Qual. 17: 1-26.
29 Youngs, W. D.; Rutzke, M.; Gutenmann, W. H.; Lisk, D. J. (1993) Nickel and vanadium in foliage in the vicinity of
30 an oil-fired power plant. Chemosphere 27: 1269-1272.
31 Zannetti, P.; Tombach, L; Balson, W. (1990) Calculation of visual range improvements from SO2 emission
32 controls--!, semi-empirical methodology. Atmos. Environ. Part A 24: 2361-2368.
33 Zannetti, P.; Tombach, L; Cvencek, S.; Balson, W. (1993) Calculation of visual range improvements from SO2
34 emission controls--!!, an application to the eastern United States. Atmos. Environ. Part A 27: 1479-1490.
35 Zappia, G.; Sabbioni, C.; Gobbi, G. (1993) Non-carbonate carbon content on black and white areas of damaged
36 stone monuments. Atmos. Environ. 27A: 1117-1121.
37 Zappia, G.; Sabbioni, C.; Pauri, M. G.; Gobbi, G. (1994) Mortar damage due to airborne sulfur compounds in a
38 simulation chamber. Mater. Struct. 27: 469-473.
39 Zappia, G.; Sabbioni, C.; Riontino, C.; Gobbi, G.; Favoni, O. (1998) Exposure tests of building materials in urban
40 atmosphere. Sci. Total Environ. 224: 235-244.
41 Zepp, R. G.; Callaghan, T. V.; Erickson, D. J. (1998) Effects of enhanced solar ultraviolet radiation on
42 biogeochemical cycles. J. Photochem. Photobiol. B 46: 69-82.
43 Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
44 submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
45 27: 1593-1607.
46 Zhang, X.; Turpin, B. J.; McMurry, P. H.; Hering, S. V.; Stolzenburg, M. R. (1994) Mie theory evaluation of
47 species contributions to 1990 wintertime visibility reduction in the Grand Canyon. Air Waste 44: 153-162.
48
March 2001 4-188 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 4A
2
3 Excerpted Key Points from the Executive Summary of the
4 World Meteorological Organization 1998
5 Assessment of Stratospheric Ozone Depletion
6 (World Meteorological Organization, 1999)
7
8 Among the provisions of the 1987 Montreal Protocol on Substances that Deplete the Ozone
9 Layer was the requirement that the Parties to the Protocol base their future decisions on available
10 scientific, environmental, technical, and economic information, as assessed by worldwide expert
11 communities. Advances in the understanding of ozone science over this decade were assessed in
12 1988, 1989, 1991, and 1994. This information was input to the subsequent Amendments and
13 Adjustments of the 1987 Protocol. The 1998 assessment summarized below is the fifth in that
14 series.
15
16 Recent Major Scientific Findings and Observations
17 Since the Scientific Assessment of Ozone Depletion: 1994, significant advances have
18 continued to be made in understanding of the impact of human activities on the ozone layer, the
19 influence of changes in chemical composition on the radiative balance of the Earth's climate, and,
20 indeed, the coupling of the ozone layer and the climate system. Numerous laboratory
21 investigations, atmospheric observations, and theoretical and modeling studies have produced
22 several key ozone- and climate-related findings that are discussed below.
23
24 • The total combined abundance of ozone-depleting compounds in the lower atmosphere
25 peaked in about 1994 and now is slowly declining. Total chlorine is declining, but total
26 bromine is still increasing. As forecast in the 1994 Assessment, the long period of increasing
27 total chlorine abundances—primarily from the chlorofluorocarbons (CFCs), carbon
28 tetrachloride (CC14), and methyl chloroform (CH3CCI3)—has ended. The peak total
29 tropospheric chlorine abundance was 3.7 ± 0.1 parts per billion (ppb) between mid-1992 and
30 mid-1994. The declining abundance of total chlorine results principally from reduced
31 emissions of methyl chloroform. Chlorine from the major CFCs is still increasing slightly. The
March 2001 4A-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 abundances of most of the halons continue to increase (for example, Hal on-1211, almost 6%
2 per year in 1996), but the rate has slowed in recent years. These halon increases likely are
3 caused by emissions in the 1990s from the halon "bank," largely in developed countries, and
4 new production of halons in developing countries. The observed abundances of CFCs and
5 chlorocarbons in the lower atmosphere are consistent with reported emissions.
6 • The observed abundances of the substitutes for the CFCs are increasing Abundances of
7 the hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HFCs) are increasing as a
8 result of continuation of earlier uses and of their use as substitutes for the CFCs. In 1996, the
9 HCFCs contributed about 5% to the tropospheric chlorine from the long-lived gases. This
10 addition from the substitutes offsets some of the decline in tropospheric chlorine associated
11 with methyl chloroform, but is still about 10 times less than that from the total tropospheric
12 chlorine growth rate throughout the 1980s. The atmospheric abundances of HCFC-141b and
13 HCFC-142b calculated from reported emissions data are factors of 1.3 and 2, respectively,
14 smaller than observations. Observed and calculated abundances agree for HCFC-22 and
15 HFC-134a.
16 • The combined abundance of stratospheric chlorine and bromine is expected to peak
17 before the year 2000. The delay in this peak in the stratosphere compared with the lower
18 atmosphere reflects the average time required for surface emissions to reach the lower
19 stratosphere. The observations of key chlorine compounds in the stratosphere up through the
20 present show the expected slower rate of increase and show that the peak had not occurred at
21 the time of the most recent observations that were analyzed for this assessment.
22 • The role of methyl bromide as an ozone-depleting compound is now considered to be less
23 than was estimated in the 1994 Assessment, although significant uncertainties remain.
24 The current best estimate of the Ozone Depletion Potential (ODP) for methyl bromide (CH3Br)
25 is 0.4, compared with an ODP of 0.6 estimated previously. The change is caused primarily by
26 both an increase in estimates of ocean removal processes and identification of an uptake by
27 soils, with a smaller contribution from change in our estimate of the atmospheric removal rate.
28 Recent research has shown that the science of atmospheric methyl bromide is complex and still
29 not well understood. Current understanding of the sources and sinks of atmospheric methyl
30 bromide is incomplete.
31
March 2001 4A-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 • The rate of decline in stratospheric ozone at mid-latitudes has slowed; hence, the
2 projections of ozone loss made in the 1994 assessment are larger than what has actually
3 occurred. Total column ozone decreased significantly at mid-latitudes (25 to 60°) between
4 1979 and 1991, with estimated linear downward trends of 4.0, 1.8, and 3.8% per decade,
5 respectively, for northern mid-latitudes in winter/spring, northern mid-latitudes in summer/fall,
6 and southern mid-latitudes year round. However, since 1991, the linear trend observed during
7 the 1980s has not continued, rather, total column ozone has been almost constant at mid-
8 latitudes in both hemispheres since the recovery from the 1991 Mt. Pinatubo eruption. The
9 observed total column ozone losses from 1979 to the period 1994 to 1997 are about 5.4, 2.8,
10 and 5.0%, respectively, for northern mid-latitudes in winter/spring, northern mid-latitudes in
11 summer/fall, and southern mid-latitudes year round, rather than the values projected in the 1994
12 assessment assuming a linear trend: 7.6, 3.4, and 7.2%, respectively. Understanding of how
13 changes in stratospheric chlorine/bromine and aerosol loading affect ozone suggests some of
14 the reasons for the unsuitability of using a linear extrapolation of the pre-1991 ozone trend to
15 the present.
16 • The springtime Antarctic ozone hole continues unabated. The extent of ozone depletion has
17 remained essentially unchanged since the early 1990s. This behavior is expected given the
18 near-complete destruction of ozone within the Antarctic lower stratosphere during springtime.
19 The factors contributing to the continuing depletion are well understood.
20 • The link between the long-term buildup of chlorine and the decline of ozone in the upper
21 stratosphere has been firmly established. Model predictions based on the observed buildup
22 of stratospheric chlorine in the upper stratosphere indicate a depletion of ozone that is in good
23 quantitative agreement with the altitude and latitude dependence of the measured ozone decline
24 during the past several decades, which peaks at about 7% per decade near 40 km at mid-
25 latitudes in both hemispheres.
26 • The late-winter/spring ozone values in the Arctic were unusually low in six out of the last
27 nine years, the six being years that are characterized by unusually cold and protracted
28 stratospheric winters. The possibility of such depletions was predicted in the 1989
29 assessment. Minimum Arctic vortex temperatures are near the threshold for large chlorine
30 activation. Therefore, the year-to-year variability in temperature, which is driven by
31 meteorology, leads to particularly large variability in ozone for current chlorine loading. As a
March 2001 4A-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 result, it is not possible to forecast the behavior of Arctic ozone for a particular year. Elevated
2 stratospheric halogen abundances over the next decade or so imply that the Arctic will continue
3 to be vulnerable to large ozone losses.
4 • The understanding of the relation between increasing surface UV-B radiation and
5 decreasing column ozone has been further strengthened by ground-based observations,
6 and newly developed satellite methods show promise for establishing global trends in UV
7 radiation. The inverse dependence of surface UV radiation and the overhead amount of ozone,
8 which was demonstrated in earlier assessments, has been further demonstrated and quantified
9 by ground-based measurements under a wide range of atmospheric conditions. In addition, the
10 influences of other variables, such as clouds, particles, and surface reflectivity, are better
11 understood. These data have assisted the development of a satellite-based method to estimate
12 global UV changes, taking into account the role of cloud cover. The satellite estimates for 1979
13 through 1992 indicate that the largest UV increases occur during spring at high latitudes in both
14 hemispheres.
15 • Stratospheric ozone losses have caused a cooling of the global lower stratosphere and
16 global-average negative radiative forcing of the climate system. The decadal temperature
17 trends in the stratosphere have now been better quantified. Model simulations indicate that
18 much of the observed downward trend in lower stratospheric temperatures (about 0.6 °C per
19 decade from 1979 to 1994) is attributed to the ozone loss in the lower stratosphere. A lower
20 stratosphere that is cooler results in less infrared radiation reaching the surface/troposphere
21 system. Radiative calculations, using extrapolations based on the ozone trends reported in the
22 1994 assessment for reference, indicate that stratospheric ozone losses since 1980 may have
23 offset about 30% of the positive forcing because of increases in the well-mixed greenhouse
24 gases (i.e., carbon dioxide, methane, nitrous oxide, halocarbons) over the same time period. The
25 climatic impact of the slowing of mid-latitude ozone trends and the enhanced ozone loss in the
26 Arctic has not yet been assessed.
27 • Based on past emissions of ozone-depleting substances and a projection of the maximum
28 allowances under the Montreal Protocol into the future, the maximum ozone depletion is
29 estimated to lie within the current decade or the next two decades, but its identification
30 and the evidence for the recovery of the ozone layer lie still further ahead. The falloff of
31 total chlorine and bromine abundances in the stratosphere in the next century will be much
March 2001 4A-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 slower than the rate of increase observed in past decades, because of the slow rate at which
2 natural processes remove these compounds from the stratosphere. The most vulnerable period
3 for ozone depletion will be extended into the coming decades. However, extreme
4 perturbations, such as natural events like volcanic eruptions, could enhance the loss from
5 ozone-depleting chemicals. Detection of the beginning of the recovery of the ozone layer could
6 be achievable early in the next century if decreasing chlorine and bromine abundances were the
7 only factor. However, potential future increases or decreases in other gases important in ozone
8 chemistry (such as nitrous oxide, methane, and water vapor) and climate change will influence
9 the recovery of the ozone layer. When combined with the natural variability of the ozone layer,
10 these factors imply that unambiguous detection of the beginning of the recovery of the ozone
11 layer is expected to be well after the maximum stratospheric loading of ozone-depleting gases.
12
13
14 REFERENCE
15 World Meteorological Organization (WMO). (1999) Scientific assessment of ozone depletion: 1998. Geneva,
16 Switzerland: World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
17
March 2001 4A-5 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 4B
2
3 Excerpted Key Points from the Executive Summary of the United Nations
4 Environmental Programme 1998 Assessment of Environmental Effects of
5 Ozone Depletion (United Nations Environmental Programme, 1998)
6
7 Decreased quantities of total-column ozone now are observed over large parts of the globe,
8 permitting increased penetration of solar UV-B radiation (280 to 315 nm) to the Earth's surface.
9 The present assessment deals with the possible consequences. The Atmospheric Science Panel
10 predicts that the ozone layer will be in its most vulnerable state during the coming two decades.
11 Some of the effects are expected to occur during most of the next century. Recent studies show
12 that the effects of ozone depletion would have been dramatically worse without protective
13 measures taken under the 1987 Montreal Protocol. The assessment is given in seven papers,
14 summarized below:
15
16 (1) Changes in Ultraviolet Radiation
17 • Stratospheric ozone levels are near their lowest points since measurements began, so
18 current UV-B radiation levels are thought to be close to their maximum. Total
19 stratospheric content of ozone-depleting substances is expected to reach a maximum before the
20 year 2000. All other things being equal, the current ozone losses and related UV-B increases
21 should be close to their maximum. Increases in surface erythemal (sunburning) UV radiation
22 relative to the values in the 1970s are estimated to be
23 • about 7% at Northern Hemisphere mid-latitudes in winter/spring;
24 • about 4% at Northern Hemisphere mid-latitudes in summer/fall;
25 • about 6% at Southern Hemisphere mid-latitudes on a year-round basis;
26 • about 130% in the Antarctic in the spring; and
27 • about 22% in the Arctic in the spring.
28 • The correlation between increases in surface UV-B radiation and decreases in overhead
29 ozone has been demonstrated further and quantified by ground-based instruments under
30 a wide range of conditions. Improved measurements of UV-B radiation are now providing
March 2001 4B-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 better geographical and temporal coverage. Surface UV-B radiation levels are highly variable
2 because of sun angle, cloud cover, and, also, because of local effects, including pollutants and
3 surface reflections. With a few exceptions, the direct detection of UV-B trends at low- and
4 mid-latitudes remains problematic because of this high natural variability, the relatively small
5 ozone changes, and the practical difficulties of maintaining long-term stability in networks of
6 UV-measuring instruments. Few reliable UV-B radiation measurements are available from
7 pre-ozone-depletion days.
8 • Satellite-based observations of atmospheric ozone and clouds are being used, together
9 with models of atmospheric transmission, to provide global coverage and long-term
10 estimates of surface UV-B radiation. Estimates of long-term (1979 to 1992) trends in zonally
11 averaged UV irradiances that include cloud effects are nearly identical to those for clear-sky
12 estimates, providing evidence that clouds have not influenced the UV-B trends. However, the
13 limitations of satellite-derived UV estimates should be recognized. To assess uncertainties
14 inherent in this approach, additional validations involving comparisons with ground-based
15 observations are required.
16 • Direct comparisons of ground-based UV-B radiation measurements between a few
17 mid-latitude sites in the Northern and Southern Hemispheres have shown larger
18 differences than those estimated using satellite data. Ground-based measurements show that
19 summertime erythemal UV irradiances in the Southern Hemisphere exceed those at comparable
20 latitudes of the Northern Hemisphere by up to 40%, whereas corresponding satellite-based
21 estimates yield only 10 to 15% differences. Atmospheric pollution may be a factor in this
22 discrepancy between ground-based measurements and satellite-derived estimates. UV-B
23 measurements at more sites are required to determine whether the larger observed differences
24 are globally representative.
25 • High levels of UV-B radiation continue to be observed in Antarctica during the recurrent
26 spring-time ozone hole. For example, during ozone-hole episodes, measured biologically
27 damaging radiation at Palmer Station, Antarctica (64 ° S) has been found to approach and
28 occasionally even exceed maximum summer values at San Diego, CA (32 °N).
29 • Long-term predictions of future UV-B levels are difficult and uncertain. Nevertheless,
30 current best estimates suggest that a slow recovery to pre-ozone-depletion levels may be
31 expected during the next half-century. Although the maximum ozone depletion, and hence
March 2001 4B-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 maximum UV-B increase, is likely to occur in the current decade, the ozone layer will continue
2 to be in its most vulnerable state into the next century. The peak depletion and the recovery
3 phase could be delayed by decades because of interactions with other long-term atmospheric
4 changes (e.g., increasing concentrations of greenhouse gases). Other factors that could influence
5 the recovery include nonratification or noncompliance with the Montreal Protocol and its
6 Amendments and Adjustments and future volcanic eruptions. The recovery phase for surface
7 UV-B irradiances probably will not be detectable until many years after the ozone minimum.
8
9 (2) Effects on Human and Animal Health
10 • Recent estimates suggest that the increase in the risk of cataract and skin cancer because
11 of ozone depletion would not have been controlled adequately by implementation of the
12 Montreal Protocol (1987) alone, but can be achieved through implementation of its later
13 provisions. Risk assessments for the United States and northwestern Europe indicate large
14 increases in cataracts and skin cancers under either the "no Protocol" or the early Montreal
15 Protocol scenarios. Under scenarios based on later amendments (Copenhagen, 1992) and
16 Montreal (1997), increases in cataracts and skin cancer attributable to ozone depletion return
17 almost to zero by the end of the next century.
18 • The increases in UV-B radiation associated with ozone depletion are likely to lead to
19 increases in the incidence or severity of a variety of short- and long-term health effects, if
20 current exposure practices are not modified by changes in behavior.
21 • Adverse effects on the eye will affect all populations irrespective of skin color. Adverse
22 impacts could include more cases of acute reactions such as "snowblindness", increases in
23 cataract incidence or severity (and thus the incidence of cataract-associated blindness), and
24 increases in the incidence (and mortality) from ocular melanoma and squamous cell carcinoma
25 of the eye.
26 • Effects on the immune system also will affect all populations but may be both adverse and
27 beneficial. Adverse effects include depressed resistance to certain tumors and infectious
28 diseases, potential impairment of vaccination responses, and possibly increased severity of
29 some autoimmune and allergic responses. Beneficial effects could include decreases in the
30 severity of certain immunologic disease conditions, such as psoriasis and nickel allergy.
March 2001 4B-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Effects on the skin could include increases in photoaging and skin cancer with risk
2 increasing with fairness of skin. Increases in UV-B are likely to accelerate the rate of
3 photoaging, as well as increase the incidence (and associated mortality) of melanoma and
4 nonmelanoma skin cancer, basal cell carcinoma, and squamous cell carcinoma.
5 • Research is generating much new information being used to help reduce uncertainties
6 associated with current risk estimates. Evaluation of the impact of susceptibility genes is
7 helping to identify highly susceptible populations so that their special risk can be assessed.
8 Examination of the impacts of behavior changes, such as consuming diets that are high in
9 antioxidants, avoiding sun exposure during the 4 h around solar noon, and wearing of covering
10 apparel (e.g., hats, sunglasses), is beginning to identify important exposure patterns, as well as
11 possible mitigation strategies.
12 • Quantitative risk assessments for a variety of other effects, such as UV-B-induced
13 immunosuppression of infectious diseases, are not yet possible. New information continues
14 to confirm the reasonableness of these concerns, but data that is adequate for quantitative risk
15 assessment are not yet available.
16
17 (3) Effects on Terrestrial Ecosystems
18 • Increased UV-B can be damaging for terrestrial organisms including plants and
19 microbes, but all these organisms also have protective and repair processes. The balance
20 between damage and protection varies among species and even varieties of crop species; many
21 species and varieties can accommodate increased UV-B. Tolerance of elevated UV-B by some
22 species and crop varieties provides opportunities for genetic engineering and breeding to deal
23 with potential crop-yield reductions because of elevated UV-B in agricultural systems.
24 • Research in the past few years indicates that increased UV-B exerts effects more often
25 through altered patterns of gene activity rather than damage. These UV-B effects on
26 regulation manifest themselves in many ways including changes in life-cycle timing, changes in
27 plant form, and production of plant chemicals not directly involved in primary metabolism.
28 These plant chemicals play a role in protecting plants from pathogens and insect attack and
29 affect food quality for humans and grauine animals.
30 • Terrestrial ecosystem responses to increased UV-B are evident primarily in interactions
31 among species, rather than in the performance of individual species. Much of the recent
March 2001 4B-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 experimentation indicates that increased UV-B affects the balance of competition among higher
2 plants, the degree to which higher plants are consumed by insects, and susceptibility of plants to
3 pathogens. These effects can be mediated, in large part, by changes in plant form and
4 chemistry, but effects of UV-B on insects and microbes are also possible. The direction of
5 these UV-B mediated interactions among species is often difficult to predict based only on
6 single-organism responses to increased UV-B.
7 • Effects of increased UV-B radiation can accumulate from year to year in long-lived
8 perennial plants and from generation to generation in annual plants. This effect has been
9 shown in a few recent studies, but the generality of this accumulation among species is not
10 presently known. If this phenomenon is widespread, this would amplify otherwise subtle
11 responses to UV-B seen in a single growing season, for example, in forest trees.
12 • Effects of increased UV-B must be taken into account together with other environmental
13 factors including those associated with global change. Responses of plants and other
14 organisms to increased UV-B are modified by other environmental factors (e.g., CO, water
15 stress, mineral nutrient availability, heavy metals, temperature). Many of these factors also are
16 changing as the global climate is altered.
17
18 (4) Effects on Aquatic Ecosystems
19 • Recent studies continue to demonstrate that solar UV-B and UV-A have adverse effects on
20 the growth, photosynthesis, protein and pigment content, and reproduction of
21 phytoplankton, thus affecting the food web. These studies have determined biological
22 weighting functions and exposure-response curves for phytoplankton and have developed new
23 models for the estimation of UV-related photoinhibition. In spite of this increased
24 understanding and enhanced ability to model aquatic impacts, considerable uncertainty remains
25 with respect to quantifying effects of ozone-related UV-B increases at the ecosystem level.
26 • Macroalgae and sea grasses show a pronounced sensitivity to solar UV-B. They are
27 important biomass producers in aquatic ecosystems. Most of these organisms are attached and
28 so cannot avoid being exposed to solar radiation at their growth site. Effects have been found
29 throughout the top 10 to 15 m of the water column.
30 • Zooplankton communities, as well as other aquatic organisms including sea urchins,
31 corals, and amphibians, are sensitive to UV-B. There is evidence that, for some of these
March 2001 4B-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 populations, even current levels of solar UV-B radiation, acting in conjunction with other
2 environmental stresses, may be a limiting factor, but quantitative evaluation of possible effects
3 remains uncertain.
4 • UV-B radiation is absorbed by and breaks down dissolved organic carbon (DOC) and
5 particulate organic carbon (POC) and makes the products available for bacterial
6 degradation and remineralization. The degradation products are of importance in the cycling
7 of carbon in aquatic ecosystems. Because UV-B breaks down DOC as it is absorbed, increase
8 in UV-B can increase the penetration of both UV-B and UV-A radiation into the water column.
9 As a consequence, the quantity of UV-B penetrating to a given depth both influences and is
10 influenced by DOC. Warming and acidification result in faster degradation of these substances
11 and, thus, enhance the penetration of UV radiation into the water column.
12 • Polar marine ecosystems, where ozone-related UV-B increases are the greatest, are
13 expected to be the oceanic ecosystems most influenced by ozone depletion. Oceanic
14 ecosystems are characterized by large spatial and temporal variabilities that make it difficult to
15 select out UV-B-specific effects on single species or whole phytoplankton communities.
16 Although estimates of reduction in both Arctic and Antarctic productivity are based on
17 measurable short-term effects, there remain considerable uncertainties in estimating long-term
18 consequences, including possible shifts in community structure. Reduced productivity offish
19 and other marine crops could have an economic impact, as well as affect natural predators;
20 however, quantitative estimation of the possible effects of reduced production remain
21 controversial.
22 • Potential consequences of enhanced levels of exposure of aquatic ecosystems to UV-B
23 radiation include reduced uptake capacity for atmospheric carbon dioxide (CO2),
24 resulting in the potential augmentation of global warming. The oceans play a key role with
25 respect to the budget of greenhouse gases. Marine phytoplankton are a major sink for
26 atmospheric CO2 and they have a decisive role in the development of future trends of CO2
27 concentrations in the atmosphere. The relative importance of the net uptake of CO2 by the
28 biological pump and the possible role of increased UV-B in the ocean are still controversial.
29
30
March 2001 4B-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 (5) Effects on Biogeochemical Cycles
2 • Effects of increased UV-B on emissions of CO2 and carbon monoxide (CO) and on
3 mineral nutrient cycling in the terrestrial biosphere have been confirmed by recent
4 studies of a range of species and ecosystems. The effects, both in magnitude and direction, of
5 UV-B on trace gas emissions and mineral nutrient cycling are species specific and operate on a
6 number of processes. These processes include changes in the chemical composition in living
7 plant tissue; photodegradation (breakdown by light) of dead plant matter, including litter;
8 release of CO from vegetation previously charred by fire; changes in the communities of
9 microbial decomposers; and effects on nitrogen-fixing micro-organisms and plants. Long-term
10 experiments are in place to examine UV-B effects on carbon capture and storage in biomass
11 within natural terrestrial ecosystems.
12 • Studies in natural aquatic ecosystems have indicated that organic matter is the primary
13 regulator of UV-B penetration. Enhanced UV-B can affect the balance between the
14 biological processes that produce the organic matter and the chemical and microbial processes
15 that degrade it. Changes in the balance have broad impacts on the effects of enhanced UV-B on
16 biogeochemical cycles. These changes, which are reinforced by changes in climate and
17 acidification, result from clarification of the water and changes in light quality.
18 • Increased UV-B has positive and negative impacts on microbial activity in aquatic
19 ecosystems that can affect carbon and mineral nutrient cycling, as well as the uptake and
20 release of greenhouse and chemically reactive gases. Photoinhibition of surface aquatic
21 micro-organisms by UV- B can be offset partially by photodegradation of dissolved organic
22 matter to produce substrates, such as organic acids and ammonium, that stimulate microbial
23 activity.
24 • Modeling and experimental approaches are being developed to predict and measure the
25 interactions and feedbacks between climate change in UV-B-induced changes in marine
26 and terrestrial biogeochemical cycles. These interactions include alterations in the oxidative
27 environment in the upper ocean and in the marine boundary layer and oceanic production and
28 release of CO, volatile organic compounds (VOC), and reactive oxygen species (ROS, such as
29 hydrogen peroxide and hydroxyl radicals). Climate-related changes in temperature and water
30 supply in terrestrial ecosystems interact with UV-B radiation through biogeochemical processes
31 operating on a wide range of time scales.
March 2001 4B-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 (6) Effects on Air Quality
2 • Increased UV-B will increase the chemical activity in the lower atmosphere (the
3 troposphere). Troposphere ozone levels are sensitive to local concentrations of nitrogen
4 oxides (NOx) and hydrocarbons. Model studies suggest that additional UV-B radiation reduces
5 tropospheric ozone in clean environments (low NOX) and increases tropospheric ozone in
6 polluted areas (high NOX).
7 • Assuming other factors remain constant, additional UV-B will increase the rate at which
8 primary pollutants are removed from the troposphere. Increased UV-B is expected to
9 increase the concentration of hydroxyl radicals (OH) and result in faster removal of pollutants.
10 Increased concentrations of oxidants such as hydrogen peroxide and organic peroxides also are
11 expected. The effects of UV-B increases on tropospheric ozone, OH, methane, CO, and
12 possibly other tropospheric constituents, although not negligible, will be difficult to detect
13 because the concentrations of these species also are influenced by many other variable factors
14 (e.g., emissions).
15 • No significant effects on humans or the environment have been identified from
16 trifluoroacetic acid (TFA) produced by atmospheric degradation of HCFCs and HFCs.
17 Numerous studies have shown that TFA has, at most, moderate short-term toxicity. Insufficient
18 information is available to assess potential chronic, developmental, or reproductive effects. The
19 atmospheric degradation mechanisms of most substitutes for ozone-depleting substances are
20 well established. HCFCs and HFCs are two important classes of substitutes. Atmospheric
21 degradation of HCFC-123 (CF3CHC12), HCFC-124 (CF3CHFCI), and H FC-I34a (CF3CH2F)
22 produces TFA. Reported measurements of TFA in rain, rivers, lakes, and oceans show it to be
23 an ubiquitous component of the hydrosphere, present at levels much higher than can be
24 explained by currently reported sources. The levels of TFA currently produced by the
25 atmospheric degradation of HFCs and HCFCs are estimated to be orders of magnitude below
26 those of concern and make only a minor contribution to the current environmental burden of
27 TFA.
28
29 (7) Effects on Materials
30 • Physical and mechanical properties of polymers are affected negatively by increased
31 UV-B in sunlight. Increased UV-B reduces the useful lifetimes of synthetic polymer products
March 2001 4B-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 used outdoors and of biopolymer materials such as wood, paper, wool, and cotton. The
2 reduction in service life of materials depends on the synergistic effect of increased UV-B and
3 other factors, especially the temperature of the material during exposure to sunlight. Even
4 under harsh UV exposure conditions, the higher temperatures largely determine the extent of
5 increased UV-induced damage to photostabilized polyethylenes. However, accurate assessment
6 of such damage to various materials is presently difficult to make because of limited availability
7 of technical data, especially on the relationship between the dose of UV-B radiation and the
8 resulting damage of the polymer or other material.
9 • Conventional photostabilizers are likely to be able to mitigate the effects of increased UV
10 levels in sunlight. More effective photostabilizers for plastics have been commercialized in
11 recent years. The use of these compounds allows plastic polymer products to be used in a wide
12 range of different UV environments found worldwide. It is reasonable to expect existing
13 photostabilizer technologies to be able to mitigate these effects of an increased UV-B on
14 polymer materials. This, however, would increase the cost of the relevant polymer products,
15 surface coatings, and treated biopolymer materials. However, the efficiencies of even the
16 conventional photostabilizers under the unique exposure environments resulting from an
17 increase in solar UV-B have not been well studied.
18
19
20 REFERENCE
21 United Nations Environment Programme (UNEP). (1998) Environmental effects of ozone depletion: 1998
22 assessment. J. Photochem. Photobiol. B 46: 1-4.
23
24
25
March 2001 4B-9 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 4C
2
3 Excerpted Key Points from the Executive Summary of the Special Report
4 of the International Panel on Climate Change Working Group II on the
5 Regional Impacts of Climate Change: An Assessment of Vulnerability
6
7
8 Excerpts from Executive Summary materials from a Special Report of the IPCC Working
9 Group II, The Regional Impacts of Climate Change: An Assessment of Vulnerability (IPCC,
10 1998) are incorporated below in this appendix to provide an overview of key points regarding the
11 vulnerability of aquatic and terrestrial ecosystems, water resources, agriculture, and human
12 habitability in North America to climate change.
13
14 Scope of the Assessment
15 The report was prepared at the request of the Conference of the Parties to the United
16 Nations Framework Convention on Climate Change (UNFCCC) and its subsidiary bodies
17 (specifically, the Subsidiary Body for Scientific and Technological Advice-SBSTA). The special
18 report provides, on a regional basis, a review of state-of-the-art information on the vulnerability
19 to potential changes in climate of ecological systems, socioeconomic sectors (agriculture,
20 fisheries, water resources, and human settlements), and human health. The report reviews the
21 sensitivity of these systems as well as options for adaptation. Though the report draws heavily on
22 the sectoral impact assessments of the Second Assessment Report (SAR), it also draws on more
23 recent peer-reviewed literature (inter alia, country studies programs).
24
25 Nature of the Issue
26 Human activities (primarily burning of fossil fuels and changes in land use and land cover)
27 are increasing atmospheric concentrations of greenhouse gases, which alter radiative balances
28 and tend to warm the atmosphere, and, in some regions, aerosols, which have an opposite effect
29 on radiative balances and tend to cool the atmosphere. At present, in some locations primarily in
30 the Northern Hemisphere, the cooling effects of aerosols can be large enough to more than offset
31 the warming caused by greenhouse gases. Because aerosols do not remain in the atmosphere for
March 2001 4C-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 long periods and global emissions of their precursors are not projected to increase substantially,
2 aerosols will not offset the global long-term effects of greenhouse gases, which are long-lived.
3 Aerosols can have important consequences for continental-scale patterns of climate change.
4 These changes in greenhouse gases and aerosols, taken together, are projected to lead to
5 regional and global changes in temperature, precipitation, and other climate variables, resulting
6 in global changes in soil moisture; an increase in global mean sea level; and prospects for more
7 severe extreme high-temperature events, floods, and droughts in some places. Based on the
8 range of sensitivities of climate to changes in the atmospheric concentrations of greenhouse gases
9 (IPCC, 1996; WG I) and plausible changes in emissions of greenhouse gases and aerosols
10 (lS92a-f, scenarios that assume no climate policies), climate models project that the mean annual
11 global surface temperature will increase by 1 to 3.5 °C by 2100, that global mean sea level will
12 rise by 15 to 95 cm, and that changes in the spatial and temporal patterns of precipitation will
13 occur. The average rate of warming probably would be greater than any seen in the past 10,000
14 years, although the actual annual to decadal rate would include considerable natural variability,
15 and regional changes could differ substantially from the global mean value. These long-term,
16 large-scale, human-induced changes will interact with natural variability on time scales of days to
17 decades (e.g., the El Nino-Southern Oscillation [ENSO] phenomenon) and, thus, influence social
18 and economic well-being. Possible local climate effects caused by unexpected events such as a
19 climate-change-induced change of flow pattern of marine water streams (e.g., the Gulf Stream)
20 have not been considered, because such changes cannot be predicted with confidence at present.
21 Scientific studies show that human health, ecological systems, and socioeconomic sectors
22 (e.g., hydrology and water resources, food and fiber production, coastal systems, human
23 settlements), all of which are vital to sustainable development, are sensitive to changes in
24 climate, including both the magnitude and rate of climate change, as well as to changes in
25 climate variability. Whereas many regions are likely to experience adverse effects of climate
26 change, some of which are potentially irreversible, some effects of climate change are likely to be
27 beneficial. Climate change represents an important additional stress on those systems already
28 affected by increasing resource demands, unsustainable management practices, and pollution,
29 which in many cases may be equal to or greater than those of climate change. These stresses will
30 interact in different ways across regions but can be expected to reduce the ability of some
31 environmental systems to provide, on a sustained basis, key goods and services needed for
March 2001 4C-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 successful economic and social development, including adequate food, clean air, and water;
2 energy; safe shelter; low levels of disease; and employment opportunities. Climate change also
3 will take place in the context of economic development, which may make some groups or
4 countries less vulnerable to climate change, for example, by increasing the resources available for
5 adaptation. Those that experience low rates of growth, rapid increases in population, and
6 ecological degradation may become increasingly vulnerable to potential changes.
7
8 Approach of the Assessment
9 The report assesses the vulnerability of natural and social systems of major regions of the
10 world to climate change. Vulnerability is defined as the extent to which a natural or social
11 system is susceptible to sustaining damage from climate change. Vulnerability is a function of
12 the sensitivity of a system to changes in climate (the degree to which a system will respond to a
13 given change in climate, including both beneficial and harmful effects) and the ability to adapt
14 the system to changes in climate (the degree to which adjustments in practices, processes, or
15 structures can moderate or offset the potential for damage or take advantage of opportunities
16 created because of a given change in climate). Under this framework, a highly vulnerable system
17 would be one that is highly sensitive to modest changes in climate, where the sensitivity includes
18 the potential for substantial harmful effects, and one for which the ability to adapt is severely
19 constrained.
20 Because the available studies have not employed a common set of climate scenarios and
21 methods, and because of uncertainties regarding the sensitivities and adaptability of natural and
22 social systems, the assessment of regional vulnerabilities is necessarily qualitative. However, the
23 report provides substantial and indispensable information on what currently is known about
24 vulnerability to climate change.
25 In a number of instances, quantitative estimates of impacts of climate change are cited in
26 the report. Such estimates are dependent on the specific assumptions employed regarding future
27 changes in climate, as well as on the particular methods and models applied in the analyses.
28 In interpreting these estimates, it is important to bear in mind that uncertainties regarding the
29 character, magnitude, and rates of future climate change remain. These uncertainties impose
30 limitations on the ability of scientists to project impacts of climate change, particularly at
31 regional and smaller scales.
March 2001 4C-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 It is in part because of the uncertainties regarding how climate will change that the report
2 takes the approach of assessing vulnerabilities rather than assessing quantitatively the expected
3 impacts of climate change. The estimates are interpreted best as illustrative of the potential
4 character and approximate magnitudes of impacts that may result from specific scenarios of
5 climate change. They serve as indicators of sensitivities and possible vulnerabilities. Most
6 commonly, the estimates are based on changes in equilibrium climate that have been simulated to
7 result from an equivalent doubling of carbon dioxide (CO2) in the atmosphere. Usually, the
8 simulations have excluded the effects of aerosols. Increases in global mean temperatures
9 corresponding to these scenarios mostly fall in the range of 2 to 5 °C. To provide a temporal
10 context for these scenarios, the range of projected global mean warming by 2100 is 1 to 3.5 °C,
11 accompanied by a mean sea-level rise of 15 to 95 cm, according to the IPCC Second Assessment
12 Report. General circulation model (GCM) results are used in this analysis to justify the order of
13 magnitude of the changes used in the sensitivity analyses. They are not predictions that climate
14 will change by specific magnitudes in particular countries or regions. The amount of literature
15 available for assessment varies in quantity and quality among the regions.
16
17 Overview of Regional Vulnerabilities to Global Climate Change
18 The report's assessment of regional vulnerability to climate change focuses on ecosystems,
19 hydrology and water resources, food and fiber production, coastal systems, human settlements,
20 human health, and other sectors or systems (including the climate system) important to
21 10 regions that encompass the Earth's land surface. Wide variation in the vulnerability of similar
22 sectors or systems is to be expected across regions, as a consequence of regional differences in
23 local environmental conditions; preexisting stresses to ecosystems; current resource-use patterns;
24 and the framework of factors affecting decision making, including government policies, prices,
25 preferences, and values. Nonetheless, some general observations, based on information
26 contained in the IPCC Second Assessment Report (SAR) (IPCC, 1995) and synthesized from the
27 regional analyses in the 1998 assessment, provide a global context for assessment of each
28 region's vulnerability. The general types of vulnerabilities are discussed first below, followed by
29 more specific discussion of projected likely regional vulnerabilities most directly applicable to
30 the United States.
31
March 2001 4C-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ecosystems
2 Ecosystems are of fundamental importance to environmental function and to sustainability,
3 and they provide many goods and services critical to individuals and societies. These goods and
4 services include (1) providing food, fiber, fodder, shelter, medicines, and energy; (2) processing
5 and storing carbon and nutrients; (3) assimilating wastes; (4) purifying water, regulating water
6 runoff, and moderating floods; (5) building soils and reducing soil degradation; (6) providing
7 opportunities for recreation and tourism; and (7) housing the Earth's entire reservoir of genetic
8 and species diversity. In addition, natural ecosystems have cultural, religious, aesthetic, and
9 intrinsic existence values. Changes in climate have the potential to affect the geographic location
10 of ecological systems, the mix of species that they contain, and their ability to provide the wide
11 range of benefits on which societies rely for their continued existence. Ecological systems are
12 intrinsically dynamic and are constantly influenced by climate variability. The primary influence
13 of anthropogenic climate change on ecosystems is expected to be through the rate and magnitude
14 of change in climate means and extremes; climate change is expected to occur at a rapid rate
15 relative to the speed at which ecosystems can adapt and reestablish themselves; and through the
16 direct effects of increased atmospheric CO2 concentrations, which may increase the productivity
17 and efficiency of water use in some plant species. Secondary effects of climate change involve
18 changes in soil characteristics and disturbance regimes (e.g., fires, pests, diseases), which would
19 favor some species over others and thus change the species composition of ecosystems.
20 Based on model simulations of vegetation distribution, which use GCM-based climate
21 scenarios, large shifts of vegetation boundaries into higher latitudes and elevations can be
22 expected. The mix of species within a given vegetation class likely will change. Under
23 equilibrium GCM climate scenarios, large regions show drought-induced declines in vegetation,
24 even when the direct effects of CO2 fertilization are included. By comparison, under transient
25 climate scenarios, in which trace gases increase slowly over a period of years, the full effects of
26 changes in temperature and precipitation lag the effects of a change in atmospheric composition
27 by a number of decades; hence, the positive effects of CO2, precede the full effects of changes in
28 climate.
29 Climate change is projected to occur at a rapid rate relative to the speed at which forest
30 species grow, reproduce, and reestablish themselves (past tree species' migration rates are
31 believed to be on the order of 4 to 200 km per century). For mid-latitude regions, an average
March 2001 4C-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 warming of 1 to 3.5 °C over the next 100 years would be equivalent to a poleward shift of the
2 present geographic bands of similar temperatures (or "isotherms") by approximately 150 to
3 550 km, or an altitude shift of about a 150 to 550 m. Therefore, the species composition of
4 forests is likely to change; in some regions, entire forest types may disappear, and new
5 assemblages of species and, hence, new ecosystems may be established. As a consequence of
6 possible changes in temperature and water availability under doubled equivalent-CO2 equilibrium
7 conditions, a substantial fraction (a global average of one-third, varying by region from
8 one-seventh to two-thirds) of the existing forested area of the world likely would undergo major
9 changes in broad vegetation types, with the greatest changes occurring in high latitudes and the
10 least in the tropics. In tropical rangelands, major alterations in productivity and species
11 composition would occur because of altered rainfall amount and seasonality and increased
12 evapotranspiration, although a mean temperature increase alone would not lead to such changes.
13 Inland aquatic ecosystems will be influenced by climate change through altered water
14 temperatures, flow regimes, water levels, and thawing of permafrost at high latitudes. In lakes
15 and streams, warming would have the greatest biological effects at high latitudes, where
16 biological productivity would increase and lead to expansion of cool-water species' ranges, and
17 at the low-latitude boundaries of cold- and cool-water species ranges, where extinctions would be
18 greatest. Increases in flow variability, particularly the frequency and duration of large floods and
19 droughts, would tend to reduce water quality, biological productivity, and habitat in streams. The
20 geographical distribution of wetlands is likely to shift with changes in temperature and
21 precipitation, with uncertain implications for net greenhouse gas emissions from nontidal
22 wetlands. Some coastal ecosystems (saltwater marshes, mangrove ecosystems, coastal wetlands,
23 coral reefs, coral atolls, and river deltas) are particularly at risk from climate change and other
24 stresses. Changes in these ecosystems would have major negative effects on freshwater supplies,
25 fisheries, biodiversity, and tourism.
26 Adaptation options for ecosystems are limited, and their effectiveness is uncertain. Options
27 include establishment of corridors to assist the "migration" of ecosystems, land-use management,
28 plantings, and restoration of degraded areas. Because of the projected rapid rate of change
29 relative to the rate at which species can reestablish themselves, the isolation and fragmentation
30 of many ecosystems, the existence of multiple stresses (e.g., land-use change, pollution), and
March 2001 4C-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 limited adaptation options, ecosystems (especially forested systems, montane systems, and coral
2 reefs) are vulnerable to climate change.
3
4 Hydrology and Water Resources
5 Water availability is an essential component of welfare and productivity. Currently,
6 1.3 billion people do not have access to adequate supplies of safe water, and 2 billion people do
7 not have access to adequate sanitation. Although these people are dispersed throughout the
8 globe, reflecting subnational variations in water availability and quality, some 19 countries
9 (primarily in the Middle East and northern and southern Africa) face such severe shortfalls that
10 they are classified as either water-scarce or water-stressed; this number is expected to roughly
11 double by 2025, in large part because of increases in demand resulting from economic and
12 population growth. For example, most policy makers now recognize drought as a recurrent
13 feature of Africa's climate. However, climate change will further exacerbate the frequency and
14 magnitude of droughts in some places.
15 Changes in climate could exacerbate periodic and chronic shortfalls of water, particularly in
16 arid and semi-arid areas of the world. Developing countries are highly vulnerable to climate
17 change because many are located in arid and semi-arid regions, and most derive their water
18 resources from single-point systems such as bore holes or isolated reservoirs. These systems, by
19 their nature, are vulnerable because there is no redundancy in the system to provide resources,
20 should the primary supply fail. Also, given the limited technical, financial, and management
21 resources possessed by developing countries, adjusting to shortages or implementing adaptation
22 measures will impose a heavy burden on their national economies. There is evidence that
23 flooding is likely to become a larger problem in many temperate and humid regions, requiring
24 adaptations not only to droughts and chronic water shortages but also to floods and associated
25 damages, raising concerns about dam and levee failures.
26 The impacts of climate change will depend on the baseline condition of the water supply
27 system and the ability of water resources managers to respond not only to climate change but also
28 to population growth and changes in demands; technology; and economic, social, and legislative
29 conditions.
30 Various approaches are available to reduce the potential vulnerability of water systems to
31 climate change. Options include pricing systems, water efficiency initiatives, engineering and
March 2001 4C-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 structural improvements to water supply infrastructure, agriculture policies, and urban planning
2 and management. At the national/regional level, priorities include placing greater emphasis on
3 integrated, cross-sectoral water resources management, using river basins as resource
4 management units, and encouraging sound pricing and management practices. Given increasing
5 demands, the prevalence and sensitivity of many simple water management systems to
6 fluctuations in precipitation and runoff, and the considerable time and expense required to
7 implement many adaptation measures, the water resources sector in many regions and countries
8 is vulnerable to potential changes in climate.
9
10 Food and Fiber Production
11 Currently, 800 million people are malnourished; as the world's population increases and
12 incomes in some countries rise, food consumption is expected to double over the next three to
13 four decades. The most recent doubling in food production occurred over a 25-year period and
14 was based on irrigation, chemical inputs, and high-yielding crop varieties. Whether the
15 remarkable gains of the past 25 years will be repeated is uncertain. Problems associated with
16 intensifying production on land already in use (e.g., chemical and biological runoff, waterlogging
17 and salinization of soils, soil erosion and compaction) are becoming increasingly evident.
18 Expanding the amount of land under cultivation (including reducing land deliberately taken out
19 of production to reduce agricultural output) also is an option for increasing total crop production,
20 but it could lead to increases in competition for land and pressure on natural ecosystems,
21 increased agricultural emissions of greenhouse gases, a reduction in natural sinks of carbon, and
22 expansion of agriculture to marginal lands, all of which could undermine the ability to
23 sustainably support increased agricultural production.
24 Changes in climate will interact with stresses that result from actions to increase
25 agricultural production, affecting crop yields and productivity in different ways, depending on the
26 types of agricultural practices and systems in place. The main direct effects will be through
27 changes in factors such as temperature, precipitation, length of growing season, and timing of
28 extreme or critical threshold events relative to crop development, as well as through changes in
29 atmospheric CO2 concentration (which may have a beneficial effect on the growth of many crop
30 types). Indirect effects will include potentially detrimental changes in diseases, pests, and weeds,
31 the effects of which have not yet been quantified in most available studies. Evidence continues
March 2001 4C-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 to support the findings of the IPCC SAR that "global agricultural production could be maintained
2 relative to baseline production" for a growing population under 2xCO2 equilibrium climate
3 conditions. In addition, the regional findings of this special report lend support to concerns over
4 the "potential serious consequences" of increased risk of hunger in some regions, particularly the
5 tropics and subtropics. Generally, middle to high latitudes may experience increases in
6 productivity, depending on crop type, growing season, changes in temperature regimes, and the
7 seasonality of precipitation. In the tropics and subtropics, where some crops are near their
8 maximum temperature tolerance and where dry-land, nonirrigated agriculture predominates,
9 yields are likely to decrease. The livelihoods of subsistence farmers and pastoral peoples, who
10 make up a large portion of rural populations in some regions, also could be affected negatively.
11 In regions where there is a likelihood of decreased rainfall, agriculture could be significantly
12 affected.
13 Fisheries and fish production are sensitive to changes in climate and currently are at risk
14 from overfishing, diminishing nursery areas, and extensive inshore and coastal pollution.
15 Globally, marine fisheries production is expected to remain about the same in response to
16 changes in climate; high-latitude freshwater and aquaculture production is likely to increase,
17 assuming that natural climate variability and the structure and strength of ocean currents remain
18 about the same. The principal impacts will be felt at the national and local levels, as centers of
19 production shift. The positive effects of climate change, such as longer growing seasons, lower
20 natural winter mortality, and faster growth rates in higher latitudes, may be offset by negative
21 factors such as changes in established reproductive patterns, migration routes, and ecosystem
22 relationships.
23 Given the many forces bringing profound change to the agricultural sector, adaptation
24 options that enhance resilience to current natural climate variability and potential changes in
25 means and extremes and address other concerns (e.g., soil erosion, salinization) offer no- or
26 low-regret options. For example, linking agricultural management to seasonal climate
27 predictions can assist in incremental adaptation, particularly in regions where climate is strongly
28 affected by ENSO conditions. The suitability of these options for different regions varies, in part
29 because of differences in the financial and institutional ability of the private sector and
30 governments in different regions to implement them. Adaptation options include changes in
31 crops and crop varieties; development of new crop varieties; changes in planting schedules and
March 2001 4C-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 tillage practices; introduction of new biotechnologies; and improved water-management and
2 irrigation systems, which have high capital costs and are limited by availability of water
3 resources. Other options, such as minimum- and reduced-tillage technologies, do not require
4 such extensive capitalization but do require high levels of agricultural training and support.
5 In regions where agriculture is well adapted to current climate variability or where market
6 and institutional factors are in place to redistribute agricultural surpluses to make up for
7 shortfalls, vulnerability to changes in climate means and extremes generally is low. However, in
8 regions where agriculture is unable to cope with existing extremes, where markets and
9 institutions to facilitate redistribution of deficits and surpluses are not in place, or where
10 adaptation resources are limited, the vulnerability of the agricultural sector to climate change
11 should be considered high. Other factors also will influence the vulnerability of agricultural
12 production in a particular country or region to climate change, including the extent to which
13 current temperatures or precipitation patterns are close to or exceed tolerance limits for important
14 crops, per capita income, the percentage of economic activity based on agricultural production,
15 and the preexisting condition of the agricultural land base.
16
17 Coastal Systems
18 Coastal zones are characterized by a rich diversity of ecosystems and a great number of
19 socioeconomic activities. Coastal human populations in many countries have been growing at
20 double the national rate of population growth. Currently, it is estimated that about half of the
21 global population lives in coastal zones, although there is large variation among countries.
22 Changes in climate will affect coastal systems through sea-level rise and an increase in
23 storm-surge hazards and possible changes in the frequency or intensity of extreme events.
24 Coasts in many countries currently face severe sea-level rise problems as a consequence of
25 tectonically and anthropogenically induced subsidence. An estimated 46 million people per year
26 currently are at risk of flooding from storm surges. Climate change will exacerbate these
27 problems, leading to potential impacts on ecosystems and human coastal infrastructure. Large
28 numbers of people also potentially are affected by sea-level rise, for example, tens of millions of
29 people in Bangladesh would be displaced by a 1-m increase (the top of the range of IPCC
30 Working Group I estimates for 2100) in the absence of adaptation measures. A growing number
31 of extremely large cities are located in coastal areas, which means that large amounts of
March 2001 4C-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 infrastructure may be affected. Although annual protection costs for many nations are relatively
2 modest, about 0.1% of gross domestic product (GDP), the average annual costs to many small
3 island states total several percent of GDP. For some island nations, the high cost of providing
4 storm-surge protection would make it essentially infeasible, especially given the limited
5 availability of capital for investment.
6 Beaches, dunes, estuaries, and coastal wetlands adapt naturally and dynamically to changes
7 in prevailing winds and seas, as well as sea-level changes; in areas where infrastructure
8 development is not extensive, planned retreat and accommodation to changes may be possible.
9 It also may be possible to rebuild or relocate capital assets at the end of their design life. In other
10 areas, however, accommodation and planned retreat are not viable options, and protection using
11 hard structures (e.g., dikes, levees, floodwalls, barriers) and soft structures (e.g., beach
12 nourishment, dune restoration, wetland creation) will be necessary. Factors that limit the
13 implementation of these options include inadequate financial resources, limited institutional and
14 technological capability, and shortages of trained personnel. In most regions, current coastal
15 management and planning frameworks do not take account of the vulnerability of key systems to
16 changes in climate and sea level or long lead times for implementation of many adaptation
17 measures. Inappropriate policies encourage development in impact-prone areas. Given
18 increasing population density in coastal zones; long lead times for implementation of many
19 adaptation measures; and institutional, financial, and technological limitations (particularly in
20 many developing countries), coastal systems should be considered vulnerable to changes in
21 climate.
22
23 Human Health
24 In much of the world, life expectancy is increasing; in addition, infant and child mortality
25 in most developing countries is droping. Against this positive backdrop, however, there appears
26 to be a widespread increase in new and resurgent vectorborne and infectious diseases, such as
27 dengue, malaria, hantavirus, and cholera. In addition, the percentage of the developing world's
28 population living in cities is expected to increase from 25% (in 1960) to more than 50% by 2020,
29 with percentages in some regions far exceeding these averages. These changes will bring
30 benefits only if accompanied by increased access to services such as sanitation and potable water
31 supplies; they also can lead to serious urban environmental problems, including air pollution
March 2001 4C-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 (e.g., particulates, surface ozone, lead), poor sanitation, and associated problems in water quality
2 and potability, if access to services is not improved.
3 Climate change could affect human health through increases in heat-stress mortality,
4 tropical vector-borne diseases, urban air pollution problems, and decreases in cold-related
5 illnesses. Compared with the total burden of ill health, these problems are not likely to be large.
6 In the aggregate, however, the direct and indirect impacts of climate change on human health do
7 constitute a hazard to human population health, especially in developing countries in the tropics
8 and subtropics; these impacts have considerable potential to cause significant loss of life, affect
9 communities, and increase health-care costs and lost work days. Model projections (which entail
10 necessary simplifying assumptions) indicate that the geographical zone of potential malaria
11 transmission would expand in response to global mean temperature increases at the upper part of
12 the IPCC-projected range (3 to 5 °C by 2100), increasing the affected proportion of the world's
13 population from approximately 45% to approximately 60% by the latter half of the next century.
14 Areas where malaria is currently endemic could experience intensified transmission (on the order
15 of 50 to 80 million additional annual cases, relative to an estimated global background total of
16 500 million cases). Some increases in non-vector-borne infectious diseases, such as
17 salmonellosis, cholera, and giardiasis, also could occur as a result of elevated temperatures and
18 increased flooding. However, quantifying the projected health impacts is difficult because the
19 extent of climate induced health disorders depends on other factors, such as migration, provision
20 of clean urban environments, improved nutrition, increased availability of potable water,
21 improvements in sanitation, the extent of disease vector-control measures, changes in resistance
22 of vector organisms to insecticides, and more widespread availability of health care. Human
23 health is vulnerable to changes in climate, particularly in urban areas, where access to space
24 conditioning may be limited, as well as in areas where exposure to vector-borne and
25 communicable diseases may increase and health-care delivery and basic services, such as
26 sanitation, are poor.
27
28 Regional Vulnerability to Global Climate Change
29 Discussions about two geographic regions (North American and Polar regions) assessed in
30 the report are included here because of their relevance to the continental United States and
31 Alaska, respectively.
March 2001 4C-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 North American Region
2 This region consists of Canada and the United States south of the Arctic Circle. Within the
3 region, vulnerability to and the impacts of climate change vary significantly from sector to sector
4 and from subregion to subregion. This "texture" is important in understanding the potential
5 effects of climate change on North America, as well as in formulating and implementing viable
6 response strategies.
7 Ecosystems. Most ecosystems are moderately to highly sensitive to changes in climate.
8 Effects are likely to include both beneficial and harmful changes. Potential impacts include
9 northward shifts of forest and other vegetation types, which would affect biodiversity by altering
10 habitats and would reduce the market and nonmarket goods and services they provide; declines in
11 forest density and forested area in some subregions, but gains in others; more frequent and larger
12 forest fires; expansion of arid land species into the great basin region; drying of prairie pothole
13 wetlands that currently support over 50% of all waterfowl in North America; and changes in
14 distribution of habitat for cold-, cool-, and warm-water fish. The ability to apply management
15 practices to limit potential damages is likely to be low for ecosystems that are not already
16 intensively managed.
17 Hydrology and Water Resources. Water quantity and quality are particularly sensitive to
18 climate change. Potential impacts include increased runoff in winter and spring and decreased
19 soil moisture and runoff in summer. The Great Plains and prairie regions are particularly
20 vulnerable. Projected increases in the frequency of heavy rainfall events and severe flooding also
21 could be accompanied by an increase in the length of dry periods between rainfall events and in
22 the frequency or severity of droughts in parts of North America. Water quality could suffer and
23 would decline where minimum river flows decline. Opportunities to adapt are extensive, but
24 their costs and possible obstacles may be limiting.
25 Food and Fiber Production. The productivity of food and fiber resources of North
26 America is moderately to highly sensitive to climate change. Most studies, however, have not
27 fully considered the effects of potential changes in climate variability; water availability; stresses
28 from pests, diseases, and fire; or interactions with other, existing stresses. Warmer climate
29 scenarios (4 to 5 °C increases in North America) have yielded estimates of negative impacts in
30 eastern, southeastern, and corn belt regions and positive effects in northern plains and western
31 regions. More moderate warming produced estimates of predominately positive effects in some
March 2001 4C-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 warm-season crops. Vulnerability of commercial forest production is uncertain, but is likely to
2 be lower than less intensively managed systems because of changing technology and
3 management options. The vulnerability of food and fiber production in North America is thought
4 to be low at the continental scale, though subregional variation in losses or gains is likely. The
5 ability to adapt may be limited by information gaps; institutional obstacles; high economic,
6 social, and environmental costs; and the rate of climate change.
7 Coastal Systems. Sea level has been rising relative to the land along most of the coast of
8 North America, and falling in a few areas, for thousands of years. During the next century, a
9 50-cm rise in sea level from climate change alone could inundate 8500 to 19,000 square
10 kilometers of dry land, expand the 100-year flood plain by more than 23,000 square kilometers,
11 and eliminate as much as 50% of North America's coastal wetlands. The projected changes in
12 sea level because of climate change alone would underestimate the total change in sea level from
13 all causes along the eastern seaboard and Gulf Coast of North America. In many areas, wetlands
14 and estuarine beaches may be squeezed between advancing seas and dikes or seawalls built to
15 protect human settlements. Several local governments are implementing land-use regulations to
16 enable coastal ecosystems to migrate landward as sea level rises. Saltwater intrusion may
17 threaten water supplies in several areas.
18 Human Settlements. Projected changes in climate could have positive and negative impacts
19 on the operation and maintenance costs of North American land and water transportation. Such
20 changes also could increase the risks to property and human health and life as a result of possible
21 increased exposure to natural hazards (e.g., wildfires, landslides, extreme weather events) and
22 result in increased demand for cooling and decreased demand for heating energy, with the overall
23 net effect varying across geographic regions.
24 Human Health. Climate can have wide-ranging and potentially adverse effects on human
25 health via direct pathways (e.g., thermal stress, extreme weather and climate events) and indirect
26 pathways (e.g., disease vectors and infectious agents, environmental and occupational exposures
27 to toxic substances, food production). In high-latitude regions, some human health impacts are
28 expected because of dietary changes resulting from shifts in migratory patterns and abundance of
29 native food sources.
30 Conclusions. Taken individually, any one of the impacts of climate change may be within
31 the response capabilities of a subregion or sector. The fact that they are projected to occur
March 2001 4C-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 simultaneously and in concert with changes in population, technology, economics, and other
2 environmental and social changes, however, adds to the complexity of the impact assessment and
3 the choice of appropriate responses. The characteristics of subregions and sectors of North
4 America suggest that neither the impacts of climate change nor the response options will be
5 uniform.
6 Many systems of North America are moderately to highly sensitive to climate change, and
7 the range of estimated effects often includes the potential for substantial damages. The
8 technological capability to adapt management of systems to lessen or avoid damaging effects
9 exists in many instances. The ability to adapt may be diminished, however, by the attendant
10 costs, lack of private incentives to protect publicly owned natural systems, imperfect information
11 regarding future changes in climate and the available options for adaptation, and institutional
12 barriers. The most vulnerable sectors and regions include long-lived natural forest ecosystems in
13 the east and interior west, water resources in the southern plains, agriculture in the southeast and
14 southern plains, human health in areas currently experiencing diminished urban air quality,
15 northern ecosystems and habitats, estuarine beaches in developed areas, and low-latitude cool
16 and cold-water fisheries. Other sectors and subregions may benefit from opportunities associated
17 with warmer temperatures or, potentially, from CO2 fertilization, including west coast coniferous
18 forests; some western rangelands; reduced energy costs for heating in the northern latitudes;
19 reduced salting and snow-clearance costs; longer open-water seasons in northern channels and
20 ports; and agriculture in the northern latitudes, the interior west, and the west coast.
21
22 Polar (Arctic and Antarctic) Regions
23 The polar regions include some very diverse landscapes, and the Arctic and the Antarctic
24 are very different in character. The Arctic is defined here as the area within the Arctic Circle; the
25 Antarctic here includes the area within the Antarctic Convergence, including the Antarctic
26 continent, the Southern Ocean, and the sub-Antarctic islands. The Arctic can be described as a
27 frozen ocean surrounded by land, and the Antarctic as a frozen continent surrounded by ocean.
28 The projected warming in the polar regions is greater than for many other regions of the world.
29 Where temperatures are close to freezing on average, global warming will reduce land ice and
30 sea ice, the former contributing to sea-level rise. However, in the interiors of ice caps, increased
March 2001 4C-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 temperature may not be sufficient to lead to melting of ice and snow, and will tend to have the
2 effect of increasing snow accumulation.
3 Ecosystems. Major physical and ecological changes are expected in the Arctic. Frozen
4 areas close to the freezing point will thaw and undergo substantial changes with warming.
5 Substantial loss of sea ice is expected in the Arctic Ocean. As warming occurs, there will be
6 considerable thawing of permafrost, leading to changes in drainage, increased slumping, and
7 altered landscapes over large areas. Polar warming probably will increase biological production
8 but may lead to different species composition on land and in the sea. On land, there will be a
9 tendency for polar shifts in major biomes such as tundra and boreal forest and associated
10 animals, with significant impacts on species such as bear and caribou. However, the Arctic
11 Ocean geographically limits northward movement. Much smaller changes are likely for the
12 Antarctic, but there may be species shifts. In the sea, marine ecosystems will move poleward.
13 Animals dependent on ice may be disadvantaged in both polar areas.
14 Hydrology and Water Resources. Increasing temperature will thaw permafrost and melt
15 more snow and ice. There will be more running and standing water. Drainage systems in the
16 Arctic are likely to change at the local scale. River and lake ice will break up earlier and freeze
17 later.
18 Food and Fiber Production. Agriculture is severely limited by the harsh climate. Many
19 limitations will remain in the future, although some small northern extension of farming into the
20 Arctic may be possible. In general, marine ecological productivity should rise. Warming should
21 increase growth and development rates of nonmammals; ultraviolet-B (UV-B) radiation is still
22 increasing, however, which may adversely affect primary productivity as well as fish
23 productivity.
24 Coastal Systems. As warming occurs, the Arctic could experience a thinner and reduced
25 ice cover. Coastal and river navigation will increase, with new opportunities for water transport,
26 tourism, and trade. The Arctic Ocean could become a major global trade route. Reductions in
27 ice will benefit offshore oil production. Increased erosion of Arctic shorelines is expected from a
28 combination of rising sea level, permafrost thaw, and increased wave action as a result of
29 increased open water. Further breakup of ice shelves in the Antarctic peninsula is likely.
30 Elsewhere in Antarctica, little change is expected in coastlines and probably in its large ice
31 shelves.
March 2001 4C-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Human Settlements. Human communities in the Arctic will be substantially affected by the
2 projected physical and ecological changes. The effects will be particularly important for
3 indigenous peoples leading traditional lifestyles. There will be new opportunities for shipping,
4 the oil industry, fishing, mining, tourism, and migration of people. Sea ice changes projected for
5 the Arctic have major strategic implications for trade, especially between Asia and Europe.
6 Conclusions. The Antarctic peninsula and the Arctic are very vulnerable to projected
7 climate change and its impacts. Although the number of people directly affected is relatively
8 small, many native communities will face profound changes that impact on traditional lifestyles.
9 Direct effects could include ecosystem shifts, sea and river-ice loss, and permafrost thaw.
10 Indirect effects could include feedbacks to the climate system such as further releases of
11 greenhouse gases, changes in ocean circulation drivers, and increased temperature and higher
12 precipitation with loss of ice, which could affect climate and sea level globally. The interior of
13 Antarctica is less vulnerable to climate change, because the temperature changes envisaged over
14 the next century are likely to have little impact and very few people are involved. However,
15 there are considerable uncertainties about the mass balance of the Antarctic ice sheets and the
16 future behavior of the West Antarctic ice sheet (low probability of disintegration over the next
17 century). Changes in either could affect sea level and Southern Hemisphere climates.
18
19
20 REFERENCES
21 Intergovernmental Panel on Climate Change (IPCC). (1996) Climate change 1996: contribution of working group I
22 to the second assessment of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom:
23 Cambridge University Press; p. 572. [IPCC second assessment report].
24 Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
25 of vulnerability. Cambridge, United Kingdom: Cambridge University Press.
26
March 2001 4C-17 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 4D
2
3 Excerpted Materials from the U.S. Global Change Research
4 Program Assessment Overview Report on Climate Change Impacts
5 on the United States (U.S. Global Change Research Program, 2000) and
6 Subsidiary Regional Assessment Reports
9 The subject Assessment Overview on Climate Change Impacts on the United States
10 (USGCRP, 2000) was prepared by the USGCRP National Assessment Synthesis Team (NAST)
11 and represents an important landmark for the U.S. national process of research analyses and
12 dialog about coming changes in climate, their impacts, and possible adaptation measures that can
13 be taken. NAST consists of a committee of experts drawn from governments, academia industry,
14 and nongovernmental organizations (NGO's). The subject overview is based on a much more
15 extensive, detailed "foundation" report, written by NAST in coordination with independent
16 regional and sector assessment teams. The subject assessment, required by a 1990 U.S. law, was
17 conducted under the USGCRP in response to a request from the President's Science Advisor.
18 The materials presented below are excerpted from the September 2000 NSTC review draft of the
19 assessment overview and, if necessary, later will be appropriately corrected to reflect the final
20 versions of the report due out in fall 2000, after completion of all peer-review and clearance
21 precesses. Selected material derived from one or another of the specific regional assessments
22 also are presented in this appendix. The materials selected for presentation here are meant to
23 provide an informative introduction to the latest available expert assessment of potential sector
24 and regional-scale impacts of climate change in the United States and to illustrate the difficulties
25 in projecting likely varying location-specific mixes of potential deleterious and beneficial effects
26 of climate change.
27 The past record of 1000 years of global temperature and CO2 emissions change, as
28 depicted by the assessment overview, is shown in Figure 4D-1. As noted in the Figure 4D-1,
29 there appears to be a relatively close correlation between marked parallel increases in
30 anthropogenic carbon emissions starting roughly in the latter part of the 18th century, increasing
31 atmospheric CO2 concentrations, and notable increasing global average temperature trends.
March 2001 4D-1 DRAFT-DO NOT QUOTE OR CITE
-------
0.8
Temperature Change
-0.6
oinoinoi^oi^oLooinoinoiooinoifloin
8*3-a>rococNr^^cooioOT^cocoi>CNco .- _ ^
oo^^cMcNcrjt^^r^r^LnincD
CD r^ r- co co
380'
0.
Q^
11
_2
§•
>*
JD
360
340
320
C0
Q-
8
300'
280
260'
CO2 Concentrations
Soooooooooooooooooo o
goLnoLnoLnoLnoinoiooioomoLD o
OT-t-CMCMCDCY)1^tnt^ininCDCDr--r--COCOCJ)CD O
LD LO in
_ I"- co o
i^ i-^ r^- r^
mLOLOLOLOLOLOLOLOl^
g-^CMCO-^-LOCDr^OOOl
C1O)OJO)O)O)O)O>O5
Figure 4D-1. Records of Northern Hemisphere surface temperatures, CO2 concentrations,
and carbon emissions show a close correlation. Temperature Change:
reconstruction of annual-average Northern Hemisphere surface air
temperatures derived from historical records, tree rings, and corals (blue),
and air temperatures directly measured (purple). CO2 Concentrations:
record of global CO2 concentration for the last 1000 years, derived from
measurements of CO2 concentration in air bubbles in the layered ice cores
drilled in Antarctica (blue line) and from atmospheric measurements since
1957. Carbon Emissions: reconstruction of past emissions of CO2 as a result
of land clearing and fossil fuel combustion since about 1750 (in billions of
metric tons of carbon per year.
March 2001
4D-2
DRAFT-DO NOT QUOTE OR CITE
-------
1 The two primary models used to project future changes in global climate in the USGCRP
2 assessment were developed at the Canadian Climate Centre and the Hadley Centre in the United
3 Kingdom and have been peer-reviewed extensively by other scientists. Both incorporate similar
4 assumptions about future emissions of carbon dioxide and other major greenhouse gases (both
5 approximate the IPCC "business as usual" scenario with a 1% per year increase in greenhouse
6 gases and growing sulfur emissions). These models were the best fit to a list of criteria
7 developed for the U.S. National Assessment. Climate models developed at the National Center
8 for Atmospheric Research (NCAR), NOAA's Geophysical Fluid Dynamics Laboratory (GFDL),
9 NASA's Goddard Institute for Space Studies (GISS), and Max Planck Institute (MPI) in
10 Germany also were used in various aspects of the assessment. Although the physical principles
11 driving the models are similar, they differ in how they represent the effects of some important
12 processes, with the two primary models yielding different views of 21st century climate. On
13 average over the United States, the Hadley model projects a much wetter climate than does the
14 Canadian model, although the Canadian model projects a greater increase in temperature than
15 does the Hadley. Both projections are plausible, given current understanding. See Figure 4D-2
16 for plots of U.S. average temperature increases projected by the different models. In all climate
17 models, increases in temperature for the United States are significantly higher than global
18 average temperature increases (see Table 4D-1), because of the fact that all models project
19 warming to be greatest at middle to high latitudes (partly because melting snow and ice make the
20 surface less reflective of sunlight, allowing it to absorb more heat). Warming also will be greater
21 over land than over the oceans because it takes longer for the oceans to warm.
22 Uncertainties about future climate stem from a wide variety of factors (e.g., questions about
23 how to represent clouds and precipitation in climate models and uncertainties about how
24 emissions of greenhouse gases will change). These uncertainties result in differences in climate
25 model projections. Examining these differences aids in understanding the range of risk or
26 opportunity associated with a plausible range of future climate changes. These differences in
27 model projections also raise questions about how to interpret model results, especially at the
28 regional level, where projections can differ significantly.
29 One of the most important world-wide consequences of the overall global warming
30 increases projected for the 21st century is sea level rise, and it can be expected to impact Alaska,
31 coastal areas of the continental United States, and U.S. Hawaiian and Carribean islands regions,
March 2001 4D-3 DRAFT-DO NOT QUOTE OR CITE
-------
CD
O)
c
CD
.C
O
CD
CD
Q.
E
.CD
9-
8-
7-
6-
5-
4-
3-
2-
1-
0-
-1-
-2-
-3-
-4-
-5
Changes in Temperature over the U.S.
Simulated by Climate Models
Hadley Centre - Version 2
Canadian Centre
Max Planck Institute
Geophysical Fluid Dynamics Laboratory
Hadley Centre - Version 3
NCAR - Parallel Version
NCAR - Climate System Model
1850
1900
1950
I
2000
I
2050
2100
Figure 4D-2. Simulation of decadal average changes in temperature from leading climate
models on historic and projected changes in CO2 and sulfate atmospheric
concentrations. For the 21st century, the projected global temperature
increase is 4.9 °F for the Hadley model and 7.4 °F for the Canadian model.
The model with the smallest projected increase in global temperature is the
NCAR Climate System Model at 3.6 °F. By comparison, the projected
increase in temperature for the 21st century over the contiguous United
States is Canadian, 9.4 °F, Hadley, 5.5 °F, and NCAR Climate System
Model, 4.0 °F.
March 2001
4D-4
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4D-1. RANGE OF PROJECTED WARMING IN THE 21ST CENTURY
Global United States
Hadley Model +4.9 °F +5.5 °F
Canadian Model +7.4 °F +9.4 °F
NCAR Climate System Model +3.6 °F +4.0 °F
1 as well. Figure 4D-3 illustrates sea level rise predicted by the Canadian and Hadley models used
2 in the USGCRP assessment. The sea level rise projected by either model can be expected to pose
3 threats, not only in terms of potential inundation of low lying portions of the Hawaiian and
4 Carribean islands, but also in terms of shoreline erosion in portions of Alaska and the continental
5 United States. Those continental United States areas most vulnerable to future sea level rise are
6 those low lying areas already experiencing rapid erosion rates, as depicted in Figure 4D-4.
7 Substantial impacts can be expected, including losses of coastal wetlands important for migratory
8 birds and degradation of estuarine sound complexes providing shallow water fishery nurseries
9 (most immediately because of salt water incursions resulting from sea level rise and other
10 impacts resulting from more frequent and extensive algal-toxic blooms impacting coastal
11 commercial fish and shell fish harvests secondary to increased nutrient out flows caused by
12 extreme rain fall events [e.g., during hurricanes]).
13 The main climate models used all predict notable increases in the minimum and maximum
14 annual average temperatures in the United States during the next 100 years. Projected changes in
15 temperature minimum and maxima are likely more important than average temperatures, in that
16 they influence such things as human comfort, heat and cold stress in plants and animals,
17 maintenance of snow pack, and pest populations (low temperatures kill many pests and higher
18 minimum temperatures may allow increased overwinter survival of pests). The largest increases
19 in temperature are projected over much of the southern United States in summer, dramatically
20 raising the heat index (a measure of discomfort based on temperature and humidity). Also,
21 following an average 5 to 10% increase in average U.S. precipitation over the last century, the
22 climate models project notable changes in precipitation during the 21st century. The Canadian
March 2001 4D-5 DRAFT-DO NOT QUOTE OR CITE
-------
20
15-
10-
5-
- Canadian Model (Thermal Expansion)
- Hadley Model (Thermal Expansion)
- Hadley Model (I.E. + glacial melt)
1850
1900
1950
2000
2050
2100
Figure 4D-3. Historic and projected changes in sea level (in inches) based on the Canadian
and Hadley model simulations. The Canadian model projection includes
only the effects of thermal expansion of warming ocean waters. The Hadley
projection includes both thermal expansion and the additional sea-level rise
projected because of melting of land-based glaciers. Neither model includes
consideration of possible sea-level changes because of polar ice melting or
accumulation of snow on Greenland and Antarctica.
Severely eroding
Figure 4D-4. This map is a preliminary classification of annual shoreline erosion
throughout the United States, in coarse detail and resolution. The areas most
vulnerable to future sea-level change are those with low relief that are already
experiencing rapid erosion rates, such as the Southeast and Gulf Coast.
March 2001
4D-6
DRAFT-DO NOT QUOTE OR CITE
-------
1 model predicts the largest percentage increases in precipitation in California and the Southwest;
2 but east of the Rocky Mountains, the southern half of the United States is projected to have
3 decreased precipitation (with especially large decreases in eastern Colorado, western Kansas, and
4 in an arc stretching from Louisiana to Virginia). The Hadley model predicts largest percentage
5 increases in southern California and the Southwest, along with lesser increases for the rest of the
6 nation, except for small areas of the Northwest and the Gulf Coast. Both models also predict
7 increases in frequency of heavy precipitation effects, largely because of shifts in storm activity
8 and tracks. Soil moisture critical for both agriculture and natural ecosystems may, despite
9 increased precipitation, actually undergo marked decreases in some areas, because of offsetting
10 evaporation rates increased by higher temperatures during projected scenarios of likely increased
11 periods of drought for some U.S. regions.
12 The predicted changes in temperature and precipitation are expected to result in varying
13 impacts of climate change on ecosystems various U.S. regions. Such impacts will likely include
14 the following.
15 • Changes in productivity and carbon storage capacity of ecosystems (decreases in some places
16 and increases in others are very likely).
17 • Shifts in the distribution of major plant and animal species are likely.
18 • Some ecosystems, such as alpine meadows, are likely to disappear in some places because the
19 new local climate will not support them or there are barriers to their movement.
20 • In many places, it is very likely that ecosystem services, such as air and water purification,
21 landscape stabilization against erosion, and carbon storage capacity will be reduced. These
22 losses likely will occur in the wake of episodic, large-scale disturbances that trigger species
23 migrations or local extinctions.
24 • In some places, it is very likely that ecosystems services will be enhanced where climate-
25 related stresses are reduced.
26 The USGCRP assessment provides extensive detailed evaluations of the above and other
27 types of impacts projected to occur as consequences of changing weather patterns (and
28 consequent shifts in temperature, precipitation, etc.). Such evaluations are summarized in the
29 overview assessment in relation to several overall sectors (water resources, agriculture, forests,
30 coastal areas and marine resources, and human health) and in relation to different regions of the
31 United States.
March 2001 4D-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 The following concise statements highlight some of the more salient points emerging from
2 the overall sector evaluations.
3 • Water. Rising temperatures and greater precipitation are likely to lead to more evaporation
4 and greater swings between wet and dry conditions. Changes in the amount and timing of rain,
5 snow, runoff, and soil moisture are very likely. Water management, including pricing and
6 allocation, will very likely be important in determining many impacts.
7 • Agriculture. Overall productivity of American agriculture likely will remain high and is
8 projected to increase throughout the 21st century, with northern regions faring better than
9 southern ones. Though agriculture is highly dependent on climate, it is also highly adaptive
10 Weather extremes, pests, and weeds likely will present challenges in a changing climate.
11 Falling commodity prices and competitive pressures are likely to stress farmers and rural
12 communities.
13 • Forests. Rising CO2 concentrations and modest warming are likely to increase forest
14 productively in many regions. With larger increases in temperature, increased drought is likely
15 to reduce forest productivity in some regions, notably in the Southeast and Northwest. Climate
16 change is likely to cause shifts in species ranges, as well as large changes in disturbances such
17 as fire and pests.
18 • Coastal Areas and Marine Resources. Coastal wetlands and shorelines are vulnerable to
19 sea-level rise and storm surges, especially when climate impacts are combined with the
20 growing stressed of increasing human population and development. It is likely that coastal
21 communities will be affected increasingly by extreme events. The negative impacts on natural
22 ecosystems are very likely to increase.
23 • Human Health. Heat-related illnesses and deaths, air pollution, injuries and deaths from
24 extreme weather events, and diseases carried by water, food, insects, ticks, and rodents, have
25 all been raised as concerns for the United States in a warmer world. Modern public health
26 efforts will be important in identifying and adapting to these potential impacts.
27 The USGCRP Assessment also evaluated sector impacts in relation to various U.S. regions,
28 broken out as depicted in Figure 4D-5 derived from the overview assessment (USGCRP, 2000).
29 That assessment highlighted the following important points in relation to expected major impacts
30 in each of the regions evaluated.
31
March 2001 4D-8 DRAFT-DO NOT QUOTE OR CITE
-------
.
to
o
o
fe
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
U.S. Regions
Figure 4D-5. Breakout of U.S. regions as evaluated by USGCRP, based on the overview assessment depiction.
-------
1 • Alaska. Sharp winter and springtime temperature increases are very likely to cause continued
2 thawing of permafrost, further disrupting forest ecosystems, roads, and buildings.
3 • Northwest. Increasing stream temperatures are very likely to further stress migrating fish,
4 complicating restoration efforts.
5 • Mountain West. Higher winter temperatures are very likely to reduce snowpack and peak
6 runoff and shift the peak to earlier in the spring, reducing summer runoff and complicating
7 water management for flood control, fish runs, cities, and irrigation.
8 • Southwest. With an increase in precipitation, the desert ecosystems native to this region are
9 likely to decline, whereas grasslands and shrublands likely are to expand.
10 • Midwest/Great Plains. Higher CO2 concentrations are likely to offset the effects of rising
11 temperatures on forests and agriculture for several decades, increasing productivity.
12 • Southern Great Plains. Prairie potholes, which provide important habitat for ducks and other
13 migratory waterfowl, are likely to dry up in a warmer climate.
14 • Great Lakes. Lake levels are likely to decline, leading to reduced water supply and more
15 costly transportation. Shoreline damage caused by high water levels is likely to decrease.
16 • Northern and Mountain Regions. It is very probable that warm weather recreational
17 opportunities, such as hiking, will expand, whereas cold weather activities, such as skiing, will
18 contract.
19 • Northeast, Southeast, and Midwest. Rising temperatures are very likely to increase the heat
20 index dramatically in summer, with impacts to health and comfort. Warmer winters are likely
21 to reduce cold-related stresses.
22 • Appalachians. Warmer and moister air very likely will lead to more intense rainfall events,
23 increasing the potential for flash floods.
24 • Southeast. Under warmer wetter scenarios, the range of southern tree species is likely to
25 expand. Under hotter and drier scenarios, it is likely that far southeastern forests will be
26 displaced by grasslands and savannas.
27 • Southeast Atlantic Coast. It is very probable that rising sea levels and storm surge will
28 threaten natural ecosystems and human coastal development and reduce buffering capacity
29 against storm impacts.
30 • Southeast Gulf Coast. Inundation of coastal wetlands will very likely increase, threatening
31 fertile areas for marine life, migrating birds, and waterfowl.
March 2001 4D-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Islands. More intense El Nino and La Nina events are possible and are likely to create extreme
2 fluctuations in water resources for island citizens and the tourists who sustain local economies.
3 Other materials from the overview assessment summarize regional concerns with regard to
4 different types of sector impacts. Tables 4D-2 and 4D-3 present two examples drawn from the
5 overview assessment, denoting concerns or impacts regarding water resources and types of
6 ecosystems, respectively, likely to be impacted in different U.S. regions.
TABLE 4D-2. TYPES OF WATER CONCERNS PROJECTED TO BE IMPORTANT
FOR U.S. REGIONS CONSEQUENT TO FUTURE CLIMATE CHANGE3
Region
Northeast
Southeast
Midwest
Great Plains
West
Northwest
Alaska
Islands
Floods
X
X
X
X
X
X
X
Droughts
X
X
X
X
X
X
X
X
Snowpack
X
X
X
X
X
X
Groundwater
X
X
X
X
X
X
Lake, River, and
Reservoir Levels
X
X
X
X
Quality
X
X
X
X
X
X
This table identifies some of the key regional concerns about water. Many of these issues were raised and
discussed by stakeholders during regional workshops and other Assessment meetings held between 1997 and
2000.
1 As seen in Table 4D-2, different types of water impacts are projected to be of important
2 widespread concern across many different U.S. regions. It should be noted that some limited
3 beneficial effects may occur in some regions (e.g., longer periods of open-water transportation on
4 navigable rivers and sounds in and around Alaska).
5 The overview assessment notes that the information presented in Table 4D-3 represents
6 only a partial list of potential impacts for major ecosystem types and that, although the impacts
7 often are stated in terms of plant-community impacts, it is important to recognize that such plant-
8 community changes also will have animal habitat effects and consequent impacts on both
9 terrestrial and aquatic animal species. Both the plant and animal impacts can have further
10 consequent impacts on human health and welfare, which also can be expected to vary
11 considerably from region to region.
March 2001 4D-11 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4D-3. PROJECTED FUTURE CLIMATE-CHANGE-INDUCED IMPACTS
ON TYPES OF ECOSYSTEMS OF CONCERN TO DIFFERENT U.S. REGIONS
Ecosystem
Type
Impacts
U.S. Regions
Forests Changes in tree species composition and
alteration of animal habitat
Displacement of forests by woodlands and
grasslands under a warmer climate in
which soils are drier
Grasslands Displacement of grasslands by woodlands
and forests under a wetter climate
Increase in success of normative invasive
plant species
Tundra Loss of alpine meadows as their species
are displaced by lower elevation species
Loss of northern tundra as trees migrate
poleward
Changes in plant community composition
and alteration of animal habitat
Semi-arid Increase in woody species and loss of
and Arid desert species under wetter climate
Freshwater Loss of prairie pot holes with more
frequent drought conditions
Habitat changes in rivers and lakes as
amount and timing of runoff changes and
water temperatures rise
Coastal and Loss of coastal wetlands as sea level rises
Marine and coastal development prevents
landward migration
Loss of barrier islands as sea-level rise
prevents landward migration
Changes in quantity and quality of
freshwater delivery to estuaries and bays
alter plant and animal habitats
Loss of coral reefs as water temperature
increases
Changes in ice location and duration alter
marine mammal habitat
NE SE MW GP WE PNW AK IS
X X
X X
X X
X
X
X
X
X
X X
March 2001
4D-12
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
9
10
11
The USGCRP (2000) assessment included extensive detailed evaluation of projected
climate change impacts on the different U.S. regions depicted in Figure 4D-5. Overview reports
on those detailed evaluations by various regional assessment teams are in various stages of
preparation, with information pertaining to each being available via the internet at the following
address: http://www.nacc.usgcrp.gov/regions/.
The wide variation in the types of projected impacts of climate change, both deleterious and
some possible beneficial effects, can be readily illustrated by one example illustrated in
Figure 4D-6. The figure depicts projected types of changes that may occur (with varying degrees
of certainty indicated) as the consequence of climate change impacts on the Mid-Atlantic Region
(MAR) of the United States, including both potentially negative and positive impacts.
Summary of MAR Impacts
Positive Impact
Most Certain
• Agricultural production
• Coastal zones
• Temperature related health status
erosion,
saltwater intrusion
soybeans,
possibly corn
and treefruits
Moderately Certain
• Forester/ production
• Temperature related health status
more growth,
different mix
less cold stress
Uncertain
• Biodiversity
• Fresh water quantity
• Fresh water quality
• Ecological functioning
• Vector and water-borne disease health status
• Environmental effects from agriculture
migration barriers,
invasive species
forest composition,
cold water fisheries
Crypt os poridiosis,
malaria
nutrient leaching,
runoff
warmer
temperatures
more average
streamflow
Figure 4D-6. Projected climate change impacts in the Mid-Atlantic Region (MAR) of the
United States.
Source: Mid-Atlantic Regional Assessment Team (2000).
March 2001
4D-13
DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Mid-Atlantic Regional Assessment (MARA) Team. (2000) Preparing for a changing climate: the potential
3 consequences of climate variability and change (Mid-Atlantic overview). Washington, DC: U.S. Global
4 Change Research Program (USGCRP).
5 U. S. Global Change Research Program (USGCRP, 2000) Climate Change Impacts on the United States: the
6 Potential Consequences of Climate Variability and Change (Overview), Report of National Assessment
7 Synthesis Team (NAST). NSTC Review Draft (September 2000).
March 2001 4D-14 DRAFT-DO NOT QUOTE OR CITE
-------
APPENDIX 4E
Recent Model Projections of Excess Mortality Expected in U.S. Cities
During Summer and Winter Seasons Because of Future Climate Change,
Based on Kalkstein and Greene (1997)
March 2001 4E-1 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4E-1. MODELED PROJECTIONS OF DIRECT HUMAN HEALTH
IMPACTS OF CLIMATE CHANGE: ESTIMATED TOTAL EXCESS MORTALITY
IN U.S. URBAN AREAS FOR AN AVERAGE SUMMER SEASON,
ASSUMING FULL ACCLIMATIZATION3
Year 2020 Climate
SMSA
Anaheim
Atlanta
Baltimore
Birmingham
Boston
Buffalo
Chicago
Cincinnati
Cleveland
Columbus
Dallas
Denver
Detroit
Ft. Lauderdale
Greensboro
Hartford
Houston
Indianapolis
Jacksonville
Kansas City
Los Angeles
Louisville
Memphis
Miami
Minneapolis
Nassau
New Orleans
New York
Present
climate
0
25
84
42
96
33
191
14
29
33
36
42
110
0
22
38
7
36
0
49
68
17
25
0
59
29
20
307
GFDL 89
0
43
57
26
113
15
243
16
21
24
45
29
84
0
28
21
7
23
0
79
74
0
42
0
55
59
0
363
UKMO
0
62
148
47
165
52
538
90
55
83
62
41
240
0
43
42
16
93
0
173
123
2
27
0
185
84
0
753
Max Planck
0
22
63
14
134
36
421
49
44
51
45
30
164
0
27
32
7
51
0
93
83
0
57
0
148
84
0
498
Year 2050 Climate
GFDL 89
0
60
124
40
155
34
359
54
46
51
107
35
130
0
37
38
15
55
0
121
110
0
40
0
123
110
0
460
UKMO
0
138
164
47
194
73
583
81
58
90
64
39
271
0
45
50
17
86
0
156
128
1
29
0
215
116
0
999
Max Planck
0
33
131
21
160
59
550
67
53
78
44
32
256
0
29
41
6
69
0
105
116
1
49
0
186
116
0
727
March 2001
4E-2
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4E-1 (cont'd). MODELED PROJECTIONS OF DIRECT HUMAN HEALTH
IMPACTS OF CLIMATE CHANGE: ESTIMATED TOTAL EXCESS MORTALITY
IN U.S. URBAN AREAS FOR AN AVERAGE SUMMER SEASON,
ASSUMING FULL ACCLIMATIZATION3
Year 2020 Climate
SMSA
Newark
Philadelphia
Phoenix
Pittsburgh
Portland
Providence
Riverside
Salt Lake City
San Antonio
San Diego
San Francisco
San Jose
Seattle
St. Louis
Tampa
Washington, DC
Total
Present
climate
26
129
0
39
9
47
4
0
4
0
28
0
5
79
28
0
1,840
GFDL 89
83
99
0
32
13
39
6
0
0
0
24
0
1
149
68
0
1,981
UKMO
173
362
0
66
22
80
10
0
0
0
23
0
0
173
95
0
4,128
Max Planck
111
191
0
64
11
52
6
0
0
0
23
0
2
158
28
0
2,799
Year 2050 Climate
GFLD 89
150
246
0
61
23
73
8
0
0
0
18
0
0
212
95
0
3,790
UKMO
127
477
0
83
31
96
11
0
0
0
24
0
0
155
100
0
4,748
Max Planck
161
323
0
95
14
74
7
0
0
0
23
0
1
189
47
0
3,863
"Abbreviations: SMSA, standard metropolitan statistical area; GFDL, Geophysical Fluid Dynamics Laboratory
Model; UKMO, United Kingdom Meteorological Office Model; Max Planck, Max Planck Institute Model.
Values given are estimated excess deaths.
Source: Kalkstein and Green (1997).
March 2001
4E-3
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4E-2. MODELED PROJECTIONS OF DIRECT HUMAN HEALTH
IMPACTS OF GLOBAL CLIMATE CHANGE: ESTIMATED TOTAL EXCESS
MORTALITY IN U.S. URBAN AREAS FOR AN AVERAGE WINTER SEASON,
ASSUMING FULL ACCLIMATIZATION3
Year 2020 climate
SMSA
Anaheim
Atlanta
Baltimore
Birmingham
Boston
Buffalo
Chicago
Cincinnati
Cleveland
Columbus
Dallas
Denver
Detroit
Ft. Lauderdale
Greensboro
Hartford
Houston
Indianapolis
Jacksonville
Kansas City
Los Angeles
Louisville
Memphis
Miami
Minneapolis
Nassau
Present
Climate
2
37
0
25
0
7
2
0
2
12
32
9
34
36
0
0
24
16
0
12
100
16
23
46
0
24
GFDL 89
0
53
0
12
0
18
4
0
9
1
41
10
15
4
0
0
33
32
0
51
102
12
20
35
0
21
UKMO
0
48
0
8
0
8
3
0
10
2
33
11
20
4
0
0
29
28
0
36
78
17
17
35
0
4
Max Planck
1
52
0
11
0
17
4
0
10
1
43
10
15
5
0
0
35
33
0
46
100
12
19
37
0
20
Year 2050 climate
GFDL 89
0
50
0
11
0
5
4
0
15
3
36
11
18
3
0
0
29
34
0
42
77
19
19
32
0
5
UKMO
0
47
0
7
0
5
2
0
10
2
31
11
25
3
0
0
27
28
0
35
88
15
15
32
0
3
Max Planck
0
52
0
12
0
18
5
0
9
1
41
11
14
5
0
0
33
32
0
46
81
12
19
36
0
21
March 2001
4E-4
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4E-2 (cont'd). MODELED PROJECTIONS OF DIRECT HUMAN HEALTH
IMPACTS OF GLOBAL CLIMATE CHANGE: ESTIMATED TOTAL EXCESS
MORTALITY IN U.S. URBAN AREAS FOR AN AVERAGE WINTER SEASON,
ASSUMING FULL ACCLIMATIZATION3
Year 2020 climate
SMSA
New Orleans
New York
Newark
Philadelphia
Phoenix
Pittsburgh
Portland
Providence
Riverside
Salt Lake City
San Antonio
San Diego
San Francisco
San Jose
Seattle
St. Louis
Tampa
Washington, DC
Total
Present
Climate
52
102
48
85
26
19
17
27
10
5
9
17
85
o
3
13
50
21
19
1,067
GFDL 89
56
123
23
80
25
20
15
21
26
7
10
26
39
2
40
61
26
31
1,104
UKMO
51
150
8
14
26
29
12
34
29
9
6
16
30
4
45
68
24
38
984
Max Planck
54
120
20
73
25
21
15
33
26
8
11
24
42
2
37
60
26
30
1,098
Year 2050 climate
GFDL 89
51
152
10
36
26
24
12
35
27
8
5
16
30
o
3
46
53
22
20
989
UKMO
47
93
6
9
27
31
10
36
27
10
4
18
21
5
47
61
20
35
894
Max Planck
54
121
23
82
26
21
13
21
26
9
9
16
26
4
43
61
25
31
1,059
Abbreviations: SMSA, standard metropolitan statistical area; GFDL, Geophysical Fluid Dynamics Laboratory
Model; UKMO, United Kingdom Meteorological Office Model; Max Planck, Max Planck Institute Model.
Values given are estimated excess deaths.
REFERENCE
Kalkstein, L. S.; Greene, J. S. (1997) An evaluation of climate/mortality relationships in large U.S. cities and the
possible impacts of a climate change. Environ. Health Perspect. 105: 84-93.
March 2001
4E-5
DRAFT-DO NOT QUOTE OR CITE
-------
i 5. HUMAN EXPOSURE TO PARTICULATE MATTER
2 AND ITS CONSTITUENTS
3
4
5 5.1 INTRODUCTION
6 5.1.1 Purpose
7 Exposure is defined as the contact by an individual with a pollutant for a specific duration
8 of time at a visible external boundary (modified from Duan 1982, 1991). For airborne particulate
9 matter (PM), the breathing zone is considered the point of contact and the lung is the external
10 boundary of concern. An individual's exposure is measured as the PM air concentration in
11 his/her breathing zone over time. Understanding exposure is important, because it is individuals
12 who experience adverse health effects associated with elevated PM concentrations in ambient air.
13 The U.S. Environmental Protection Agency's (EPA's) regulatory authority for PM applies
14 primarily to ambient air and those sources that contribute to ambient PM air concentrations.
15 Thus, a major emphasis must be to develop an understanding of exposure to PM from ambient
16 sources. However, personal exposure to total PM may result from exposure to PM from both
17 ambient and nonambient sources. Therefore, it will be necessary to account for both in order to
18 fully understand the relationship between PM and health effects. Personal exposure to PM from
19 nonambient sources may be a confounder in community-based epidemiological studies in which
20 ambient PM measures are correlated with community health parameters. In addition, an
21 individual's personal exposure to ambient, nonambient, and total PM would provide useful
22 information for studies where health outcomes are tracked individually.
23 The overall purpose of this chapter is to provide current exposure information that will aid
24 in the understanding and interpretation of PM dosimetry, toxicology, and epidemiology studies
25 assessed in later chapters. The specific objectives of this chapter, which are described below, are
26 fourfold.
27 (1) To provide an overall conceptual framework of exposure science as applied to PM, including
28 the identification and evaluation of factors that determine personal exposure to total PM and
29 to PM from ambient and nonambient PM sources
March 2001 5-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 (2) To provide a concise summary and review of recent data (since 1996) and findings from
2 pertinent studies of personal exposure to total PM and to PM from ambient and nonambient
3 sources
4 (3) To characterize quantitative relationships between ambient air quality measurements (mass,
5 chemical components, number, etc.), as determined by a community monitoring site, and
6 total personal PM exposure as well as its ambient and nonambient components
7 (4) To evaluate the implications of using ambient PM concentrations as a surrogate for exposure
8 in epidemiological studies of PM health effects
9
10 5.1.2 Particulate Matter Mass and Constituents
11 Current EPA PM regulations are based on mass as a function of aerodynamic size.
12 However, EPA also measures the chemical composition of PM in both monitoring and research
13 studies. The composition of PM is variable and adverse health effects may be related to PM
14 characteristics other than mass. PM from ambient and nonambient sources also may have
15 differing physical and chemical characteristics and differing health effects. Ultimately, to
16 understand and control health impacts caused by PM, it is important to quantify and understand
17 exposure to those chemical constituents responsible for the adverse health effects. The National
18 Research Council (NRC) recognized the distinction between measuring exposure to PM mass
19 and to chemical constituents when setting Research Priorities for Airborne Particulate Matter I:
20 Immediate Priorities and a Long-range Research Portfolio (NRC, 1998). Specifically, NRC
21 Research Topic 1 recommends evaluating the relationship between outdoor measures versus
22 actual human exposure for PM mass. The NRC Research Topic 2 recommends evaluating
23 exposures to biologically important constituents and specific characteristics of PM that cause
24 responses in potentially susceptible subpopulations and the general population. It also was
25 recognized by the NRC that, "a more targeted set of studies under this research topic (#2) should
26 await a better understanding of the physical, chemical, and biological properties of airborne
27 particles associated with the reported mortality and morbidity outcomes" (NRC, 1999). The
28 NRC also stated that the studies "should be designed to determine the extent to which members
29 of the population contact these biologically important constituents and size fraction of concern in
30 outdoor air, outdoor air that has penetrated indoors, and air pollutants generated indoors" (NRC,
31 1999). Thus, when biologically important constituents are identified, exposure studies should
March 2001 5-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 include contributions from all sources. The emphasis in this chapter on PM mass reflects the
2 current state of the science. Where available, data also have been provided on chemical
3 constituents, although in most cases, the data are limited. As recognized by the NRC, a better
4 understanding of exposures to chemical constituents will be required to more fully identify,
5 understand, and control those sources of PM with adverse health effects and to accurately define
6 the relationship between PM exposure and health outcomes.
7
8 5.1.3 Relationship to Past Documents
9 Early versions of PM criteria documents did not emphasize total human exposure but rather
10 focused almost exclusively on outdoor air concentrations. For instance, the 1969 Air Quality
11 Criteria for Particulate Matter (PM AQCD) (National Air Pollution Control Administration,
12 1969) did not discuss either exposure or indoor concentrations. The 1982 PM AQCD (U.S.
13 Environmental Protection Agency, 1982) provided some discussion of indoor PM concentrations,
14 reflecting an increase in microenvironmental and personal exposure studies. The new data
15 indicated that personal activities, along with PM generated by personal and indoor sources (e.g.,
16 cigarette smoking), could lead to high indoor levels and high personal exposures to total PM.
17 Some studies reported indoor concentrations that exceeded PM concentrations found in the air
18 outside the monitored microenvironments or at nearby monitoring sites.
19 Between 1982 and 1996, many more studies of personal and indoor PM exposure
20 demonstrated that, in most inhabited domestic environments, indoor PM concentrations and
21 personal PM exposures of the residents were greater than ambient PM concentrations measured
22 simultaneously (e.g., Sexton et al., 1984; Spengler et al., 1985; Clayton et al., 1993). As a result,
23 the NRC (1991) recognized the potential importance of indoor sources of contaminants
24 (including PM) in causing adverse health outcomes.
25 The 1996 AQCD (U.S. Environmental Protection Agency, 1996) reviewed the human PM
26 exposure literature through early 1996. Many of the studies cited showed poor correlations
27 between personal exposure or indoor measurements of PM and outdoor or ambient site
28 measurements. Conversely, Janssen et al. (1995) and Tamura et al. (1996a) showed that in the
29 absence of major nonambient sources, total PM exposures to individuals tracked through time
30 were highly correlated with ambient PM concentrations. Analyses of these latter two studies led
31 to consideration of ambient and nonambient exposures as separate components of total personal
March 2001 5-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 exposure. As a result, the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996), for
2 the first time, distinguished between ambient and nonambient PM exposure. This chapter builds
3 on the work of the 1996 PM AQCD by further evaluating the ambient and nonambient
4 components of PM, as well as reporting research that evaluates the relationship between ambient
5 concentrations and total, ambient, and nonambient personal exposure.
6
7
8 5.2 STRUCTURE FOR THE CHAPTER
9 The chapter is organized to provide information on the principles of exposure, review the
10 existing literature, and summarize key findings and limitations in the information; the specific
11 sections are described below.
12 • Section 5.3 discusses the basic concepts of exposure, including definitions, methods for
13 estimating exposure, and methods for estimating ambient components of exposure.
14 • Section 5.4 presents PM mass data, including a description of the key available studies,
15 correlations of PM exposures with ambient concentrations, and factors that effect the
16 correlations.
17 • Section 5.5 presents data on PM constituents, including a description of the key available
18 studies, correlations with ambient concentrations, and factors that effect the correlations.
19 • Section 5.6 discusses the implications of using ambient PM concentrations in epidemiological
20 studies of PM health effects.
21 • Section 5.7 summarizes key findings and limitations of the information.
22
23
24 5.3 BASIC CONCEPTS OF EXPOSURE
25 5.3.1 Components of Exposure
26 The total exposure of an individual over a discrete period of time includes exposures to
27 many different particles from various sources while in different microenvironments Cue's). Duan
28 (1982) defined a microenvironment as "a [portion] of air space with homogeneous pollutant
29 concentration." It also has been defined as a volume in space, for a specific time interval, during
30 which the variance of concentration within the volume is significantly less than the variance
March 2001 5-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 between that jue and surrounding jue's (Mage, 1985). In general, people pass through a series of
2 jue's, including outdoor, in-vehicle, and indoor jue's, as they go through time and space. Thus,
3 total daily exposure for a single individual to PM can be expressed as the sum of various
4 microenvironmental exposures that the person occupies in the day (modified from National
5 Research Council, 1991).
6 In a given (j,e, particles may originate from a wide variety of sources. For example, in an
7 indoor (j,e, PM may be generated by (1) indoor activities, (2) outdoor PM entering the indoor (j,e,
8 (3) the chemical interaction of outdoor air pollutants and indoor air or indoor sources,
9 (4) transport from another indoor jue, or (5) personal activities. All of these disparate sources
10 have to be accounted for in a total human PM exposure assessment.
11 An analysis of personal exposure to PM mass (or constituent compounds) requires
12 definition and discussion of several classes of particles and exposure. In this chapter, PM
13 metrics may be described in terms of exposure or as an air concentration. PM also may be
14 described according to both its source (i.e., ambient, nonambient) and the microenvironment
15 where exposure occurs. Table 5-1 provides a summary of the terms used in this chapter, the
16 notation used for these terms, and their definition. These terms will be used throughout this
17 section and will provide the terminology for evaluating personal exposure to total PM and PM
18 from ambient and nonambient sources.
19
20 5.3.2 Methods To Estimate Personal Exposure
21 Personal exposure may be estimated using either direct or indirect approaches. Direct
22 approaches measure the contact of the person with the chemical concentration in the exposure
23 media over an identified period of time. Direct measurement methods include personal exposure
24 monitors (PEMs) for PM that are worn continuously by individuals as they encounter various
25 microenvironments and perform their daily activities. Indirect approaches use available
26 information on concentrations of chemicals in microenvironments, along with information about
27 the time individuals spend in those microenvironments and personal PM generating activities.
28 The indirect approach then uses models and data on microenvironmental air concentrations and
29 time spent in microenvironments to estimate personal exposure. This section describes the
30 methods to directly measure personal exposures and microenvironmental concentrations, as well
March 2001 5-5 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-1. CLASSES OF PARTICULATE MATTER EXPOSURE AND
CONCENTRATION DEFINITIONS
Term
Notation
Definition
Concentration
Personal Exposure
Microenvironment
Ambient PM
Personal Exposure to
Personal-Activity PM
Personal Exposure to
Nonambient PM
Personal Exposure to
Total PM
Ate
Ambient-Outdoor PM Cao
Indoor PM Q
Ambient-Indoor PM C^
Indoor-Generated PM CL
Personal Exposure to Elg
Indoor-Generated PM
Personal Exposure to Eag
Ambient-Generated PM
-'-'no
General Definitions
Air concentration of PM in a given microenvironment, expressed in
Contact at visible external boundaries of an individual with a pollutant
for a specific duration of time; quantified by the amount of PM available
in concentration units Cwg/m3) at the oral/nasal contact boundary for a
specified time period (At). General term for any exposure variable.
Volume in space, for a specific time interval, during which the variance
of concentration within the volume is significantly less than the variance
between that Ate and surrounding Ates
Concentration Variables
PM in the atmosphere measured at a community ambient monitoring site
either emitted into the atmosphere directly (primary PM) or formed in it
(secondary PM). Major sources of PM species are industry, motor
vehicles, commerce, domestic emissions such as wood smoke, and
natural wind-blown dust or soil.
Ambient PM in an outdoor microenvironment
All PM found indoors
Ambient PM that has infiltrated indoors (i.e., has penetrated indoors and
remains suspended)
PM generated or formed indoors
Exposure Variables
Sum of personal exposure resulting from indoor-generated PM
Sum of personal exposure caused by ambient-outdoor and ambient
indoor PM (does not include resuspended ambient PM previously
deposited indoors)
Small-scale PM-generating activities that primarily influence exposure of
the person performing the activity itself
Sum of personal exposure to indoor-generated and personal activity PM
nonag ig pact
Sum of all personal exposures to ambient and nonambient PM
F = F+F +F=F +F
-^t J-'jg ^ -'-'pact ^ -'-'ag -'-'nonag J-'ag
1 as the models used to estimate exposure. Several approaches to estimate personal exposure to
2 ambient PM also are described.
March 2001
5-6
DRAFT-DO NOT QUOTE OR CITE
-------
1 5.3.2.1 Direct Measurement Methods
2 5.3.2.1.1 Personal Exposure Monitoring Methods
3 In theory, personal exposure to total PM is measured by sampling the concentration of PM
4 in inhaled air entering the nose or mouth. Practically, it is defined as that PM collected by a
5 PEM worn by a person and sampling from a point near the breathing zone (but not impacted by
6 exhaled breath). The inlet to a PEM normally is placed at the outer limit of the breathing zone to
7 avoid a negative sampling bias resulting from dilution of the collected air by exhaled breath
8 depleted of PM. However, such placement does not allow for the sampling of directly inhaled
9 cigarette smoke or inhaled air that passes through a dust mask. PEMs for PM use measurement
10 techniques similar to those used for ambient PM. The PEM is a filter-based mass measurement
11 of a particle size fraction (PM10 or PM2 5), usually integrated over either a 24- or 12-h period at
12 flow rates of 2 to 4 L/min using battery-operated pumps. PEMs must be worn by study
13 participants and, therefore, they must be quiet, compact, and battery-operated. These
14 requirements limit the type of pumps and the total sample volume that can be collected.
15 Generally, small sample volumes limit personal exposure measurements to PM mass and a few
16 elements detected by XRF. In most studies, PM25 and PM10 have not been collected
17 concurrently.
18 Other methods used for ambient PM also have been adapted for use as a personal exposure
19 monitor. For example, a personal nephelometer that measures particle number within a specific
20 particle size range using light scattering has been used in personal exposure studies to obtain
21 real-time measurements of PM.
22
23 5.3.2.1.2 Microenvironmental Monitoring Methods
24 Direct measurements of microenvironmental PM concentrations, which are used with
25 models to estimate personal exposure to PM, also use methods similar to those for ambient PM.
26 These methods differ from PEMs in that they are stationary with respect to the microenvironment
27 (such as a stationary PEM). Microenvironmental monitoring methods include filter-based mass
28 measurements of particle size fractions (PM10, PM25), usually integrated over either a 24- or 12-h
29 period. Flow rates vary between various devices from 4 to 20 L/min. Larger sample volumes
30 allow more extensive chemical characterization to be conducted on microenvironmental samples.
31 Because more than one pumping system can be used in a microenvironment, PM2 5 and PM10 can
March 2001 5-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 be collected simultaneously. Other continuous ambient PM measurement methods that have
2 been utilized for microenvironmental monitoring are the Tapered Element Oscillating
3 Microbalance (TEOM) and nephelometers. Various continuous techniques for counting particles
4 by size also have also been used (Climet, LASX, SMPS, APS). Measurement techniques are
5 discussed in Chapter 2.
6
7 5 3.2.2 Indirect Methods (Modeling Methods)
8 5.3.2.2.1 Personal Exposure Models
9 Exposure modeling for PM2 5 mass and chemical constituents is a relatively new field
10 facing significant methodological challenges and input data limitations. Exposure models
11 typically use one of two general approaches: (1) a time-series approach that estimates
12 microenvironmental exposures sequentially as individuals go through time or (2) a time-averaged
13 approach that estimates microenvironmental exposures using average microenvironmental
14 concentrations and the total time spent in each microenvironment. Although the time-series
15 approach to modeling personal exposures provides the appropriate structure for accurately
16 estimating personal exposures (Esmen and Hall, 2000; Mihlan et al., 2000), a time-averaged
17 approach typically is used when the input data needed to support a time-series model are not
18 available. In addition, the time-varying dose profile of an exposed individual can be modeled
19 only by using the time-series approach (McCurdy, 1997, 2000). We define the personal
20 exposure of an individual to a chemical in air to be (NRC, 1991)
21
E
i —i2
= \C(t}dt,
22 where
23 E is the personal exposure during the time period from tt to t2, and
24 C(t) is the concentration near the nose and mouth not impacted by
25 exhaled air, at time t.
26 In general, personal exposure models combine microenvironmental concentration data with
27 human activity pattern data to estimate personal exposures. Time-averaged models also can be
28 used to estimate personal exposure for an individual or for a defined population. Total personal
March 2001 5-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 exposure models estimate exposures for all of the different microenvironments in which a person
2 spends time, and total average personal exposure is calculated from the sum of these
3 microenvironmental exposures:
4
5 where Ey. is the personal exposure in each microenvironment, y (Duan, 1982). Example
6 microenvironments include outdoors, indoors at home, indoors at work, and in transit. Each
7 microenvironmental exposure, Ey, is calculated from the average concentration in
8 microenvironment y, C , weighted by the time spent in microenvironment j, t,. T is the sum of ty
9 over ally. It is important to note that, although measurement data may be an average
10 concentration over some time period (i.e., 24 h), significant variations in PM concentrations can
1 1 occur during that time period. Thus, an error may be introduced if real-time concentrations are
12 highly variable, and an average concentration for a microenvironment is used to estimate
13 exposure when the individual is in that microenvironment for only a fraction of the total time.
14 This exposure formulation has been applied to concentration data in a number of studies (Ott,
15 1984; Ott et al., 1988, 1992; Miller et al., 1998; Klepeis et al., 1994; Lachenmyer and Hidy,
16 2000).
17 Microenvironmental concentrations used in the exposure models can be measured directly
18 or estimated from one or more microenvironmental models. Microenvironmental models vary in
19 complexity, from a simple indoor/outdoor ratio to a multi-compartmental mass-balance model.
20 A discussion of microenvironmental models is presented below in Section 5.3.2.2.2.
21 On the individual level, the time spent in the various microenvironments is obtained from
22 time/activity diaries that are completed by the individual. For population-based estimates, the
23 time spent in various microenvironments is obtained from human activity databases. Many of
24 the largest human activity databases have been consolidated by EPA's National Exposure
25 Research Laboratory (NERL) into one comprehensive database called the Consolidated Human
26 Activity Database (CHAD). CHAD contains over 22,000 person-days of 24-h activity data from
27 11 different human activity pattern studies. Population cohorts with diverse characteristics can
March 2001 5-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 be constructed from the activity data in CHAD and used for exposure analysis and modeling
2 (McCurdy, 2000). Table 5-2 is a summary listing of the human activity studies in CHAD.
3 Methodologically, personal exposure models can be divided into three general types:
4 (1) statistical models based on empirical data obtained from one or more personal monitoring
5 study, (2) simulation models based upon known or assumed physical relationships, and
6 (3) physical-stochastic models that include Monte Carlo or other techniques to explicitly address
7 variability and uncertainty in model structure and input data (Ryan, 1991; Macintosh et al.,
8 1995). The attributes, strengths, and weaknesses of these model types are discussed by Ryan
9 (1991), National Research Council (1991), Frey and Rhodes (1996), and Ramachandran and
10 Vincent (1999). GIS-based approaches to estimate health risks of environmental concentrations
11 also have been developed (e.g., Beyea and Hatch, 1999; Jensen, 1999). A recent summary
12 review of the logic of exposure modeling is found in Klepeis (1999).
13 Personal exposure models that have been developed for PM are summarized in Table 5-3.
14 The regression-based models (Johnson et al., 2000; Janssen et al., 1997; Janssen et al., 1998a)
15 were developed for a specific purpose (i.e., to account for the observed difference between
16 personal exposure and microenvironmental measurements) and are based on data from a single
17 study, which limits their utility for broader purposes. Other types of models in Table 5-3 were
18 limited by a lack of data for the various model inputs. For example, ambient PM monitoring data
19 is not generally of adequate spatial and temporal resolution for these models. Lurmann and Korc
20 (1994) used site-specific coefficient of haze (COH) information to stochastically develop a time
21 series of 1-h PM10 data from every sixth day 24-h PM10 measurements. A mass-balance model
22 typically was used for indoor microenvironments when sufficient data was available, such as for
23 a residence. For most other microenvironments, indoor/outdoor ratios were used because of the
24 lack of data for a mass-balance model. In addition, only the deterministic model PMEX included
25 estimation of inhaled dose from activity-specific breathing rate information. Data from recent
26 PM personal exposure and microenvironmental measurement studies will help facilitate the
27 development of improved personal exposure models for PM.
28 An integrated human exposure source-to-dose modeling system that will include exposure
29 models to predict population exposures to environmental pollutants such as PM currently is
30 being developed by NERL. A first-generation population exposure model for PM, called the
31 Stochastic Human Exposure and Dose Simulation (SHEDS-PM) model, recently has been
March 2001 5-10 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-2. ACTIVITY PATTERN STUDIES INCLUDED IN THE CONSOLIDATED HUMAN
o
to
o
o
1
o
H
6
o
0
H
O
o
H
W
O
HH
H
W
Calendar
Study Time Period
Name of the Study Age1
Baltimore Jan-Feb 1997 65+
Jul-Aug 1998
CARB: Adolescents Oct 1987- 12 - 94
and Adults Sept 1988
CARB: Children Apr 1989- 0-11
Feb 1990
Cincinnati (EPRI) Mar-Apr and 0 - 86
Aug 1985
Denver (EPA) Nov 1982- 18-70
Feb 1983
Los Angeles: Elem. Oct 1989 10 - 12
School Children
Los Angeles: High Sept-Oct 1990 13 - 17
School Adoles.
National: NHAPS-A8 Sept 1992- 0 - 93
Oct 1994
National: NHAPS-B8 As above 0 - 93
University of Feb-Decl997 0-13
Michigan: Children
Valdez, AK Nov 1990- 11-71
Oct 1991
Washington, DC Nov 1982- 18-98
(EPA) Feb 1983
AC11V11Y UALAKA
Diary
Days2 TvDe3 Time4
391 Diary; 15 -min 24h Standard
blocks
1,762 Retrospective 24h Standard
1,200 Retrospective 24h Standard
2,614 Diary 24h; nominal
7 p.m. -7 a.m.
805 Diary 24h; nominal
7 p.m.-7 a.m.
51 Diary 24h Standard
43 Diary 24h Standard
4,723 Retrospective 24h Standard
4,663 Retrospective 24h Standard
5,616 Retrospective 24h Standard
40 1 Retrospective Varying 24-h
period
699 Diary 24h; nominal
7 p.m. -7 a.m.
Notes: 'All studies included both genders. The age range depicted is for the subjects actually included; in most c
Age 0 = babies < 1 year old.
2The actual number of person-days of data in CHAD after the "flagging" and removal of questionable dati
'Retrospective: a "what did you do yesterday" type of survey; also known as an ex post survey. Diary: a "
"Standard = midnight-to-midnight.
'Was activity-specific breathing rate data collected?
6PEM = a personal monitoring study. In addition to the diary, a subject carried a small CO or PM2 5 moni
7P-D removed = The number of person-days of activity pattern data removed from consolidated CHAD bt
in the text.
8National Human Activity Pattern Study; A = the air version; B = the water version. The activity data ob1
9A question was asked regarding which activities (within each 6-h time block in the day) involved "heavy
L»U
Documentation
Rate5 or Reference Notes
No Williams et al. (2000a,b) Multiple days, varying from 5-15; part
of a PM25 PEM study
No Robinson etal. (1989)
Wiley etal. (1991b)
No Wiley etal. (1991a)
Yes Johnson (1989) 3 consecutive days; 1 86 P-D removed7
No Akland etal. (1985) Part of CO PEM6 study; 2 consec.
Johnson (1984) days; 5 5 P-D removed7
Yes Spier etal. (1992) 7 P-D removed7
Yes Spier et al. (1992) 23 P-D removed7
No9 Klepeis et al. (1995) A national random-probability survey
Tsang and Klepeis (1996)
No9 As above As above
No Institute for Social 2 days of data: one is a weekend day
Research (1997)
No Goldstein etal. (1992) 4 P-D removed7
No Akland et al. (1985) Part of a CO PEM6 study; 6 P-D
Hartwell et al. (1984) removed7
;ases, there was not an upper limit for the adult studies. Ages are inclusive.
\. See the text for a discussion of these procedures.
real-time" paper diary that a subject carried as he or she went through the day.
tor throughout the sampling period.
;cause of missing activity and location information; completeness criteria are listed
:ained on the two versions are identical.
• breathing", lifting heavy objects, and running hard.
-------
TABLE 5-3. PERSONAL EXPOSURE MODELS FOR PARTICIPATE MATTER
.
to
o
o
Study Citation Model Name Model Type
Microenvironments or Predictors
Output
Notes
Time-series Models:
Hayes and Marshall (1999) PMEX Deterministic
Johnson et al. (2000)
Klepeisetal. (1994)
Regression-based
Stochastic
Lurmann and Korc (1994) REHEX-II Stochastic
Koontz and Niang (1998) CPIEM Stochastic
to
Indoors: residential, work, school
Outdoors: near roadway, other
Motor vehicle
Auto travel, roadside, ETS, food
prep, grilling, high ambient PM
ETS, cooking, cleaning, attached
garage, wood burning
12 residential with different
sources, restaurant/bar,
nonresidential indoors, in transit,
outdoors
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
et al., 2000)
Fixed I/O ratio of 0.7 for indoors w/o sources
and 1.2 for in transit
Reduced form mass balance model for indoors
with PM sources
Used California activity pattern and breathing
rate data. Used either a mass balance model or
I/O ratio distribution for indoor
microenvironments. Indoor sources included.
Time-averaged Models:
fe
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Clayton et al. (1999a) SIM
Janssenetal. (1997)
Janssenetal. (1998a)
Ott et al. (2000)
Burke etal. (2001)
Stochastic
Regression-based
Regression-based
RCS
SHEDS-PM Stochastic
Smoking parent, ETS exposure,
outdoor physical activity
Number of cigarettes smoked,
hours of ETS exposure, residence
on busy road, time in vehicle
Not separated
Outdoors, indoors: residence,
office, stores, school, in vehicle,
restaurant/lounge,
Distribution of annual PM25
exposures
Accounts for difference between
personal and microenvironmental
PM10
Accounts for difference between
personal and microenvironmental
PM10
Distribution of PM10 exposure for
population
PM25 exposure distributions for
population, by age, gender,
smoking and employment
status; PM2 5 exposure uncertainty
predictions. Percent contribution
from PM of ambient origin to
total personal exposures
Based on 3-day ambient measurements
Children only
Adults only
A random-component superposition (RCS)
model that uses distribution of ambient PM10
and estimated nonambient PM10
concentrations
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.
-------
1 developed. The SHEDS-PM model uses a 2-stage Monte Carlo sampling technique previously
2 applied by Macintosh et al. (1995) for benzene exposures. This technique allows for separate
3 characterization of variability and uncertainty in the model predictions (to predict the distribution
4 of total exposure to PM for the population of an urban/metropolitan area and to estimate the
5 contribution of ambient PM to total PM exposure). This model is yet to be evaluated and is
6 discussed for information purposes only because results from the case study have been only
7 recently reported in a journal article submitted for peer review (Burke et al., 2001).
8
9 5.3.2.2.2 Microenvironmental Models
10 The mass balance model has been used extensively in exposure analysis to estimate PM
1 1 concentrations in indoor microenvironments (Calder, 1957; Sexton and Ryan, 1988; Duan, 1982,
12 1991; McCurdy, 1995; Johnson, 1995; Klepeis et al., 1995; Dockery and Spengler, 1981; Ott,
13 1984; Ott et al., 1988, 1992, 2000; Miller et al., 1998; Mage et al., 1999; Wilson et al., 2000).
14 The mass balance model describes the infiltration of particles from outdoors into the indoor
15 microenvironment and the generation of particles from indoor sources:
16
17
V dC; /dt = vPCa -vQ-kVCi + Qi, (5-2)
18
19
20 where V = volume of the well-mixed indoor air (cubic meters),
21 Q = concentration of indoor PM;
22 v = volumetric air exchange rate between indoors and outdoors (cubic
23 meters per hour);
24 P = penetration ratio, the fraction of ambient (outdoor) PM that is not
25 removed from ambient air during its entry into the indoor volume;
26 Ca = concentration of PM in the ambient air (micrograms per cubic meter);
27 k = removal rate (per hour); and
28 Q; = indoor sources of particles (micrograms per hour).
29
30 Q; contains a variety of indoor, particle-generating sources, including combustion or
3 1 mechanical processes, condensation of vapors formed by combustion or chemical reaction,
March 2001 5-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 suspension from bulk material, and resuspension of previously deposited PM. The removal rate,
2 k, includes dry deposition to interior surfaces by diffusion, impaction, electrostatic forces, and
3 gravitational fallout. It may include other removal processes such as filtration by forced air
4 heating, ventilation, or air-conditioning (HVAC) or by independent air cleaners. All parameters
5 except V are functions of time. P and k also are functions of particle aerodynamic diameter
6 and v.
7 In addition to the mass balance model, a number of single-source or single-
8 microenvironment models exist. However, most are used to estimate personal exposures to
9 environmental tobacco smoke (ETS). These models include both empirically based statistical
10 models and physical models based on first principles; some are time-averaged, whereas others
11 are time-series. These models evaluate the contribution of ETS to total PM exposure in an
12 enclosed microenvironment and can be applied as activity-specific components of total personal
13 exposure models. Examples of ETS-oriented personal exposure models are Klepeis (1999),
14 Klepeis et al. (1996, 2000), Mage and Ott (1996), Ott (1999), Ott et al. (1992, 1995), and
15 Robinson et al. (1994).
16
17 5.3.2.3 Methods of Estimating Personal Exposure to Ambient Particulate Matter
18 In keeping with the various components of PM exposure described above in Section 5.3.1,
19 personal exposure to PM can be expressed as the sum of exposure to particles from different
20 sources summed over all microenvironments in which exposure occurs. Total personal exposure
21 may be expressed as
22
t/t = t/ag + t/ig + t/pact
T3 - T3 + T3 (5'3)
C/i — -t/ag ~r _t/nonag,
23
24 where Et is the total personal exposure to ambient and nonambient PM, Eag is personal exposure
25 to ambient PM (the sum of ambient PM while outdoors and ambient PM that has infiltrated
26 indoors, while indoors), Eig is personal exposure to indoor-generated PM, Epact is personal
27 exposure to PM from personal activity, and Enonag is personal exposure to nonambient PM.
28 Although personal exposure to ambient and nonambient PM cannot be measured directly, they
March 2001 5-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 can be calculated or estimated from other measurement data. Approaches for estimating these
2 components of PM exposure are described in the following section.
3
4 5.3.2.3.1 Mass Balance Approach
5 Ambient-Indoor Concentrations of Particulate Matter
6 The mass balance model described above (Equation 5-2) has been used to estimate PM
7 concentrations in indoor microenvironments. This model also may be used to estimate ambient-
8 indoor (Cai) and indoor-generated (Cig) PM concentrations. The mass balance model can be
9 solved for Cai and Cig assuming equilibrium conditions, and assuming that all variables remain
10 constant (Ott et al., 2000; Dockery and Spengler, 1981; Koutrakis et al., 1992). By substituting
11 dCai + dCig for dC; in equation 5-2 and assuming dCai and dCig = 0, ambient-indoor PM (Cai) and
12 indoor-generated PM (Cig), at equilibrium, are given by
13
14
Cai=(C/aoPa)//(a+k) (5-4)
(5-5)
15 where a = v/V, the number of air exchanges per hour. Equations 5-4 and 5-5 assume equilibrium
16 conditions and, therefore, are valid only when the parameters &, a, Cao, and Q; are not changing
17 rapidly and when the Cs are averaged over several hours. Under certain conditions (e.g.,
18 air-conditioned homes, homes with HVAC or air cleaners that cycle on and off, ambient
19 pollutants with rapidly varying concentrations), nonequilibrium versions of the mass balance
20 model (Ott et al., 2000; Freijer and Bloeman, 2000; Isukapalli and Georgopoulos, 2000) are
21 likely to provide a more accurate estimate of Cai and Cig. However, the equilibrium model
22 provides a useful, if simplified, example of the basic relationships (Ott et al., 2000).
23 Equation 5-4 may be rearranged further to give Cai/Cao, the equilibrium fraction of ambient
24 PM that is found indoors, defined as the infiltration factor (FINF) (Dockery and Spengler, 1981).
25
March 2001 5-15 DRAFT-DO NOT QUOTE OR CITE
-------
C Pa
FlNF = 7^ = ^7? (5'6)
^ao a + K
1
2 The penetration ratio (P) and the decay rate (k) can be estimated using a variety techniques.
3 A discussion of these variables and estimation techniques is given in Section 5.4.3.2.2. Because
4 both P and k are a function of particle aerodynamic diameter, FINF also will be a function of
5 particle aerodynamic diameter.
6
7 Personal Exposure to Ambient-Generated Particulate Matter
8 Personal exposure to ambient-generated PM (Eag) may be estimated using ambient-indoor
9 PM concentration (Cai) from the mass balance model, ambient outdoor PM concentrations (Cao)
10 and information on the time an individual spent in the various microenvironments.
11 Mathematically, this may be expressed as
12
13 Eag = yCao + (1 - j)Cai
Pa
(5-7)
(. 1 \ ? V J
a + k)
14
15 where y is the fraction of time that an individual spent outdoors, and (1 -y) is the fraction of time
16 spent indoors.
17 It is convenient to express personal exposure to ambient generated PM (Eag) as the product
18 of the ambient PM concentration (Cao or Ca) and a personal exposure or attenuation factor.
19 Following the usage in several recent papers (Zeger et al., 2000; Dominici et al., 2000; Ott et al.,
20 2000), the symbol a will be used for this attenuation factor. Equation 5-7 can be rearranged to
21 obtain an expression for a:
22
March 2001 5-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Substituting equation 5-6 in equation 5-8 gives a relationship for a in terms of the infiltration
2 factor FINF and the fraction of time spent in the various microenvironments:
3
a = y+(1- y) FINF. (5-9)
4
5 Thus, personal exposures to ambient PM (Eag) may be calculated from measurable quantities:
6
7 Eag = aCao. (5-10)
8
9 The factor a can be measured directly or calculated from measured or estimated values of the
10 parameters a, k, and P and the time spent in various microenvironments from activity pattern
11 diaries (Wilson et al., 2000).
12 The use of a mass balance model to separate personal exposure into two components
13 because of exposure to ambient and nonambient concentrations is not novel. This approach,
14 based on Equation 5-3 as given in Duan (1982) and called superposition of component
15 concentrations, has been applied using multiple microenvironments to carbon monoxide (Ott,
16 1984; Ott et al., 1988, 1992), volatile organic compounds (Miller et al., 1998), and particles
17 (Koutrakis et al., 1992; Klepeis et al., 1994). However, in these studies, and in most of the
18 exposure literature, the ambient and nonambient components are added to yield a personal
19 exposure from all sources of the pollutant. The use of the mass balance model, ambient
20 concentrations, and exposure parameters to estimate exposure to ambient-generated PM and
21 exposure to indoor-generated PM separately as different classes of exposure has been discussed
22 in Wilson and Suh (1997) and in Wilson et al. (2000).
23
24 5.3.2.3.2 Tracer Species as Surrogates of Ambient-Generated Particulate Matter
25 The ratio of personal exposure to ambient concentration for a PM component that has no
26 indoor sources may be used as a measure of the ratio of personal exposure to ambient PM to the
27 ambient concentration of PM for PM of similar aerodynamic diameter (Wilson et al., 2000).
28 Sulfate, in particular, often is used as a marker of outdoor air in indoor microenvironments
March 2001 5-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 (Jones et al., 2000). It is found primarily in the PM2 5 fraction of the aerosol (Cohen et al., 2000).
2 Ozkaynak et al. (1996a, b) and Janssen et al. (1999a) report, in the PTEAM and Netherlands
3 studies respectively, that XRF analyses of indoor PM and the immediate outdoor PM show that
4 sulfur is the only element reported with virtually identical mass concentrations in both indoor and
5 outdoor air. Therefore, where there are no indoor sources of fine-mode sulfates, one may deduce
6 that the ambient-to-personal relationship found for sulfates probably would be the same as that
7 for unspeciated particulate matter of the same aerodynamic size range. This assumption has not
8 been validated, however, and ambient PM with different physical or chemical characteristics may
9 not behave similarly to sulfate.
10 Parti culate sulfate is formed in the ambient air via photochemical oxidation of gaseous
11 sulfur dioxide arising from the primary emissions from the combustion of fossil fuels containing
12 sulfur. They also arise from the direct emissions of sulfur-containing particles from
13 nonanthropogenic sources (e.g., volcanic activity, wind-blown soil). In the indoor environment,
14 the only common sources of sulfate may be resuspension by human activity of deposited PM
15 containing ammonium sulfates or soil sulfates that were tracked into the home. In some homes
16 an unvented kerosene heater using a high-sulfur fuel may be a major contributor during winter
17 (Leaderer et al., 1999). Use of matches to light cigarettes or gas stoves are also a source of
18 sulfates. Studies that have used sulfate as a surrogate for ambient PM are discussed in
19 Section 5.4.3.1 (i.e., Oglesby et al., 2000a; Sarnat et al., 2000; Ebelt, 2000).
20
21 5.3.2.3.3 Source-Apportionment Techniques
22 Source apportionment techniques provide a method for determining personal exposure to
23 PM from specific sources. If a sufficient number of samples are analyzed with sufficient
24 compositional detail, it is possible to use statistical techniques to derive source category
25 signatures, identify indoor and outdoor source categories, and estimate their contribution to
26 indoor and personal PM. Daily contributions from sources that have no indoor component can
27 be used as tracers to generate exposure to ambient PM of similar aerodynamic size or directly as
28 exposure surrogates in epidemiologic analyses. Studies that have used source-apportionment are
29 discussed in Section 5.4.3.3 (i.e., Ozkaynak and Thurston, 1987; Yakovleva et al., 1999; Mar
30 et al. 2000; Laden et al., 2000).
31
March 2001 5-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.4 SUMMARY OF PARTICULATE MATTER MASS DATA
2 5.4.1 Types of Particulate Matter Measurement Studies
3 A variety of field measurement studies have been conducted to quantify personal exposure
4 to PM mass, measure microenvironmental concentrations of PM, and evaluate the relationship
5 between personal exposure to PM and PM air concentrations measured at ambient sites.
6 In general, exposure measurement studies are of two types depending on how the participants are
7 selected for the study. In ^probability study, participants are selected using a probability
8 sampling design where every member of the defined population has a known, positive probability
9 of being included into the sample. Probability study results can be used to make statistical
10 inferences about the target population. In & purposeful or nonprobability design, any convenient
11 method may be used to enlist participants and the probability of any individual in the population
12 being included in the sample is unknown. Participants in purposeful samples (also referred to as
13 a "convenience" samples) may not have same the characteristics that would lead to exposure as
14 the general population. Thus, results of purposeful studies apply only to the subjects sampled on
15 the days that they were sampled. In a purposeful study, statistically valid inferences cannot be
16 made to any other population or period of time. Although such studies may report significant
17 differences, confidence intervals, and/? values, they have no inferential validity (Lessler and
18 Kalsbeek,1992). However, most purposeful studies of PM personal exposure can provide data to
19 develop relationships on important exposure factors and useful information for developing and
20 evaluating either statistical or physical/chemical human exposure models.
21 Regardless of the sampling design (probability or purposeful) there are three general
22 categories of study design that can be used to measure personal exposure to PM and evaluate the
23 relationship between personal PM exposure levels and ambient PM concentrations measured
24 simultaneously: (1) longitudinal, (2) daily-average, and (3) pooled. These are discussed in
25 Section 5.4.3.1.1.
26
27 5.4.2 Available Data
28 5.4.2.1 Personal Exposure Data
29 Table 5-4 gives an overview of the personal exposure studies that have been conducted and
30 are reviewed in this section. This includes studies that have been reported since the 1996 AQCD.
March 2001 5-19 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-4. SUMMARY OF RECENT PERSONAL EXPOSURE STUDIES
tr
to
o
o
1
to
o
o
^d
fe
H
6
o
2|
0
H
O
Cj
O
H
W
O
^
o
H
W
Study Design
Pooled
Pooled
Pooled
Pooled
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Study Location and
Population
Riverside, CA,
PTEAM
Basel, Switzerland,
EXPOLIS
Toronto, Canada
Mexico City
Wageningen,
Netherlands, school
children
Amsterdam (Am),
Helsinki (His), elderly
angina or coronary
heart disease
Baltimore, elderly
healthy and COPD
Fresno I
Fresno II (elderly)
Los Angeles, elderly
COPD subjects
Boston, COPD
subjects
Nashville, TN, COPD
subjects
Vancouver, British
Columbia, COPD
No. of
Subjects
178
50
732
66
13
41 (Am)
49 (His)
21
5
16
30
18
10
16
Study Period
Fall 1990
1997
9/ 1995 -
8/1996
1992
1995
Winter 1998
Spring 1999
7-8/1998
Feb, 1999
Apr-May 1999
Summer/ Fall
1996
Winter 1996-7
Summer 1996
Summer 1995
April-Sept,
1998
Days per
Age Subject
Probability
10-70 l(12h)
1 (48 h)
16+ 3
<65 1
Purposeful
10-12 6
50-84 22 (Am)
27 (His)
72-93 5-22
60+ 24
24
56-83 4
12
36-88 6
54-86 7
PM Exposure"
Metrics
Studies
P, I, O, A
P, I, P
P, I, O, A
P, I, O
Studies
P, A, School
P, I, O
P, I, O , A
P, I, O, A
P, I, O, A
P, I, O
P, I, O, A
P, I, O
P, A
Co-Pollutant
PM Size Measured1" Metrics
PM10
PM25 VOC, CO,
NO2, S, K, Pb,
Br, Ca
PM25 (12 mo)
PM10 (3 mo)
PM10
PM2 5, PM10
PM25
PM25, PM10 CO, O3, NO2,
SO2
PM25, PM10 CO, O3
PM2 5, PM10
PM2,
PM2.5,PM10
PM2.5,PM10
PM2.5, PM10
Reference
Clayton etal. (1993)
Ozkaynak et al. (1996a,b)
Oglesby et al. (2000a)
Jantunen etal. (1998)
Clayton et al. (1999a),
Pellizzari etal. (1999)
Santos-Burgoa et al. (1998)
Janssenetal. (1999b)
Janssen etal. (2000)
Williams et al. (2000a,b)
Evans et al. (2000)
Linn etal. (1999)
Rojas-Bracho et al. (2000)
Bahadori etal. (2001)
Ebelt et al. (2000)
-------
TABLE 5-4 (cont'd). SUMMARY OF RECENT PERSONAL EXPOSURE STUDIES (since 1996)
*-"
g.
to
o
o
L/l
1
Study Design
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Pooled
Study Location and No. of Days per PM Exposure' Co-Pollutant
Population Subjects Study Period Age Subject Metrics PM Size Measured1" Metrics
Purposeful Studies (cont'd)
Amsterdam and 45 1994, 1995 10-12 4-8 P, A, School PM10
Wageningen, Neth.,
school children
Amsterdam, adults 37 1994 51-70 5-8 P, I, A PM10
Baltimore, elderly 15 Summer 1998, 75 ± 6.8 12 P PM25,PM10 O3, NO2, SO2
subjects Spring 1999 VOCs
Tokyo, Japan, 18 1992 3 P, I, O, A SPM NO2
elderly housewives
Osaka, Japan 26 Fall 1990-1995 Multiple P, I, O PM2,PM2.10,PM>10
days
Milan, Italy, office 100 Spring/summer 1 P, Home, PM10 NO2, CO,
workers and winter Office, VOCs
Commuting
Reference
Janssenetal. (1997)
Janssenetal. (1998a)
Sarnat et al. (2000)
Tamuraetal. (1996a)
Tamuraetal. (1996b)
Carreretal. (1998)
"All based on gravimetric measurements.
O
£
H
I
0
o
0
H
O
o
H
W
O
O
H
W
bP = personal,
I = indoors, O = outdoors, A = ambient.
-------
1 Major studies that were reported before that time also have been included to provide a
2 comprehensive evaluation of data in this area. Table 5-4 gives information on the sampling and
3 study designs, the study population, the season, number of participants, PM exposure metric, and
4 the PM size fraction measured.
5 Although there are a number of studies listed in the table, the data available to answer the
6 important questions related to exposure are limited. Few are based on probability sampling
7 designs that allow study results to be inferred to the general population. Unfortunately, none of
8 these probability studies uses a longitudinal study design. This limits our ability to provide
9 population estimates on the relationship between personal PM exposures and ambient site
10 measurements. In addition, most of the probability studies of PM exposure were conducted
11 during a single season, thus variations in ambient concentrations, air exchange rates, and
12 personal activities are not accounted for across seasons. In these cases, study results are only
13 applicable to a specific time period. Longitudinal studies, on the other hand, generally have
14 small sample sizes and use a purposeful sampling design. Many of these studies did not include
15 ambient site measurements to allow comparisons with the exposure data, and approximately half
16 of these studies monitored PM2 5.
17 Four large-scale probability studies that quantify personal exposure to PM under normal
18 ambient source conditions have been reported in the literature. These include the EPA's Particle
19 Total Exposure Assessment Methodology (PTEAM) study (Clayton et al., 1993; Ozkaynak et al.,
20 1996a,b); the Toronto, Ontario, study (Clayton et al., 1999a and Pellizzari et al., 1999); the Air
21 Pollution Exposure Distribution within Adult Urban Populations in Europe (EXPOLIS) exposure
22 study (Jantunen et al., 1998, 2000; Oglesby, et al., 2000); and a study of a small, highly polluted,
23 area in Mexico City (Santos-Burgoa et al., 1998). Only preliminary results have been reported
24 for the EXPOLIS study. A fifth study conducted in Kuwait during the last days of the oil-well
25 fires (Al-Raheem et al., 2000) is not reported here because the ambient PM levels were not
26 representative of normal ambient source conditions.
27 Recent longitudinal exposure studies have focused on potentially susceptible
28 subpopulations such as the elderly with preexisting respiratory and heart diseases (hypertension,
29 chronic obstructive pulmonary disease, and congestive heart disease). This is in keeping with air
30 pollution analyses that indicate mortality associated with high levels of ambient PM2 5 is greatest
31 for elderly people with cardiopulmonary disease (U.S. Environmental Protection Agency, 1996).
March 2001 5-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 Longitudinal studies were conducted in the Netherlands by Janssen (1998) and Jassen et al.
2 (1997, 1998a,b, 1999a,b) on purposefully selected samples of adults (50 to 70 years old) and
3 children (10 to 12 years old). Several additional studies have focused on nonsmoking elderly
4 populations in Amsterdam and Helsinki (Janssen et al., 2000), Tokyo (Tamura et al., 1996a),
5 Baltimore (Liao et al., 1999; Williams et al., 2000a,b,c), and Fresno, CA (Evans et al. 2000).
6 These cohorts were selected because of the low incidence of indoor sources of PM (such as
7 combustion or cooking). This should allow an examination of the relationship between personal
8 and ambient PM concentrations without the large influences caused by smoking, cooking, and
9 other indoor particle-generating activities. The EPA has a research program focused on
10 understanding PM exposure characteristics and relationships. Within the program, longitudinal
11 studies are being conducted on elderly participants with underlying heart and lung disease
12 (COPD, patients with cardiac defibrillator, and myocardial infarction), an elderly environmental
13 justice cohort, and asthmatics. These studies are being conducted in several cities throughout the
14 United States and over several seasons. Only preliminary data are currently available, and results
15 are not reported in this document.
16 A series of studies by Phillips et al. (1994, 1996, 1997a,b, 1998a,b, 1999) examined
17 personal ETS exposure in several European cities. Participants varied by age and occupation.
18 Respirable Particulate Matter (RSP) concentrations were reported. These studies are not
19 included in Table 5-4 because of their focus on ETS exposure, which is not the focus of this
20 chapter. A small personal exposure study in Zurich, Switzerland, was reported by Monn et al.,
21 (1997) for PM10. This study also is not listed in Table 5-4 because indoor and outdoor
22 measurements were not taken simultaneously with the personal measurements, and other details
23 of the study were not published.
24
25 5.4.2.2 Microenvironmental Data
26 Usually, personal PM monitoring is conducted using integrated measurements over a 12- or
27 24-h period. As such, total PM exposure estimates based on PEM measurements do not capture
28 data from individual microenvironments. Recent studies have examined PM concentrations in
29 various microenvironments using a number of different types of instruments ranging from filter-
30 based to continuous particle monitors. Details on the instruments used, measurements collected,
31 and findings of these studies according to microenvironment (residential indoor, nonresidential
March 2001 5-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 indoor, and traffic-related) are summarized in Table 5-5. Studies that collected
2 microenvironmental data as part of a personal exposure monitoring study are summarized in
3 Table 5-4. In general, the studies listed in Table 5-5 are relatively small, purposeful studies
4 designed to provide specific data on the factors that effect microenvironmental concentration of
5 PM from both ambient and nonambient sources.
6 Recently published studies have used various types of continuous monitors to examine
7 particle concentrations in specific microenvironments and resulting from specific activities.
8 Continuous particle monitors such as the SMPS, APS, and Climet have been used to measure
9 particle size distributions in residential microenvironments (Abt et al., 2000a; Long et al., 2000a;
10 Wallace et al., 1997; Wallace, 2000a; McBride et al., 1999; Vette et al., 2001). These studies
11 have been able to assess penetration efficiency for ambient particles and microenvironments
12 indoors as well as penetration factors and deposition rates. Continuous instruments are also a
13 valuable tool for assessing the impact of particle resuspension caused by human activity.
14 A semi-quantitative estimate of PM exposure can be obtained using personal nephelometers that
15 measure PM using light-scattering techniques. Recent PM exposure studies have used personal
16 nephelometers (1 min avg time) to measure PM continuously (Howard-Reed et al., 2000;
17 Quintana et al., 2000) in various microenvironments. These data have been used to identify the
18 most important ambient and nonambient sources of PM, to provide an estimate of source
19 strength, and to compare modeled time activity data and PEM 24-h mass data to nephelometer
20 measurements (Rea et al., 2001). Several studies also have examined PM exposure in vehicles
21 using both continuous and filter-based techniques.
22
23 5.4.2.3 Interpretation of Particulate Matter Exposure Data
24 Papers that have reanalyzed and interpreted the data collected in previous PM exposure
25 studies are summarized in Table 5-6. These analyses are directed towards understanding the
26 personal cloud, the variability in total PM exposure, and the personal exposure-to-ambient
27 concentration relationships for PM. Results are highlighted here and given in more detail in
28 Section 5.4.3. Brown and Paxton (1998) determined that the high variability in personal
29 exposure to PM makes the personal-to-ambient PM relationship difficult to predict. Wallace
30 (2000b) used data from a number of studies to test two hypotheses: elderly COPD patients have
31 (1) smaller personal clouds and (2) higher correlations between personal exposure and ambient
March 2001 5-24 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-5. SUMMARY OF RECENT MICROENVIRONMENTAL MEASUREMENT STUDIES
.
to
o
o
Reference
Study Size
Description Instrument(s) Fraction
Summary of Measurements
Notes/Findings
to
fe
H
6
o
o
H
O
o
H
W
O
O
HH
H
W
Residential Indoor: Nonsmoking Homes
Abt et al. (2000a)
Boston, MA
Long et al. (2000a)
Boston, MA
2 homes,
2 seasons,
6 days
9 homes,
2 seasons
SMPS
APS
SMPS
TEOM
Detailed indoor/outdoor traces of PM in
various size classes for different air
exchange rates (< Ih "' to > 2 h"1)
0.02-10 Continuous PM distributions and size
distributions obtained for indoor and
outdoor air
Major indoor sources of PM: cooking, cleaning, human
activity
Sources of fine particles: cooking and outdoor particles;
Sources of coarse particles: cooking, cleaning, indoor
activities. 50% of particles by volume generated by
Anuszewski et al. (1998)
Seattle, WA
Leaderer et al. (1999)
Southwest, VA
9 homes,
18 days
58 homes, summer
Nephelometer 2.5
(radiance)
2.5
10
Wallace et al. (1997, 2000b) 1 home, 4 years
Reston, VA
Howard-Reed et al. (2000) 15 participants
Fresno, CA
Baltimore, MD
SMPS
Climet
PAHs
Black carbon
Reaetal. (2001)
Baltimore, MD
Fresno, CA
15 participants
6 size bins;
100 size
channels
0.01-
0.4 ,um
Nephelometer 0.1-10
(personal
MIE) 2.5
PEM
Nephelometer 0.1-10
(personal
MIE) 2.5 and 10
PEM
Simultaneous indoor and outdoor PM
measured continuously; 1-h avg time,
I/O = 0.98; air exchange rate: 0.7-1.7 h'1
24 h mean: Regional air 26.0 ± 11.5 ,ug/m3
(n = 47); Outdoor homes 28.0 ± 17.7 ,ug/m3
(n = 43); Indoor w/ AC 28.9 ± 18.7 ,ug/m3
(n = 49); Indoor w/o AC 33.3 ± 14.2 ^g/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 //g/m3
(n = 9)
Time activity data, whole-house air
exchange rates
Continuous carbon monoxide: descriptive
data for monitored pollutants; size profiles
for six indoor particle sources
Continuous (15-min avg) PM and time
activity data; 24-h PM mass; participants
from Baltimore and Fresno PM panel
studies.
Descriptive statistics from each study for
five microenvironments
Continuous (15-min avg) PM and time
activity data; 24-h PM mass;
Modeled PM mass and time activity data to
apportion time spent in a location. Good
comparison with nephelometer mass (6-
20%)
indoor events were ultrafine particles.
Homes contained asthmatic children, heavy wood
burning.
Dominant source of fine particles was outdoor air.
Epidemiological study of maternal and infant health
effects associated with indoor air pollution
Indoor PM concentrations were lower for homes with air
conditioning (AC) than non-air-conditioned homes.
0.3- to 0.5-,um particles linked to outdoor concentrations,
frying, broiling; 0.5- to 2.5-^m particles related to
cooking events; >2.5-^m particles influenced by physical
movement.
Time-series plots of personal nephelometer data showed
that each participant's PM exposure consisted of a series
of short-term peaks, imposed on a background caused by
ambient PM concentrations.
54 ± 31% of average daily PM2 5 exposure occurred
indoor residences, where participants spent 83 ± 10% of
their time. A significant portion of PM2 5 exposure
occurred where participants spent 4-13% of their time.
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL MEASUREMENT STUDIES
h-^
O
to
O
O
Reference
Residential Indoor:
Quintana et al. (2000)
San Diego, CA
Chang et al. (2000)
Baltimore, MD
Study
Description
Nonsmoking Homes (cont'd)
Asthmatic children
indoor and outdoor
9 homes
1 person
performing
predetermined
activities
Size
Instruments) Fraction
Nephelometer 0.1-10
(personal
MIE) 2.5 and 10
Harvard
impactors
TEOM
"Roll around" 2.5
monitor
(RAS)(PM25,
CO, VOC, O3
NO2 SO2)
Summary of Measurements
Indoor and outdoor measurements collected
using passive, active, and active heated
nephelometers for comparison to PM mass
measurements.
1-h personal exposures measured
simultaneously. Personal and ambient
concentrations were compared.
Notes/Findings
Nephelometer correlates best with PM2 5
vs. Indoor PM25 r = 0.66 vs. indoor PM10 r = 0.13
vs. outdoor PM25 r = 0.42 vs. outdoor PM10 r = 0.20
1-h personal O3 exposures were significantly lower in
indoor than outdoor microenvironments.
1-h personal CO exposures were highest in vehicles.
Personal and ambient PM2 5 correlations were strongest
for outdoor microenvironments and those with high air
exchange rates (i.e., vehicles).
1
to
O
£
1
O
0
2|
O
H
0
0
H
W
O
H
W
Lioy etal. (1999)
NA
Residential Indoor: Other
Brauer etal. (1996)
Mexico
Jenkins etal. (1996 a,b)
16 U.S. Cities
McBride etal. (1999)
NA
Vette etal. (2001)
Fresno, CA
10 vacuum cleaners
Home Types
22 rural Mexican
homes
(smoking and
nonsmoking)
Smoking and
nonsmoking homes
Combustion source
(incense) and
walking
(1 room, carpeted)
Detached
semioccupied
residence
0.3-0.5
Inertial 10
impactor
Nephelometer 2.5
(Radiance)
Fluoropore Particle
membrane phase ETS
filters markers
Met-One laser
particle
counter
SMPS 0.01-2.5
LASX
Vacuum cleaners ranged in collection
efficiency from 29-99%.
Indoor PM25: 132-555 ^g/m3 PM10: 282-
768 Mg/m3. Outdoor PM25: 37 ,ug/m3 PM10:
68,ug/m3; I/OPM25: 1.8-12.4; PM10: 4.7-
10.0
Mean PM3 5 concentrations were 17-20
l/g/m3 in smoking homes over nonsmoking
homes
Ratios of particle counts a 1.0 and 5.7 m
from the combustion source/activity were
obtained
Temporal relationships between indoor and
outdoor aerosol concentrations evaluated;
penetration factors and deposition rates
estimated. Fresno panel study empty
residence
Substantial fine particle emissions from motors with
emission rates from 0.028 - 128.8 /^g/min.
Variety of cooking fuels used
Nephelometer data were highly correlated with PM2 5
PM10 indoors (r = ~0. 87-0.95)
and
Proximity to source may help explain the existence of a
personal cloud.
Diurnally variable indoor/outdoor aerosol concentration
ratios because of resuspension from daytime activities.
Penetration factors ranged from 0.5 to 0.9.
Nonresidential Microenvironments
Bohadana et al. (2000)
Manufacturing
plant, woodworkers
Not given
443 personal time-weighted average
occupations samples of airborne dust
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL MEASUREMENT STUDIES
h-^
o
tr
o
o
1
to
o
£
H
6
o
2|
O
H
0
0
H
W
O
T^
O
k__l
H
W
Reference
Study Size
Description Instruments) Fraction
Summary of Measurements
Notes/Findings
Nonresidential Microenvlronnients (cont'd)
Donham et al. (2000)
San Francisco, CA
Klepeisetal. (1996)
San Francisco, CA
Nieuwenhuij sen et al. (1999)
Teschkeetal. (1999)
Baeketal. (1997)
Korea
Ottetal. (1996)
California
Houseman et al. (in press)
Boston, MA
Brauer and 't Mannetje
(1998)
Vancouver, BC
Lee and Chang (1999)
Hong Kong
34 poultry workers NIOSH 5
Method 0600
monitors
probed
respirators
Airport lounge, TSI8510 3.5
ETS piezobalance
Agricultural 4
activities
Wood production =50
wood finishing
wood construction
workers
Indoor and outdoor 3 . 5
smoking
restaurants
Bar before and after Piezobalance 3.5
smoking prohibited
Indoor and outdoor TSI 10
restaurants, DUSTRAK
stores
indoor restaurants, 2.5
various smoking 10
policies
indoor and outdoor 10
5 classrooms
Total dust sampled indoor respiratory
masks.
Personal monitoring: 630 ± 980 Mg/m3
(n = 210) ranging from 10-7,730 Mg/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
Indoor concentrations: 33-475 Mg/m3
Outdoor concentrations: 12-172 Mg/m3
I/O: 2.4
Smoking permitted:
indoor 26.3-182 Mg/m3; outdoor <5-67
Mg/m3
Smoking prohibited:
indoor 4-82 Mg/m3; outdoor 2-67 Mg/m3
Indoor restaurants: 14-278 Mg/m3
Outdoor restaurants: 7-281 Mg/m3
Indoor stores: 12-206 Mg/m3
Outdoor stores: 7-281 Mg/m3
Nonsmoking: PM25 7-65 Mg/m3;
PM10<10-74Mg/m3
Restricted smoking (>40% nonsmoking)
PM2.5 11-163 Mg/m3 ; PM10 24-89 Mg/m3
Unrestricted smoking: PM25 47-253 Mg/m3 ;
PM10 51-268 Mg/m3
Indoor PM10: 30-470 Mg/m3
Outdoor PM10: 20-617 Mg/m3
Respirable dust constituted about W,
measured.
'o of total dust
Personal exposures to ETS can be modeled in these types
of microenvironments .
No significant correlation between indoor and outdoor
measurements.
I/O nonsmoking: 2.2.
I/O smoking: 3.4
Avg I/O for restaurants: 2.3
Not known if the restaurants allowed
smoking
In stores, indoor and outdoor measurements were
correlated, avg I/O: 0.83
-------
TABLE 5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL MEASUREMENT STUDIES
Study Size
Reference Description Instruments) Fraction Summary of Measurements Notes/Findings
O
to
O
O
Traffic-Related Microeiwironments
to
oo
fe
H
6
o
o
H
O
O
H
W
O
O
H
W
Praml and Schierl (2000)
Munich, Germany
Aim etal. (1999)
Kuopio, Finland
Monnetal. (1997)
Switzerland
Rodes etal. (1998)
Saevanenso,
Los Angeles, CA
Roorda-Knape et al. (1998)
van Vliet etal. (1997)
Netherlands
Houseman et al. (in press)
Boston, MA
Brauer etal. (1999)
Vancover, BC
Janssenetal. (1997)
Netherlands
Trams and buses, Continuous 10
rural and urban millipore
polycarbonate
filter
9-km commuter
route, rush hours
1/mo
Climet
Spatial scale from a Harvard /je
city street monitor
In-vehicle, various
road types, 2-h trips
Gradient in
distance from
roadway
Harvard
impactor
6 channels
10
2.5
10
2.5
10
Black
smoke
n = 201 4-h trips, mean concentration
155 //g/m3 range: 13-686 //g/m3
I/O: 2.8
Windows closed, vents open
air exchange rate 36-47 h"1
48- or 72-h avg times; horizontal distance
from street: 0, 15, 50, and 80 m; vertical
distance from street: 20 m
Air exchange rates measured at various
ventilation settings and speeds.
Monitoring vehicle followed a diesel bus or
truck.
PM monitoring at 50, 100, 150, and 300 m
from roadway; 1-week avg time
Indoor and outdoor TSI PM10 Vehicle concentrations ranged from
vehicles DUSTRAK 33-170//g/m3. Outdoor vehicle
buses, subways concentrations ranged from 40-144 /^g/m3.
Bus concentrations: 17-268 /^g/m3; outdoor
10-203 //g/m3
Subway: 28-174 //g/m3; outdoor
8-203 Mg/m3
Commuting APC-1000 PM < 5: greatest concentrations by
environments combustion powered vehicles
PM > 5: greatest concentrations by
bicycling and buses
Background and 2.5 PM25 background: 21-35 //g/m3; roadway
roadway 10 23-43 ,ug/m3
PM10 background: 13-32 and 29-62 Mg/m3;
roadway 16-56 and 30-75 //g/m3
Tram > circular bus route:
Day > night
radial bus route
Morning commutes were generally higher than afternoon
commutes; relationships determined between PM and
wind speed and vehicle speed
MeanPM1027.3±3.0,ug/m3
No vertical gradient (0-20 m) and horizontal gradient
(0-80 m) in distance from road, each about 13%.
No significant differences between wet and dry periods.
Vehicles in front of the monitored vehicle accounted for
most of the in-vehicle commuting exposure; average I/O:
0.6-0.8 h"1 for PM25; carpool lane concentrations were
30-60% lower than noncarpool lane concentrations
No concentration gradient with increasing distance from
the roadways for PM2 5 and PM10; concentration gradient
did exist for black smoke, also found an effect with wind
direction
The average in-vehicle to outdoor ratio was 0.99.
Average I/O: 3; subway values were correlated with
outdoor concentrations.
Average roadway/background ratio: 3 for PM25 and PM10.
Average increase in concentration at the roadway
7.2-12.7//g/m3.
-------
TABLE 5-6. PAPERS INTERPRETING PARTICIPATE MATTER EXPOSURE STUDIES
O
to
O
O
Reference
Study Cited
Obj ectives/Hypotheses
to
VO
fe
H
6
o
o
H
O
O
H
W
O
O
H
W
Ozkaynak and
Spengler (1996)
Wallace (2000a) PTEAM (Ozkaynak et al., 1990; Spengler
et al., 1989; Wiener 1988, 1989; Wiener
etal., 1990)
THEES (Lioy et al., 1990)
Nashville COPD (Bahadori et al., 2001)
Amsterdam COPD (Janssen et al., 1997,
1998a)
Boston COPD (Rojas-Bracho et al., 2000)
Dockeryand Spengler, 1981
PTEAM (Ozkaynak et al., 1996a,b)
Netherlands (Janssen et al., 1995)
Brown and Paxton THEES (Lioy et al., 1990)
(1998) PTEAM pilot (Wallace, 1996)
Boston and Nashville COPD (Rojas-
Bracho et al., 2000); Bahadori et al., 2001)
Wilson and Suh Philadelphia (Burton et al., 1996; Suggs
(1997) and Burton, 1983)
EPA AIRS database
Wilson et al. (2000) New Jersey (Lioy et al., 1990)
Japan (Tamura et al., 1996a)
PTEAM (Clayton et al., 1993; Ozkaynak
etal., 1996a,b)
Netherlands (Janssen, 1998a;
Suh etal., 1992)
Mage et al. (1999) Japan (Tamura et al., 1996a)
State College (Suh et al., 1995)
Netherlands (Janssen et al., 1997, 1998a,
1999a)
New Jersey (Lioy et al., 1990)
PTEAM (Clayton et al., 1993; Ozkaynak
etal., 1996a,b)
Mage (1998) PTEAM (Clayton et al., 1993; Ozkaynak
etal., 1993, 1996a,b)
Examines the differences between pooled and longitudinal
correlations in personal and ambient (or outdoor) data for
PM25 and PM10.
Discusses the personal cloud for PM25 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 indoor/outdoor
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 indoor/outdoor
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 epidemiological health effects.
Uses personal exposure equation, mass balance, regression
analysis, and deductive logic.
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 the average person is
exposed to.
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 Mg/m3; PM2 5 ~ 6 g/m3) than for other
healthy populations (PM10: 27-56 Mg/m3; PM25:
11-27 //g/m3) of elderly, children, and the general
population. However, higher correlations of personal
exposure with ambient concentrations were not observed.
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.
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 PM25
and >64% of ambient PM10 measured by the community
monitor.
-------
1 concentrations, compared to healthy elderly, children, and the general population. The analysis
2 by Wallace (2000a) and three subsequent longitudinal studies (Williams 2000a,b,c; Ebelt et al.,
3 2000; Sarnat et al., 2000) support hypothesis 1 but not hypothesis 2. Ozkaynak and Spengler
4 (1996) show that at least 50% of personal PM10 exposure for the general population is because of
5 ambient particles that significantly contribute to inhaled particles. Wilson and Suh (1997)
6 conclude that fine and coarse particles should be treated as separate classes of pollutants because
7 of differences in characteristics and potential health effects. Wilson et al. (2000) give a review of
8 what they call the "exposure paradox" and determine that personal PM needs to be divided into
9 different classes according to source type, and that correlations between personal and ambient
10 PM will be higher when nonambient sources of PM are removed from the personal PM
11 concentration. Mage (1998) conducted analysis using the PTEAM data and showed that on
12 average a person is exposed to >75% of ambient PM2 5 and >64% of ambient PM10. Mage et al.
13 (1999) use an algorithm to fill in missing data and outliers to analyzed data sets and show that
14 variation in daily personal exposures for subjects with similar activity patterns and no ETS
15 exposure are driven by variation in ambient PM concentrations.
16
17 5.4.3 Factors Influencing and Key Findings on Particulate Matter Exposures
18 5.4.3.1 Correlations of Personal/Microenvironmental Particulate Matter with Ambient
19 Particulate Matter
20 The relationship between measured personal PM exposure and PM concentrations
21 measured at ambient sites has been of interest to exposure analysts. Many of the studies,
22 summarized above in Table 5-4, have analyzed this relationship using measurements of personal
23 PM exposures and ambient PM concentrations. The statistical correlation between these
24 measurements for the various personal exposure studies is discussed in this section.
25
26 5.4.3.1.1 Types of Correlations
27 The three types of correlation data that will be discussed in this section are longitudinal,
28 "pooled", and daily-average correlations. Longitudinal correlations are calculated when data
29 from a study includes measurements over multiple days for each subject (longitudinal study
30 design). Longitudinal correlations describe the temporal relationship between daily personal PM
31 exposure and daily ambient PM concentration for each individual subject. The longitudinal
March 2001 5-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 correlation coefficient, r, may differ for each subject, and an analysis of the variability in r across
2 subjects can be performed with this type of data. Typically, the median r is reported along with
3 the range across subjects in the study. Pooled correlations are calculated when a study involves
4 one or only a few measurements per subject and different subjects are studied on subsequent days
5 (sometimes called a "cross-sectional" study design). The different subject/different day data are
6 combined, pooled, for the correlation calculation. Pooled correlations describe the relationship
7 between daily personal PM exposure and daily ambient PM concentration across all subjects in
8 the study. This type of correlation is sometimes called cross-sectional, but will be called pooled
9 in this chapter because only a limited number of participants are monitored on any given day.
10 Daily-average correlations are calculated using the average exposure across subjects for each
11 day. Daily-average correlations describe the relationship between daily community-averaged
12 personal PM exposure and daily ambient PM concentration. This type of correlation could be
13 called cross-sectional, but given that the pooled correlation also is referred to as cross-sectional,
14 the term daily average is used here.
15 Studies that have reported longitudinal correlations also typically have reported pooled
16 correlations. However, pooling of the data for the correlation has been handled differently across
17 the various studies. For some studies, the multiple days of measurements for each subject were
18 assumed to be independent (after autocorrelation and sensitivity analysis) and combined together
19 in the correlation calculation (Ebelt et al., 2000). In other studies, daily averages across subjects
20 were calculated and the correlation determined from the daily averages (Williams et al., 2000b).
21 A third approach also was used in other studies to simulate a cross-sectional study design
22 (Janssen et al., 1997, 1998a, 1999c). In this approach, a random-sampling procedure was used to
23 select a random day from each subject's measurements to use for the correlation. This procedure
24 was repeated many times, and statistics such as the mean and standard deviation of the pooled
25 correlation coefficient were reported.
26 The type of correlation analysis can have a substantial effect on the resulting correlation
27 coefficient. Mage et al. (1999) mathematically demonstrated that very low correlations between
28 personal exposure and ambient concentrations could be obtained when people with very different
29 nonambient exposures are pooled, even though their individual longitudinal correlations are high.
30 The longitudinal studies conducted by Tamura et al. (1996a) and Janssen et al. (1997, 1998a,
31 1999c) determined that the longitudinal correlations between personal exposure and ambient PM
March 2001 5-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations were much higher than the correlations obtained from a pooled data set. Wallace
2 (2000a) reviewed a number of longitudinal studies found that the median longitudinal correlation
3 coefficient was much higher than the pooled correlation coefficient for the same data (see
4 Tables 1 and 2 Wallace, 2000a). Williams et al. (2000a,b) and Evans et al. (2000) reported
5 higher correlation coefficients for daily-average correlations compared to longitudinal
6 correlations.
7
8 5.4.3.1.2 Correlation Data from Personal Exposure Studies
9 Measurement data and correlation coefficients for the personal exposure studies described
10 in Section 5.4.2.1 are summarized in Table 5-7. All data are based on mass measurements. The
11 studies are grouped by the type of study design, longitudinal or pooled. For each study in
12 Table 5-7, summary statistics for the total personal PM exposure measurements are presented,
13 as well as statistics for residential indoor, residential outdoor, and ambient PM concentrations,
14 where available. The correlation coefficient (r) between total personal PM exposures and
15 ambient PM concentrations also are presented and classified as longitudinal or pooled
16 correlations. When reported,/^-values for the correlation coefficients are included. Correlation
17 coefficients between personal, indoor, outdoor, and ambient also are reported, when available.
18
19 5.4.3.1.3 Correlations Between Personal Exposures, Indoor, Outdoor, and Ambient
20 Measurements
21 Longitudinal and pooled correlations between personal exposure and ambient or outdoor
22 PM concentrations varied considerably between study and study subjects. Most studies report
23 longitudinal correlation coefficients that range from <0 to ~ 1, indicating that an individual's
24 activities and residence type may have a significant effect on total personal exposure to PM.
25 General population studies tend to show lower correlations because of the higher variation in the
26 levels of PM generating activities. In contrast, the absence of indoor sources for the populations
27 in several of the longitudinal studies resulted in high correlations between personal exposure and
28 ambient PM within subjects over time for these populations. But even for these studies,
29 correlations varied by individual, depending on their activities and the microenvironments that
30 they occupied.
31
March 2001 5-32 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-7. PERSONAL MONITORING STUDIES FOR PARTICIPATE MATTER: MEASURED CONCENTRATIONS
o
to
o
o
1
OJ
OJ
HH
P3
H
I
0
o
2|
0
H
O
O
H
W
O
73
O
1 — I
H
W
AJMJ CUK
KUJ^AIHJ
1> CUUFJ
Measured Concentration Levels (,ug/m3)
Size Avg. Sample
Fraction Time Statistic Size1
Longitudinal Studies
Ebelt et al. (2000) - Vancouver, BC
PM25 24 h x ± SD 106
Range
Evans et al. (2000) - Fresno, CA
PM2J 24 h x 24
Range
PM25 24 h x 12
Range
Janssen et al. (1997) - Netherlands
PM10 24 h x±SD 301
Range
Janssen et al. (1998a) - Netherlands
PM10 24 h x±SD 262
Range
Janssen et al. (1999c) - Netherlands
PM2J 24 h x±SD 77
Range
PM2J 24 h x±SD 55
Range
PM2J 24 h x±SD 22
Range
Residential Residential
Personal Indoor Outdoor Ambient
18. 2 ±14.6
2-91
13.3 9.7 20.5
1-24 4-17 4-52
11.1 8.0 10.1
7-16 4-12 5-20
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
11.4±4.1
4-29
21.7
6-37
8.6
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
Type3
Median L
P
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
H icuiirN i s
Personal-Ambient 2 Other
Correlation Coefficients (r) Correlation Coefficients (r)
Value (Range)
0.48 (-0.68-0.83)
0.15
0.41"
0.84"
0.63(0.1-0.9)
0.63
0.59
0.28(0.12)'
0.45(0.16)'
0.20(0.19)'
0.50 (-0.41-0.92)
0.50(0.07-0.83)'
0.34 (-0.09-0.67)'
0.86 (-0.11-0.99)
0.41 (-0.28-0.93)'
0.92
0.82'
Notes Type3 Value (Range)
n = 16 COPD subjects
Fresno-1 study Pp.; 0.81"
PP-O 0.80"
Fresno-2 study Pp.; 0.95"
PP-O 0.80"
n = 45 school children
With nonsmoking parents
With smoking parents
All
With nonsmoking parents
With smoking parents
n = 37 adults Med. L^ 0.72 (-0. 10-0.98)
No ETS exposure Med. LJ., 0.73 (-0.88-0.95)
All
n = 13 school children
With nonsmoking parents
With smoking parents
-------
TABLE 5-7 (cont). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
o
to
o
o
UJ
O
£
H
O
0
H
O
O
w
o
V
o
H H
H
W
CUKKUJ^AllUrN I
.JUJiiFFK.
Measured Concentration Levels (fj.g/m1)
Size Avg.
Fraction Time Statistic
Janssen et al. (2000) - Netherlands
PM2 5 24 h x ± SD
Range
Janssen et al. (2000) - Finland
PM2 5 24 h x ± SD
Range
Linn et al. (1999) - Los Angeles
PM2J 24 h x ± SD
Range
PM10 24 h x ± SD
Range
Rojas-Bracho et al. (2000) - Boston
PM2J 12 h x ± SD
Range
PM10 12 h x ± SD
Range
PM10.2.S 12 h x ± SD
Range
Sarnat et al. (2000) - Baltimore
PM2 5 24 h x ± SD
x±SD
PM10 24 h x ± SD
x±SD
PM10.2 s 24 h x ± SD
x±SD
Sample
Size1
338
336
60
59
224
225
222
37
36
37
36
37
36
Residential Residential
Personal Indoor Outdoor Ambient
24.3 ±25. 7 28. 6 ±41. 8 20.6 ±4.0
9-134 9-239 13-31
10.8 ±4.4 11.0 ±4.0 12.6 ±2.0
4-33 3-27 10-18
23.8 ±15.1 23.5 ±15.3 24.8 ±14.5
4-65 4-92 4-63
34.8 ±14.8 32.6 ±15.6 39.8 ±18.3 33 ± 15
5-85 9-105 7-97 9-??
21.6±13.6 17.5±14.1 14.2±11.2
1-128 2-73 1-57
37.2 ±22.8 31.9 ±25.2 22.2 ±18.7
9-211 2-329 3-76
15.6 ±14.6 14.5 ±9.2 8.1 ±6.8
-11-103 -3-255 -2-64
26.7 ±13. 7 25. 2 ±11. 5
18.5±11.2 5.6±49.0
33. 9 ±11. 7 34.0 ±12.8
28.0 ±16.5 7.5 ±73.2
7.2 ±4.0 8. 4 ±2.3
9.6 ±7.9 -1.3 ±24.2
Type3
Median L
Median L
Median L
P
P
Median L
Median L
Median L
Median L
Median L
P
P
P
P
Median L
Median L
Median L
Median L
;iJi;rN i s
Personal-Ambient 2
Correlation Coefficients (r)
Value (Range) Notes
0.79 (-0.41-0.98) n = 36 elderly w/CV disease
0.85 No ETS exposures
0.76 (-0. 12-0.97) n = 46 elderly w/CV disease
0.266
0.226
0.61(0.10-0.93)' n= 17 adults
0.35 (0.0-0.72)6
0.30 (0.0-0.97)6
0.76 (-0.21-0.95)7 n = 15 adults; summer
0.25 (-0.38-0.81)7 n = 15 adults; winter
0.89s Lfigh ventilation; summer
0.75s Med. ventilation; summer
0.50s Low ventilation; summer
0.44s WINTER
0.64(0.08-0.86)7 SUMMER
0.53 (-0.79-0.89)7 WINTER
0.11(-0.60-0.64)7 SUMMER
0.32 (-0.48-0.68)7 WINTER
Other
Correlation Coefficients (r)
Type3 Value (Range)
Med. Lp.; 0.91 (-0.28-1.0)
Med. L- 0.84(-0.00-0.98)
Med. Lp.; 0.89 (0.14-1.0)
Med. L- 0.70 (-0.15-0.94)
P;., 0.266
PO- 0.47s
P;., 0.32'
PO- 0.66s
Med. Lp.; 0.876
Med. L-. 0.746
Med. L,,.; 0.716
Med. L-. 0.506
Med. L,,.; 0.426
Med. L-. 0.206
-------
TABLE 5-7 (cont). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
o
to
o
o
Size Avg.
Fraction Time
Tamura et al. (1996a)
PM10 48 h
CUKKUJ^AllUrN CUUFFlt
Measured Concentration Levels (,ug/m3)
Sample Residential Residential
Statistic Size1 Personal Indoor Outdoor Ambient Type3
- Tokyo
P
;iJi;rN i s
Personal-Ambient 2
Correlation Coefficients (r)
Value (Range) Notes
0.83 n = 7 elderly adults
Other
Correlation Coefficients (r)
Type3 Value (Range)
Williams et al. (2000a,b) - Baltimore
i
OJ
L/l
o
^
H
I
0
o
0
H
O
o
H
W
O
7s
O
i — i
H
W
PM2J 24 h
PM10 24 h
PM10.2.5 24 h
Pooled Studies
x 23 13.0 9.4 22.0 22.0 Median L
Range 7-25 4-19 7-52 8-59 P
x 28 11.0 30.0 29.9
Range 4-23 13-66 13-74
x 26 1.0 8.0 8.0
Range -3-5 -2-16 1-15
0.80 (0.38-0.98)6 n = 21 elderly adults
0.89"
Pp.; 0.90"
PP-O 0.95"
P;-o 0.94"
P;., 0.87"
PO- 0.96"
P;.. 0.82"
P;- 0.81"
PO- 0.94"
Pi.. 0.18"
P;., 0.08"
PO- 0.45"
Bahadori (1998) - Nashville
PM2J 12 h
PM10 12 h
Pellizzari et al. (1999)
PM2 5 3 d
PM10 3 d
Oglesby et al. (2000a)
PM2! 48 h
x±SD 30 21.7 ±10.5 15.5 ±6.6 23.4 ±6.8 P
Range 10-67 5-40 3-61
x±SD 30 33.0 ±16.9 21.6 ±10.7 32.5 ±8.1 P
Range 5-88 9-77 7-76
- Toronto
x 922 28.4 21.1 15.1 P
x 141 67.9 29.8 24.3
- EXPOIIS Basel
x±SD 44 23.7 ±17.1 19.0 ±11.7 P
20 17.5 ±13.0 17.7 ±7.1 P
0.09 n = 10 COPD subjects;
daytime
-0.08 n = 10 COPD subjects;
daytime
0.23 n = 178; n for indoor,
outdoor lower than personal
No correlations reported
0.07 All
0.21 No ETS exposure
Pp, 0.72
P;-o 0.31
PP, 0.43
Pi-. 0.06
Pp.; 0.79
P;.o 0.33
-------
TABLE 5-7 (cont). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
o CUKKUl^A 1 1U1> CUU* V IClJiirN 1 S
to
o
o
Size
Fraction
Avg.
Time Statistic
Personal-Ambient 2
Measured Concentration Levels (,ug/m3) Correlation Coefficients (r)
Sample Residential Residential
Size1 Personal Indoor Outdoor Ambient Type3 Value (Range) Notes
Other
Correlation Coefficients (r)
Type3 Value (Range)
Santos-Burgoa et al. (1998) - Mexico City
PM10
24 h x ± SD
66 97 ± 44 99 ± 50 P 0.26
Pp.; 0.47
P;., 0.23
Tamura et al. (1996b) - Osaka
PM2 48 h
PM10 48 h
P
P
0.74
0.67
Abbreviations used:
i
OJ
ON
Avg.
Cone.
CV
d
ETS
h
= Averaging (time)
= Concentration
= Cardiovascular
= Day
= Environmental tobacco smoke
= Hour
i-a
i-o
L
Med.
o-a
P
Indoor-ambient correlation
= Indoor-outdoor correlation
= Longitudinal correlation
= Median
= Outdoor-ambient correlation
= Pooled correlation
p-i
p-o
SD
Stat.
x
= Personal-indoor correlation
= Personal-outdoor correlation
= Standard deviation
= Statistic
= Mean
H
6
o
o
H
O
o
H
W
O
Notes:
1. Sample size is for personal concentrations; indoor, outdoor and ambient sample sizes may differ.
2. Correlation coefficient is for personal-residential outdoor if no ambient concentration data reported.
3. See text for description of types of correlations.
4. Daily-averaged correlation (values for individual subjects averaged for each day).
5. Pooled correlations estimated using a Monte Carlo sampling procedure, n = 1000. If mean P is shown, then SD given; if median P is shown, then range is given.
6. Obtained from a regression equation; r = ^/(R2) .
6. Spearman rank correlations.
7. Calculated, r = V(^ ) , from R2 from a mixed model regression.
O
HH
H
W
-------
1 Probability Studies
2 In the Toronto study (Pellizzari et al., 1999), pooled correlations were derived for personal,
3 indoor, outdoor, and fixed site ambient measurements. This study was conducted in Toronto on
4 a probability sample of 732 participants who represented the general population, 16 years and
5 older. The study included between 185 and 203 monitoring periods with usable PM data for
6 personal, residential indoor, and outdoor measurements. For PM10, measurements, the mean
7 concentrations were 67.9 //g/m3 for personal, 29.8 //g/m3 for indoor air, and 24.3 //g/m3 for
8 outdoor air samples. For PM2 5, the mean concentrations were 28.4 //g/m3 for personal,
9 21.1 //g/m3 for indoor air, and 15.1 //g/m3 for outdoor air samples. A low but significant
10 correlation (r = 0.23, p < 0.01) was reported between personal exposure and ambient
11 measurements. The correlations between indoor concentrations and the various outdoor
12 measurements of PM25 ranged from 0.21 to 0.33. The highest correlations were for outdoor
13 measurements at the residences with the ambient measurements made at the roof site (0.88) and
14 the other fixed site (0.82). Pellizzari et al. (1999) state that much of the difference among the
15 data for personal/indoor/outdoor PM
16
17 ... can be attributed to tobacco smoking, since all variables reflecting smoking . . . were found to be
18 highly correlated with the personal (and indoor) particulate matter levels, relative to other variables that
19 were measured . . . none of the outdoor concentration data types (residential or otherwise) can
20 adequately predict personal exposures to particulate matter, (p. 729)
21
22 Santos-Burgoa et al. (1998) describe a 1992 study of personal exposures and indoor
23 concentrations to a randomly sampled population near Mexico City. The sample of 66 monitored
24 subjects included children, students, office and industrial workers, and housewives. None of the
25 people monitored were more than 65 years old. The mean 24-h personal exposure and indoor
26 concentrations were 97 ± 44 (SD) and 99 ± 50 //g ~3, respectively, with an rPersonai/Ambient = 0.26
27 (p = 0.099). Other correlations of interest were rPereonaWndoor = 0.47 (p = 0.002) and rIndoor/Ambient =
28 0.23 (p = 0.158). A strong statistical association was found between personal exposure and
29 socioeconomic class (p = 0.047) and a composite index of indoor sources at the home
30 (p = 0.039).
March 2001 5-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 Correlation analysis for personal exposure has not yet been reported for EXPOLIS. Some
2 preliminary results (Jantunen et al., 2000) show that in Basel and Helsinki, a single ambient
3 monitoring station was sufficient to characterize the ambient PM2 5 concentration in each city.
4 Using microenvironmental concentration data collected while the subjects were at home, at work,
5 and outdoors, they calculated the sum of the time-weighted-averages of these data and found the
6 results closely match the personal PM2 5 exposure data collected by the monitors carried by most
7 of the subjects, with a few subjects, mostly smokers, being noticeable exceptions.
8
9 Longitudinal Studies
10 A number of longitudinal studies using a purposeful sampling design have been conducted
11 and reported in the literature since 1996. A number of these studies (Janssen et al., 1998a,
12 1999b, 2000; Williams et al., 2000b; Evans et al., 2000) support the previous work by Janssen
13 et al. (1995) and Tamura et al. (1996a) and demonstrate that, for individuals with little exposure
14 to nonambient sources of PM, correlations between total PM exposure and ambient PM
15 measurements are high. Other studies (Ebelt et al., 2000; Sarnat et al., 2000) show strong
16 correlations for the SO4"2 component of PM2 5 but poorer correlations for PM2 5 mass. Still other
17 studies show only weak correlations (Rojas-Bracho et al., 2000; Linn et al., 1999; Bahadori et al.,
18 2001). Even when strong longitudinal correlations are demonstrated for individuals in a study,
19 the variety of living conditions may lead to variations in the fraction of ambient PM contributing
20 to personal exposure. Groups with similar living conditions, especially if measurements are
21 conducted during one season, may have similar a and, therefore, very high correlations between
22 personal exposure and ambient concentrations. However, when a panel contains subjects with
23 homes of very different ventilation characteristics or covers more than one season, variations in a
24 can be high across subjects.
25 Elderly Subjects. Janssen et al. (2000) continued their longitudinal studies with
26 measurements of personal, indoor, and outdoor concentrations of PM25 for elderly subjects with
27 doctor-diagnosed angina pectoris or coronary heart disease. Studies were conducted in
28 Amsterdam and Helsinki, Finland, in the winter and spring of 1998 and 1999. In the Amsterdam
29 study, with 338 to 417 observations, the mean concentrations were 24.3, 28.6, and 20.6 //g/m3 for
30 personal, indoor, and outdoor samples, respectively. Jf the measurements with ETS in the home
31 were excluded, the mean indoor concentration dropped to 16 //g/m3, which was lower than
March 2001 5-38 DRAFT-DO NOT QUOTE OR CITE
-------
1 outdoors. In the Helsinki study, the mean PM25 concentrations were 10.8 //g/m3 for personal,
2 11.0 //g/m3 for indoor air, and 12.6 //g/m3 outdoor air samples. The authors note that for this
3 group of subjects, personal exposure, indoor concentrations, and ambient concentrations of PM25
4 were highly correlated within subjects over time. Median Pearson's correlation coefficients
5 between personal exposure and outdoor concentrations were 0.79 in Amsterdam and 0.76 in
6 Helsinki. The median Pearson's r for the indoor/outdoor relationship was 0.85 for the
7 Amsterdam study, excluding homes with ETS. The correlation for indoors versus outdoors was
8 0.70 for all homes.
9 A series of PM personal monitoring studies involving elderly subjects was conducted in
10 Baltimore County, MD, and Fresno, CA. The first study was a 17-day pilot (January-February
11 1997) to investigate daily personal and indoor PMl 5 concentrations, and outdoor PM25 and
12 PM2 5_10 concentrations experienced by nonsmoking elderly residents of a retirement community
13 located near Baltimore (Liao et al., 1999; Williams et al., 2000c). The 26 residents were aged
14 65 to 89 (mean = 81), and 69% of them reported a medical condition, such as hypertension or
15 coronary heart disease. In addition, they were quite sedentary; less than 5 h day"1, on average,
16 was spent on ambulatory activities. Because most of the residents ate meals in a communal
17 dining area, the average daily cooking time in the individual apartments was only 0.5 h (range 0
18 to4.5h). About 96% of the residents'time was spent indoors (Williams et al., 2000c). Personal
19 monitoring, conducted for five subjects, yielded longitudinal correlation coefficients between
20 ambient concentrations and personal exposure ranging from 0.00 to 0.90.
21 Subjects with COPT). Linn et al. (1999) describe a 4-day longitudinal assessment of
22 personal PM25 and PM10 exposures (on alternate days) in 30 COPD subjects aged 56 to 83;
23 concurrent indoor and outdoor monitoring were conducted at their residences. This study
24 occurred in the summer and autumn of 1996 in the Los Angeles area. PM10 data from the nearest
25 fixed-site monitoring station to each residence also was obtained. Pooled correlations for
26 personal exposure to outdoor measurements were 0.26 and 0.22 for PM2 5 and PM10, respectively.
27 Day-to-day changes in PM2 5 and PM10 measured outside the homes tracked concurrent PM10
28 measurements at the nearest ambient monitoring location, with R2 values of 0.22 and 0.44,
29 respectively. Day to day changes in PM mass measured indoors also tracked outdoors at the
30 homes with R2 values of 0.27 and 0.19 for PM10 and PM2 5 respectively.
March 2001 5-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 Personal, indoor, and outdoor PM2 5 PM10 and PM2 5_10 correlations were reported by
2 Rojas-Bracho et al. (2000) for a study conducted in Boston, MA, on 18 individuals with COPD.
3 Both the mean and median personal exposure concentrations were higher than the indoor
4 concentrations, which were higher than outdoor concentrations for all three PM measurement
5 parameters. Geometric mean indoor/outdoor ratios were 1.4 ± 1.9 for PM10, 1.3 ± 1.8 for PM25,
6 and 1.5 ± 2.7 for PM25.10. Median longitudinal R2s between personal exposure and ambient PM
7 measurements were 0.12 for PM10, 0.37 for PM2 5 and 0.07 for PM2 5_10. The relationship between
8 the indoor and outdoor concentrations was strongest for PM2 5, with a median R2 of 0.55 and
9 11 homes having significant R2 values. For PM10 the median R2 value was 0.25, with significant
10 values for eight homes. Only five homes had significant indoor/outdoor associations for PM25.10,
11 with an insignificant median R2 value of 0.04.
12 Bahadori et al. (2001) report a pilot study of the PM exposure of 10 nonrandomly chosen
13 chronic obstructive pulmonary disease (COPD) patients in Nashville, TN, during the summer of
14 1995. Each subject alternately carried a personal PM2 5 or PM10 monitor for a 12-h daytime
15 period (8 a.m. to 8 p.m.) for 6 consecutive days. These same pollutants were monitored
16 simultaneously indoors and outdoors at their homes. All of the homes were air-conditioned and
17 had low air exchange rates (mean = 0.57 h"1), which may have contributed to the finding that
18 mean indoor PM2 5 was 66% of the mean ambient PM2 5. This can be contrasted with the
19 PTEAM study in Riverside, CA, where no air conditioners were in use and the mean indoor
20 PM25 was 98% of the mean ambient PM25 (Clayton et al., 1993). Data sets were pooled for
21 correlation analysis. Resulting pooled correlations between personal and outdoor concentrations
22 were r = 0.09 for PM25 and r = -0.08 for PM10.
23
24 5.4.3.1.4 A Correlation Between a Daily-Average Exposure and Ambient Concentrations
25 A recent biostatistical analysis (Zeger et al., 2000) suggests that the community mean
26 exposure is the appropriate parameter for analyzing exposure error in community time-series
27 epidemiology. Ott et al. (2000) suggest that the correlation of the community mean exposure
28 with ambient concentrations will approach 1.0 for a large community and demonstrated this
29 using data from the PTEAM study. Mage et al. (1999) calculated the daily-average exposure for
30 three earlier studies with sufficient data and found that the coefficients for the correlation of daily
March 2001 5-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 averages with the ambient concentrations were high (R2 ~ 0.9). Two new studies have obtained
2 sufficient data to calculate a daily-average correlation.
3 The 1997 Baltimore pilot study was followed up in July and August 1998 by a more
4 extensive study in which identical samplers were used for personal, indoor, and outdoor
5 measurements. The participants, aged 72 to 93 (mean 81) years, included healthy members, as
6 well as subjects with COPD and cardiovascular disease. The participants lived in an 18-story
7 retirement facility that provided a self-contained living environment. There was a central HVAC
8 system for common areas but each apartment had an individually controlled HVAC system. The
9 subjects had limited exposures to indoor-generated sources of PM because of their low
10 frequency/duration of activities like cooking, cleaning, or interacting with tobacco smokers
11 (Williams et al., 2000a,b). As a result, the daily-average correlation coefficient was very high
12 (r = 0.89) between personal exposure and ambient concentrations of PM25. Median longitudinal
13 correlations were also high (r = 0.81; range = 0.38 to 0.98).
14 Evans et al. (2000) report two panel studies in Fresno, CA, with daily-average correlation
15 coefficients of 0.41 and 0.84.
16
17 5.4.3.1.5 Correlations Using Sulfate as a Surrogate for Personal Exposure to Ambient
18 Particulate Matter
19 A study, conducted in Vancouver, involving sixteen COPD patients aged 54 to 86, reported
20 low median longitudinal (r = 0.48) and pooled (r = 0.15) correlation coefficients between
21 personal exposure and ambient concentrations of PM25 (Ebelt et al., 2000). However, the
22 correlation between personal exposure and ambient concentrations of SO42" was much higher.
23 The results for PM2 5 and sulfate are compared in Figure 5-1. Ebelt et al. (2000) conclude the
24 following.
25
26 We found S042" to be a good measure of exposure to accumulation mode PM of ambient
27 origin. Personal and ambient measures of S042" were highly correlated over time, unlike the
28 moderate correlation found for PM25. The individual correlations demonstrated that ambient
29 S042" was a consistently strong predictor across all individuals and all levels of exposure, whereas
30 for PM25 correlations varied by individual and were dependent upon the level of personal
31 exposure. Although indoor sources likely contribute to personal exposures of PM25, accounting
32 for such variables did not lead to models with the same predictive power as found for S042".
March 2001 5-41 DRAFT-DO NOT QUOTE OR CITE
-------
Correlation Coefficient for Individuals
1.00 -
0.75 -
0.50 -
0.25 -
P 0.00 -
-0.25 -
-0.50 -
-0.75 -
1 nn -
Ebeltetal., 2000
Pearson's r
^
p
T Sulfate
—
Percentile
90th Percentile
75th Percentile
Median
25th Percentile
10th Percentile
Sarnatetal., 2000
Spearman's r
PM_ _
1
x
x
1
h 1
Sulfate
I
i"N">i Summer
i i Winter
PM25 Sulfate
PM25 Sulfate
Figure 5-1. Comparison of correlation coefficients for longitudinal analysis of personal
exposure versus ambient concentrations for individual subjects for PM2 5 and
sulfate.
1
2
3
4
5
Similarly, we found that accounting for spatial variability in ambient levels did not improve the
relationship between ambient concentrations and measured personal exposures. Overall, we have
shown that a personal measure of exposure to outdoor source PM is highly related to variation in
ambient levels of PM.
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
March 2001
5-42
DRAFT-DO NOT QUOTE OR CITE
-------
1 et al., 2000). All subjects (nonrandom selection) were retired, physically healthy, and lived in
2 nonsmoking private residences. Each residence, except one, was equipped with central
3 air-conditioning; however, not all residences used air-conditioning throughout the summer. The
4 average age of the subjects was 75 years (±6.8 years). Sarnat et al. (2000) reported higher
5 longitudinal and pooled correlations for PM2 5 during summer than winter. Similar to Ebelt et al.
6 (2000), Sarnat et al. (2000) reported stronger associations between personal exposure to SO42"
7 and ambient concentrations of SO42". The ranges of correlations are shown in Figure 5-1 along
8 with similar data from Ebelt et al. (2000).
9 The study conducted by Sarnat et al. (2000) also illustrates the importance of ventilation on
10 personal exposure to PM. During the summer, subjects recorded the ventilation status of every
11 visited indoor location (e.g., windows open, air-conditioning use). As a surrogate for the
12 air-exchange rate, personal exposures were classified by the fraction of time the windows were
13 open while a subject was in an indoor environment (Fv). Sarnat et al. (2000) report regression
14 analyses for personal exposure on ambient concentration for total PM2 5 and for sulfate for each
15 of the three ventilation conditions. Personal exposure to sulfate may be taken as a surrogate for
16 personal exposure to ambient accumulation-mode PM in the absence of indoor sulfate sources.
17 Figure 5-2 shows a comparison of the regressions and indicates how the use of a sulfate tracer as
18 a surrogate for PM of ambient origin improves the correlation coefficient. The improvement is
19 especially pronounced for the lowest ventilation conditions. For the lowest ventilation condition,
20 R2 improves from 0.25 to 0.72.
21 The Ebelt et al. (2000) and Sarnat et al. (2000) studies did not use their sulfate data to
22 develop relationships between personal exposure to ambient PM and ambient PM concentrations
23 for individual subjects, as suggested by Wilson et al. (2000). However, the higher correlation
24 coefficients and the narrower range of the correlation coefficient for sulfate suggest that
25 removing indoor-generated and personal activity PM from total personal PM would result in a
26 higher correlation with ambient concentrations. However, the variation in ventilation status (and
27 thus in the attenuation coefficient a) still would cause variations between ambient concentrations
28 of PM and personal exposure to ambient PM, especially if the study continued long enough to
29 extend through more than one season.
30
March 2001 5-43 DRAFT-DO NOT QUOTE OR CITE
-------
60
JE
D)
in
g.
o
S2
Q>
Q_
50-
40-
30-
20-
Well Ventilated Indoor Environment
35
R2 = 0.80
30-
25-
20-
15-
10-
5-
0
R2 = 0.88
60
o
Q.
X
LU
"CD
c
o
10
CD
CL
50-
40-
30-
20-
10-
Moderately Vented Indoor Environment
35-
R2 = 0.57
30-
25-
20-
15-
10-
5-
R2 = 0.73
Poorly Ventilated Indoor Environment
60
E 50-
IL
CD 40-
.2
30H
20-
o
U)
CD ..n j
Q_ 10 -
R2 = 0.25
35'
30-
25-
20-
15-
10-
5-
10 20 30 40 50 60
PM2.5
R2 = 0.72
*. *
0 5 10 15 20 25 30 40
S04
Ambient Concentration (|jg/m3)
Figure 5-2. Personal exposure versus ambient concentrations for PM2 5 and sulfate. (Slope
estimated from mixed models).
Source: Sarnat et al. (2000).
March 2001
5-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.4.3.1.6 Correlations Between Personal Exposure to A mbient and Nonambient
2 Particulate Matter
3 The utility of treating personal exposure to ambient PM, Eag, and personal exposure to
4 nonambient PM, Enonag, as separate and distinct components of total personal exposure to PM, Et,
5 was pointed out by Wilson and Suh (1997). The PTEAM study measured, in addition to indoor,
6 outdoor, and personal PM, the air exchange rate for each home and collected information on the
7 time spent in various indoor and outdoor jue. This information is available for 147, 12-h daytime
8 periods. With this information, it is possible to estimate the daytime Eag and Enonag as described
9 in Section 5.3.2.3.1. Various examples of this information have been reported (Mage et al.,
10 1999; Wilson et al., 2000). Graphs showing the relationships between ambient concentration and
11 the various components of personal exposure (Et, Eag, and Enonag) are shown in Figure 5-3. The
12 correlation coefficient for the pooled data set improves from r = 0.377 for Et versus Ca
13 (Figure 5-3a) to r = 0.856 for Eag versus Ca (Figure 5-3b) because of the removal of the Enonag ,
14 which, as shown in Figure 5-3c, is highly variable and independent of Ca. The correlation
15 between Eag and Ca is less than 1 because of the day-to-day variation in ait. The regression
16 analysis with Et total PM gives (X = 0.711 and N = 81.6 //g/m3. The regression analysis with Eag
17 gives (X = 0.625. The regression with Enonag gives N = 79.2 //g/m3. The finite intercept in the
18 regression with Eag must be attributed to bias or error in some of the measurements. No studies,
19 other than PTEAM, have provided the quantity of data on Et, Ca, C;, and a required to conduct
20 an analysis comparable to that shown in Figure 5-3.
21 The higher correlations found between daily-average personal exposures and ambient PM
22 concentrations, as opposed to lower correlations found between individual exposures and
23 ambient PM levels, recently have been attributed to statistical rather than physical causes. Ott
24 et al. (2000), using their Random Component Superposition (RCS) model, solely attribute this to
25 the averaging process. Because personal exposures also include contributions from ambient
26 concentrations, the correlation between personal exposure and ambient concentrations increases
27 as the number of subjects measured daily increases. Based on theory, Ott et al. (2000) predict
28 expected correlations above 0.9 if 25 subjects had been studied during the PTEAM study and
29 above 0.70 in the New Jersey study reported by Lioy et al. (1990).
30
March 2001 5-45 DRAFT-DO NOT QUOTE OR CITE
-------
250
200-
o>
= 150
ra 100-
100
O'E 75-|
o> g>
C/> O)
O ro
O.LU
X -
LU^
roQ-
o^
00 5J
fc:"
50-
^c
°-| 25 H
0
200
(b)
»
01
S =-
£ S1
150-
100-
§ o so H
fe E
I OH
-50
r= 0.051
R"=0.0026
N = 39.6+0.086C
#= 147
(c)
*
»
•
* *
**
*
»»*
»*
*
**
50 100 150
Ambient Concentration,
200
250
Figure 5-3. Regression analyses of aspects of daytime personal exposure to PM10 estimated
using data from the PTEAM study, (a) Total personal exposure to PM, Et,
regressed on ambient concentration, Ca. (b) Personal exposure to ambient PM,
Eag regressed on Ca. (c) Personal exposure to nonambient PM, Enonag regressed
onC,.
Source: Data taken from Clayton et al. (1993).
March 2001
5-46
DRAFT-DO NOT QUOTE OR CITE
-------
1 The RCS model introduced by Ott et al.(2000) presents a modeling framework to determine
2 the contribution of ambient PM10 and indoor-generated PM10 on personal exposures in large
3 urban metropolitan areas. The model has been tested using personal, indoor and outdoor PM10
4 data from three urban areas (Riverside, CA; Toronto; and Phillipsburg, NJ). Results suggest that
5 it is possible to separate the ambient and nonambient PM contributions to personal exposures on
6 a community-wide basis. However, as discussed in the paper, the authors make some
7 assumptions that require individual consideration in each-city specific application of the model
8 for exposure or health effects investigations. Primarily, housing factors, air-conditioning,
9 seasonal differences, and complexities in time-activity profiles specific to the cohort being
10 studied have to be taken into account prior to adopting the model to a given situation.
11
12 5.4.3.2 Factors That Affect Correlations
13 A number of factors will affect the relationship between personal exposure and PM
14 measured at ambient-site community monitors. Spatial variability in outdoor microenvironments
15 and penetration into indoor microenvironments will influence the relationship for ambient-
16 generated PM, air-exchange rates and decay rates in indoor microenvironments will influence the
17 relationship for both ambient-generated and total PM, whereas personal activities will influence
18 the relationship for total PM but not ambient-generated PM. Information on these effects is
19 presented in detail in the following section.
20
21 5.4.3.2.1 Spatial Variability and Correlations Over Time
22 Chapter 3 (Section 3.2.3) presents information on the spatial variability of PM mass and
23 chemical components at fixed-site ambient monitors; for purposes of this chapter, this spatial
24 variability is called an "ambient gradient". The data presented in Section 3.2.3 indicate that
25 ambient gradients of PM and its constituents exist in urban areas to a greater or lesser degree.
26 This gradient, and any that may exist between a fixed-site monitor and the outdoor jue near where
27 people live, work, and play, obviously affects the exposure. The purpose of this section is to
28 review the available data on ambient monitor-to-outdoor microenvironmental concentration
29 gradients, or relationships, that have been measured by researchers since 1996. A few outdoor-
30 to-outdoor monitoring studies also are included to highlight relationships among important jue
31 categories. To assess spatial variability or gradients, the spatial correlations in the data are
March 2001 5-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 usually analyzed. However, it should be noted that high temporal correlation between two
2 monitoring locations does not imply low spatial variability or low ambient gradients. High
3 temporal correlation between two sites indicates that changes in concentrations at one site can be
4 estimated from data at another site.
5 Oglesby et al. (2000), in a paper on the EXPOLIS-EAS study, conclude that very little
6 spatial variability exists in Basel, Switzerland, between PM levels measured at fixed site
7 monitors and the participant's outdoor (j,e. The authors report a high correlation between home
8 outdoor PM25 levels (48-h measurements beginning and ending at 8:00 a.m.) and the
9 corresponding 24-h average PM4 (time-weighted values calculated from midnight to midnight)
10 measured at a fixed monitoring station (n = 38, rsp = 0.96, p < 0.001). They considered each
11 home outdoor monitor as a temporary fixed monitor and concluded that "the PM2 5 level
12 measured at home outdoors . . . represents the fine particle level prevailing in the city of Basel
13 during the 48-h measuring period
14 In a study conducted in Helsinki, Finland, Buzorius et al. (1999) conclude that a single
15 monitor may be used to adequately describe the ambient gradient across the metropolitan area.
16 Particle size distributions were measured using a differential mobility particle sizer (DMPS;
17 Wintlmayer) coupled with a condensation particle counter (CPC TSI 3010, 3022) at four
18 locations including the official air monitoring station, which represented a "background" site.
19 The monitoring period varied between 2 weeks and 6 mo for the sites and data were reported for
20 10-min and 1-, 8-, and 24-h averages. As expected, temporal variation decreased as the
21 averaging time increased. The authors report that particle number concentration varied in
22 magnitude with local traffic intensity. Linear correlation coefficients computed for all possible
23 site-pairs and averaging times showed that the correlation coefficient improved with increasing
24 averaging time. Using wind speed and direction vectors, lagged correlations were calculated and
25 were generally higher than the "raw" data correlations. Weekday correlations were higher than
26 weekend correlations as "traffic provides relatively uniform spatial distribution of particulate
27 matter" (p. 565). The authors conclude that, even for time periods of 10 min and 1 h, sampling at
28 one station can describe changes across relatively large areas of the city with a correlation
29 coefficient >0.7.
30 Dubowsky et al. (1999) point out that, although the variation of PM25 mass concentration
31 across a community may be small, there may be significant spatial variations of specific
March 2001 5-48 DRAFT-DO NOT QUOTE OR CITE
-------
1 components of the total mass on a local scale. An example is given of a study of concentrations
2 of polycyclic aromatic hydrocarbons (PAH) at three indoor locations in a community; an urban
3 and a semi-urban site separated by 1.6 km, and a suburban site located further away. The authors
4 found the geometric mean PAH concentrations at these three locations varied respectively as
5 31:19:8 ng/m3, and suggest that the local variations in traffic density were responsible for this
6 gradient. Note that these concentrations are 1,000 times lower than the total PM mass
7 concentration, so that such a small gradient would not be detectable for total PM2 5 mass
8 measurements on the order of 25 //g m"3.
9 Leaderer et al. (1999) monitored 24-h PM10, PM2 5, and sulfates during the summers of
10 1995 and 1996 at a regional site in Vinton, VA (6 km from Roanoke, VA). One similar 24-h
11 measurement was made outdoors at residences in the surrounding area, at distances ranging from
12 1 km to >175 km from the Vinton site, at an average separation distance of 96 km. The authors
13 reported significant correlations for PM25 and sulfates between the residential outdoor values and
14 those measured at Vinton on the same day. In addition, the mean values of the regional site and
15 residential site PM25 and sulfates showed no significant differences in spite of the large distance
16 separations and mountainous terrain intervening in most directions. However, for the
17 concentrations of PM2 5_10, estimated as PM10-PM2 5, no significant correlation among these sites
18 was found (n = 30, r = -0.20).
19 Lillquist et al. (1998) found no significant gradient in PM10 concentrations in Salt Lake
20 City, UT, when levels were low, but a gradient existed when levels were high. PM10
21 concentrations were measured outdoor at three hospitals using a Minivol 4.01 sampler
22 (Airmetrics, Inc.) operating at 5 L min"1 and at the Utah Department of Air Quality (DAQ)
23 ambient monitoring station located between 3 and 13 km from the hospitals for a period of about
24 5 mo.
25 Pope et al. (1999) monitored ambient PM10 concentrations in Provo, UT (Utah Valley),
26 during the same time frame the following year and reported nearly identical concentrations at
27 three sites separated by 4 to 12 km. Pearson correlation coefficients for the data were between
28 0.92 and 0.96. The greater degree of variability in the Salt Lake City PM10 data relative to the
29 Provo data may be related to the higher incidence of wind-blown crustal material in Salt Lake
30 City. Pope et al. (1999) reported that increased health effects in the Utah Valley were associated
31 with stagnation and thermal inversions trapping anthropogenically derived PM10, whereas, no
March 2001 5-49 DRAFT-DO NOT QUOTE OR CITE
-------
1 increases in health effects were observed when PM10 levels were increased during events of wind
2 blown crustal material.
3 Vakeva et al. (1999) found significant vertical gradients in submicron particles existed in
4 an urban street canyon of Lahti, Finland. Particle number concentrations were measured using a
5 TSI screen diffusion battery and a condensation particle counter at 1.5 and 25 m above the street
6 at rooftop level. The authors found a fivefold decrease in concentration between the two
7 sampling heights and attributed the vertical gradient to dilution and dispersion of pollutants
8 emitted at street level.
9 White (1998) suggests that the higher random measurement error for the coarse PM
10 fraction compared to the error for the fine PM fraction may be responsible for a major portion of
11 the apparent greater spatial variability of coarse ambient PM concentration compared to fine
12 ambient PM concentration in a community (e.g., Burton et al., 1996; Leaderer et al., 1999).
13 When PM25 and PM10 are collected independently, and the coarse fraction is obtained by
14 difference (PM25_10 = PM10-PM25), then the expected variance in the coarse fraction is the sum of
15 the variances of the PM10 and PM25 measurements. When a dichotomous sampler collects PM25
16 and PM2 5_10 on two separate filters, the coarse fraction also is expected to have a larger error than
17 the fine fraction. There is a possible error caused by loss of mass below the cut-point size and a
18 gain of mass above the cut-point size that is created by the asymmetry of the product of the
19 penetration times PM concentration about the cut-point size. Because a dichotomous PM
20 sampler collects coarse mass using an upper and lower cut-point, it is expected to have a larger
21 variance than for the fine mass collected using the same lower cut-point.
22 Wilson and Suh (1997) conclude that PM25 and PM10 concentrations are correlated more
23 highly across Philadelphia than are PM2 5_10 concentrations. Ambient monitoring data from 1992
24 to 1993 was reviewed for PM2 5, PM2 5_10, and PM10, as well as for PM25 and PM25.10 dichotomous
25 data for 212 site-years of information contained in the AIRS database. The authors also observed
26 that PM10 frequently was correlated more highly with PM2 5 than with PM2 5_10. The authors note
27 that PM2 5 constitutes a large fraction of PM10, and that this is the likely reason for the strong
28 agreement between PM25 and PM10. Similar observations were made by Keywood et al. (1999)
29 in six Australian cities. The authors reported that PM10 was more highly correlated with PM2 5
30 than with coarse PM (PM25.10), suggesting that "variability in PM10 is dominated by variability in
31 PM25."
March 2001 5-50 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lippmann et al. (2000) examined the site-to-site temporal correlations in Philadelphia
2 (1981 to 1994) and found the ranking of median site-to-site correlation was O3 (0.83), PM10
3 (0.78), TSP (0.71), NO2 (0.70), CO (0.50), and SO2 (0.49). The authors explain that O3 and a
4 fraction of TSP and PM10 (e.g., sulfate) are secondary pollutants that would tend to be distributed
5 spatially more uniformly within the city than primary pollutants such as CO and SO2, which are
6 more likely to be influenced by local emission sources. Lippman et al. (2000) conclude: "Thus,
7 spatial uniformity of pollutants may be due to area-wide sources, or to transport (e.g., advection)
8 of fairly stable pollutants into the urban area from upwind sources. Relative spatial uniformity of
9 pollutants would therefore vary from city to city or region to region."
10
11 5.4.3.2.2 Physical Factors Affecting Indoor Microenvironmental Particulate Matter
12 Concentrations
13 Several physical factors affect ambient particle concentrations in the indoor jue, including
14 air exchange, penetration, and particle deposition. Combined, these factors are critical variables
15 that describe ambient particle dynamics in the indoor jue and, to a large degree, significantly
16 affect an individual's personal exposure to ambient-generated particles while indoors. The
17 relationship between ambient outdoor particles and ambient particles that have infiltrated indoors
18 is given by
19
20 Cai/Cao=Ptf/O + k), (5-10)
21
22 where C^ and Cao are the concentration of ambient indoor and outdoor particles, respectively;
23 P is the penetration factor; a is the air exchange rate; and k is the particle deposition rate (as
24 discussed in Section 5.3.2.3.1, use of this model assumes equilibrium conditions and assumes
25 that all variables remain constant). Particle penetration is a dimensionless quantity that describes
26 the fraction of ambient particles that effectively penetrates the building shell. "Air exchange" is
27 a term used to describe the rate at which the indoor air in a building or residence is replaced by
28 outdoor air. The dominant processes governing particle penetration are air exchange and
29 deposition of particles as they traverse through cracks and crevices and other routes of entry into
30 the building. Although air-exchange rates have been measured in numerous studies, very few
31 field data existed prior to 1996 to determine size-dependent penetration factors and particle
March 2001 5-51 DRAFT-DO NOT QUOTE OR CITE
-------
1 deposition rates. All three parameters (P, a, and K) may vary substantially depending on building
2 type, region of the country, and season. In the past several years, researchers have made
3 significant advancements in understanding the relationship between particle size and penetration
4 factors and particle deposition rates. This section will highlight the studies that have been
5 conducted to better understand physical factors affecting indoor particle dynamics.
6
7 Air-Exchange Rates
8 The air-exchange rate, a, in a residence varies depending on a variety of factors, including
9 geographical location, age of the building, the extent to which window and doors are open, and
10 season. Murray and Burmaster (1995) used measured values of a from households throughout
11 the United States to describe empirical distributions and to estimate univariate parametric
12 probability distributions of air-exchange rates. Figure 5-4 shows the results classified by season
13 and region. In general, a is highest in the warmest region and increases from the coldest to the
14 warmest region during all seasons. Air-exchange rates also are quite variable within and between
15 seasons, as well as between regions (Figure 5-4). Data from the warmest region in summer
16 should be viewed cautiously as many of the measurements were made in Southern California in
17 July when windows were more likely to be open than in other areas of the country where
18 air-conditioning is used. Use of air-conditioning generally results in lowering air-exchange rates.
19 In a separate analyses of these data, Koontz and Rector (1995) suggested that a conservative
20 estimate for air exchange in residential settings would be 0.18 h"1 (10th percentile) and a typical
21 air exchange would be 0.45 h"1 (50th percentile).
22 These data provide reasonable experimental evidence that a varies by season in locations
23 with distinct seasons. As a result, infiltration of ambient particles may be more efficient during
24 warmer seasons when windows are likely to be opened more frequently and air-exchange rates
25 are higher. This suggests that the fraction of ambient particles present in the indoor ^e would be
26 greater during warmer seasons than colder seasons. For example, in a study conducted in
27 Boston, MA, participants living in non-air-conditioned homes kept the windows closed except
28 during the summer (Long et al., 2000a). This resulted in higher and more variable air-exchange
29 rates in summer than during any other season (Figure 5-5). During nighttime periods, when
30 indoor sources are negligible, the indoor/outdoor concentration ratio or infiltration factor may be
31 used to determine the relative contribution of ambient particles in the indoor ^e. Particle data
March 2001 5-52 DRAFT-DO NOT QUOTE OR CITE
-------
3 -
2 -
e
«
0
Coldest Region
Colder Region
Warmer Region
Warmest Region
i
I
Winter Spring Summer
Season
Fall
Figure 5-4. Air-exchange rates measured in homes throughout the United States. Climatic
regions are based on heating-degree days: Coldest region > 7000, Colder
region = 5500 to 6999, Warmer region = 2500 to 4999, and Warmest region
< 2500 heating-degree days.
Based on data from Murray and Burmaster (1995).
1
2
3
4
5
collected during this study (Figure 5-6) shows the indoor/outdoor concentration ratios by particle
size. Data show that for these nine homes in Boston, the fraction of ambient particles penetrating
indoors is higher during summer when air exchange rates were higher than fall (Long et al.,
2000b).
March 2001
5-53
DRAFT-DO NOT QUOTE OR CITE
-------
o -
7 -
f 6-
ge Rate (exchanges
co -t*. en
c
03
-^ 2 -
0 '
X
LJJ
$ 1-
0 -
•
T
• -T-
I , j- I
Fall Winter Spring
Season
$95%
90%
75%
?ROA
* 10%
*5%
Summer
Figure 5-5. Box plots of hourly air-exchange rates stratified by season in Boston, MA,
during 1998.
Source: Long et al. (2000a).
1 Particle Deposition Rates and Penetration Factors
2 Physical factors affecting indoor particle concentrations, including particle deposition rates,
3 k, and penetration factors, P, are possibly the most uncertain and variable quantities. Although k
4 can be modeled with some success, direct measurements are difficult and results often vary from
5 study to study. Particle deposition rates vary considerably depending on particle size because of
6 the viscous drag of air on the particles hindering their movement to varying degrees. The nature
7 and composition of particles also affect deposition rates. Surface properties of particles, such as
8 their electrostatic properties, can have a significant influence on deposition rates. In addition,
9 thermophoresis can also affect k, but probably to a lesser degree in the indoor ^e because
March 2001
5-54
DRAFT-DO NOT QUOTE OR CITE
-------
a) Home NEW2
€
03
O
'•*—•
03
1.1
1.0 -
0.9 -
0.8 -
0.7 -
0.6 -
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
0.0
co
o
0.1
\
Summer Fall
S 8 § 5
9
CO
p
o
p
d
CD
<=>.
c\j co TJ- in Y
o o o o Is-;
LO CN co 4 o
^ ci o c>
Particle Size (|jm)
CO
CNI
o
cb
Figure 5-6. Geometric mean infiltration factor (indoor/outdoor ratio) for hourly nighttime,
nonsource data for two seasons. Box plots of air exchange rates are shown as
inserts for each plot. (Boston, 1998)
Source: Longetal. (2000b).
1 temperatures generally vary over a small range. Combined, these effects can produce order of
2 magnitude variations in k between particles of different size and, in the case of electrophoresis
3 and thermophoresis, particles of the same size.
4 Particle penetration efficiency into the indoor ^e depends on particle size and air exchange
5 rates. Penetration varies with particle size because of the size-dependent deposition of particles
6 caused by impact!on, interception, and diffusion of particles onto surfaces as they traverse
7 through cracks and crevices. Penetration also is affected by air exchange rates. When air
8 exchange rates are high, P approaches unity because the majority of ambient particles have less
9 interaction with the building shell. In contrast, when air exchange rates are low, P is governed by
10 particle deposition as particles travel through cracks and crevices.
March 2001
5-55
DRAFT-DO NOT QUOTE OR CITE
-------
1 Significant advancements have been made in the past few years to better characterize
2 particle deposition rates and penetration factors. Several new studies, including two in which
3 semi-continuous measurements of size distributions were measured indoors and outdoors, have
4 produced new information on these quantities, which are key to understanding the contributions
5 of ambient PM to indoor PM concentrations (Equation 5-10).
6 Studies involving semi-continuous measurements of indoor and outdoor particle size
7 distributions have been used to estimate k and P as a function of particle size (Vette et al., 2001;
8 Long et al., 2000b; Abt et al., 2000b). These studies each demonstrated that the indoor/outdoor
9 concentration ratios (Cao/C in Equation 5-11) were highest for accumulation mode particles and
10 lowest for ultrafme and coarse-mode particles. Various approaches were used to estimate size-
11 specific values for k and P. Vette et al. (2001) and Abt et al. (2000b) estimated k by measuring
12 the decay of particles at times when indoor levels were significantly elevated. Vette et al. (2001)
13 estimated P using measured values of k and indoor/outdoor particle measurements during
14 nonsource nighttime periods. Long et al. (2000b) used a physical-statistical model, based on
15 Equation 5-10, to estimate k and P during nonsource nighttime periods. The results for k
16 reported by Long et al. (2000b) and Abt et al. (2000b) are compared with other studies in
17 Figure 5-7. Although not shown in Figure 5-7, the results for k obtained by Vette et al. (2001)
18 were similar to the values of & reported by Abt et al. (2000b) for particle sizes up to 1 //m.
19 Results for P by Long et al. (2000b) show that penetration was highest for accumulation-mode
20 particles and decreased substantially for coarse-mode particles (Figure 5-8). The results for
21 P reported by Vette et al. (2001) show similar trends, but are lower than those reported by Long
22 et al. (2000b). This likely is because of lower air-exchange rates in the Fresno, CA, residence
23 (a ~ 0.5 h4; Vette et al., 2001) than the Boston, MA, residences (a > 1 h'1; Long et al., 2000b).
24 These data for P and k illustrate the role that the building shell may provide in increasing the
25 concentration of particles because of indoor sources and reducing the concentration of indoor
26 particles from ambient sources, especially for homes with low air-exchange rates.
27
28 Compositional Differences Between Indoor-Generated and Ambient-Generated
29 Particulate Matter
30 Wilson et al. (2000) discuss the differences in composition between particles from indoor
31 and outdoor sources. They note that, because of the difficulty in separating indoor PM into
March 2001 5-56 DRAFT-DO NOT QUOTE OR CITE
-------
100-
10-;
03
O
-------
0)
'o
E
LU
c
g
16
"CD
c
CD
CL
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
TO
o:
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4 £
0.3
0.2
0.1
0.0
00 •'sT
O O
9 9
CM CO
q q
o o
8
o
s
ci
o o
9 06
CD o
o d
CN
o
cq
9
-------
1 PM in the ambient/outdoor air generally is produced by burning fossil fuels (e.g., coal, gasoline,
2 fuel oil) and wood, whereas combustion-product PM from indoor sources is produced by
3 biomass burning (e.g., tobacco, wood, foods, etc.). However, some indoor sources of PM (such
4 as cigarette smoking, meat cooking, and coal burning) occur both indoors and outdoors and may
5 constitute an identifiable portion of measured ambient PM (Cha et al., 1996; Kleeman and Cass,
6 1998).
7
8 Indoor Air Chemistry
9 Gas- and aerosol-phase chemical reactions in the indoor //e are responsible for secondary
10 particle formation and modification of existing particles. Homogeneous gas phase reactions
11 involving ozone and terpenes (specifically d-limonene, a-terpinene, and a-pinene) have been
12 identified as an important source of submicron particles (Weschler and Shields, 1999). Terpenes
13 are present in several commonly available household cleaning products and d-limonene has been
14 identified in more than 50% of the buildings monitored in the BASE study (Hadwen et al., 1997).
15 Long et al. (2000a) found that when PineSol (primary ingredient is a-pinene) was used indoors,
16 indoor PM25 mass concentrations increased by 3 to 32 //g m"3 (indoor ozone concentrations
17 unknown, but ambient ozone concentrations were 44 to 48 ppb). Similarly, a 10-fold increase in
18 number counts of 0.1 to 0.2 //m particles was observed in an experimental office containing
19 supplemented d-limonene and normally encountered indoor ozone concentrations (< 5 to
20 45 ppb), resulting in an average increase in particle mass concentration of 2.5 to 5.5 //g m"3
21 (Weschler and Shields, 1999). Ozone appears to be the limiting reagent as particle number
22 concentration varied proportionally to ozone concentrations (Weschler and Shields, 1999). Other
23 studies showed similar findings (e.g., Jang and Kamens, 1999; Wainman et al., 2000).
24
25 Indoor Sources of Particles
26 The major sources of indoor PM in nonsmoking residences and buildings include
27 suspension of PM from bulk material, cooking, cleaning, and the use of combustion devices,
28 such as stoves and kerosene heaters. Human and pet activities also lead to PM detritus
29 production (from tracked-in soil, fabrics, skin and hair, home furnishings, etc.), which is found
30 ubiquitously in house dust deposited on floors and other interior surfaces. House dust and lint
31 particles may be resuspended indoors by agitation (cleaning) and turbulence (HVAC systems,
March 2001 5-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 human activities, etc.). Ambient particles that have infiltrated into the indoor jue also may be
2 resuspended after deposition to indoor surfaces. Typically, resuspension of particles from any
3 source involves coarse-mode particles (>1 //m); particles of smaller diameter are not resuspended
4 efficiently. On the other hand, cooking produces both fine- and coarse-mode particles, whereas
5 combustion sources typically produce fine-mode particles.
6 Environmental tobacco smoke (ETS) is also a major indoor source of PM. It is, however,
7 beyond the scope of this chapter to review the extensive literature on ETS. A number of articles
8 provide source strength information for cigarette or cigar smoking (e.g., Daisey et al. [1998] and
9 Nelson et al. [1998]).
10 A study conducted on two homes in the Boston metropolitan area (Abt et al., 2000a)
11 showed that indoor PM sources predominate when air exchange rates were <1 h"1, and outdoor
12 sources predominate when air exchange rates were >2 h"1. The authors attributed this to the fact
13 that when air-exchange rates were low (<1 h"1), particles released from indoor sources tend to
14 accumulate because particle deposition is the mechanism governing particle decay and not air
15 exchange. Particle deposition rates are generally <1 h"1, especially for accumulation-mode
16 particles. When air-exchange rates were higher (>2 h"1), infiltration of ambient aerosols and
17 exfiltration of indoor-generated aerosols occur more rapidly, reducing the impact of indoor
18 sources on indoor particle levels. The study also confirmed previous findings that the major
19 indoor sources of PM are cooking, cleaning, and human activity. They discuss the size
20 characteristics of these ubiquitous sources and report the following.
21
22 The size of the particles generated by these activities reflected their formation processes.
23 Combustion processes (oven cooking, toasting, and barbecuing) produced fine particles and
24 mechanical processes (sauteing, frying, cleaning, and movement of people) generated coarse
25 particles. These activities increased particle concentrations by many orders of magnitude higher
26 than outdoor levels and altered indoor size distributions. (Abt et al., 2000a; p. 43)
27
28 They also note that variability in indoor PM for all size fractions was greater than for outdoor
29 PM, especially for short averaging times (2 to 33 times higher).
30 In a separate study conducted in nine nonsmoking homes in the Boston area, Long et al.
31 (2000a) concluded that the predominant source of indoor fine particles was infiltration of outdoor
March 2001 5-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 particles, and that cooking activities were the only other significant source of fine particles.
2 Coarse particles, however, had several indoor sources, such as cooking, cleaning, and various
3 indoor activities. This study also concluded that more than 50% of the particles (by volume)
4 generated during indoor events were ultrafine particles. Events that elevated indoor particle
5 levels were found to be brief, intermittent, and highly variable, thus requiring the use of
6 continuous instrumentation for their characterization. Table 5-8 provides information on the
7 mean volume mean diameter (VMD) for various types of indoor particle sources. The
8 differences in mean VMD confirm the clear separation of source types and suggest that there is
9 very little resuspension of accumulation-mode PM. In addition, measurements of organic and
10 elemental carbon indicated that organic carbon had significant indoor sources, whereas elemental
11 carbon was primarily of ambient origin.
12 Vette et al. (2001) found that resuspension was a significant indoor source of particles
13 >1 //m, whereas fine- and accumulation-mode particles were not affected by resuspension.
14 Figure 5-9 shows the diurnal variability in the indoor/outdoor aerosol concentration ratio from an
15 unoccupied residence in Fresno. The study was conducted in the absence of common indoor
16 particle sources such as cooking and cleaning. The data in Figure 5-9 show the mean
17 indoor/outdoor concentration ratio for particles >1 //m increased dramatically during daytime
18 hours. This pattern was consistent with indoor human activity levels. In contrast, the mean
19 indoor/outdoor concentration ratio for particles <1 //m (fine- and accumulation-mode particles)
20 remain fairly constant during both day and night.
21
22 5.4.3.2.3 Time/Activity Patterns
23 Total exposure to PM is the sum of various microenvironmental exposures that an
24 individual encounters during the day and will depend on the microenvironments occupied.
25 As discussed previously, PM exposure in each microenvironment is the sum of exposures from
26 ambient sources (Eag), indoor sources (Eig), and personal activities (Epact). Eag and Eig are
27 determined by the microenvironments in which an individual spends time; whereas Epact is
28 determined by the personal activities that he/she conducts while in those microenvironments.
29 Determining microenvironments and activities that contribute significantly to human
30 exposure begins with establishing human activity pattern information for the general population,
31 as well as subpopulations. Personal exposure and time activity pattern studies have shown that
March 2001 5-61 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-8. VOLUME MEAN DIAMETER (VMD) AND MAXIMUM PM2 5
CONCENTRATIONS OF INDOOR PARTICLE SOURCESab
Size Statistics
Particle Source
Cooking
Baking (Electric)
Baking (Gas)
Toasting
Broiling
Sauteing
Stir-Frying
Frying
Barbecuing
Cleaning
Dusting
Vacuuming
Cleaning with Pine Sol
General Activities
Walking Vigorously (w/Carpet)
Sampling w/Carpet
Sampling w/o Carpet
Burning Candles
N
8
24
23
4
13
o
J
20
2
11
10
5
15
52
26
7
Indoor Activity
Mean VMD
(urn)
0.189f
0.107f
0.138f
0.1 14f
0.184f, 3.48g
0.135f
0.173f
0.159f
5.38g
3.86B
0.097f
3.96g
4.25g
4.28B
0.311f
Background3'6
Mean VMD
(urn)
0.221f
0.224f
0.222f
0.236f
0.223f, 2.93g
0.277f
0.223f
0.205f
3.53g
2.79g
0.238f
3.18g
2.63g
2.93g
0.224f
Maximum
Mean
14.8
101.2
54.9
29.3
65.6
37.2
40.5
14.8
22.6
6.5
11.0
12.0
8.0
4.8
28.0
PM25
Concentration°'d
SD
7.4
184.9
119.7
43.4
95.4
31.4
43.2
5.2
22.6
3.9
10.2
9.1
6.6
3.0
18.0
Notes:
aAll concentration data corrected for background particle levels.
Includes only individual particle events that were unique for a given time period and could be detected above
background particle levels.
°PM concentrations in ^g/m3.
dMaximum concentrations computed from 5-min data for each activity.
background data are for time periods immediately prior to the indoor event.
fSize statistics calculated for PV002.05 using SMPS data.
gSize statistics calculated for PV0 7_10 using APS data.
Source: Long et al. (2000a).
1 different populations have varying time activity patterns and, accordingly, different personal PM
2 exposures. Both characteristics will vary greatly as a function of age, health status, ethnic group,
3 socioeconomic status, season, and region of the country. Collecting detailed time activity data
4 can be very burdensome on participants but is clearly valuable in assessing human exposure and
March 2001
5-62
DRAFT-DO NOT QUOTE OR CITE
-------
o
o
3
O
~£j
o
o
-o
d
eS
o
0.0
06
Time
Figure 5-9. Mean hourly indoor/outdoor particle concentration ratio from an unoccupied
residence in Fresno, CA, during spring 1999.
Source: Vette etal. (2001).
1 microenvironments. For modeling purposes, human activity data frequently come from general
2 databases that are discussed below.
3 The gathering of human activity information, often called "time-budget" data, started in the
4 1920s; however, their use for exposure assessment purposes only began to be emphasized in the
5 1980s. Many of the largest U.S. human activity databases have been consolidated by EPA's
6 National Exposure Research Laboratory's (NERL) into one comprehensive database containing
7 over 22,000 person-days of 24-h activity known as the Consolidated Human Activity Database,
8 or CHAD (Glen et al., 1997). The information in CHAD will be accessible for constructing
9 population cohorts of people with diverse characteristics that are useful for analysis and
10 modeling (McCurdy, 2000). See Table 5-2 for a summary listing of human activity studies in
11 CHAD. Most of the databases in CHAD are available elsewhere, including the National Human
March 2001
5-63
DRAFT-DO NOT QUOTE OR CITE
-------
1 Activity Pattern Survey (NHAPS), California's Air Resources Board (CARB), and the University
2 of Michigan's Institute for Survey Research data sets.
3 Although CHAD provides a very valuable resource for time and location data, there is less
4 information on PM generating personal activities. In addition, very few of the time-activity
5 studies have collected longitudinal data within a season or over multiple seasons. Such
6 longitudinal data are important in understanding potential variability in activities and how they
7 impact correlations between PM exposure and ambient site measurements for both total PM and
8 PM of ambient origin.
9
10 5.4.3.3 Impact of Ambient Sources on Exposures to Particulate Matter
11 Different sources may generate ambient PM with different aerodynamic and chemical
12 characteristics, which may, in turn, result in different health responses. Thus, to fully understand
13 the relationship between PM exposure and health outcome, exposure from difference sources
14 should be identified and quantified. Source apportionment techniques provide a method for
15 determining personal exposure to PM from specific sources. Daily contributions from sources
16 that have no indoor component can be used as tracers to generate exposure to ambient PM of
17 similar aerodynamic size or directly as exposure surrogates in epidemiologic analyses. The
18 recent EPA PM Research Needs Document (U. S. Environmental Protection Agency, 1998)
19 recommended use of source apportionment techniques to determine daily time-series of source
20 categories for use in community, time-series epidemiology.
21 A number of epidemiological studies (discussed more fully in Chapter 6) have evaluated
22 the relationship between health outcomes and sources of particulate matter determined from
23 measurements at a community monitor. These studies suggest the importance of examining
24 sources and constituents of indoor, outdoor, and personal PM. Ozkaynak and Thurston (1987)
25 evaluated the relationship between parti culate matter sources and mortality in 36 Standard
26 Metropolitan Statistical Areas (SMSAs). Particulate matter samples from EPA's Inhalable
27 Particle (IP) Network were analyzed for SO42" and NO3" by automated colorimetry, and elemental
28 composition was determined with X-ray fluorescence (XRF). Mass concentrations from five
29 particulate matter source categories were determined from multiple regression of absolute factor
30 scores on the mass concentration: (1) resuspended soil, (2) auto exhaust, (3) oil combustion,
31 (4) metals, and (5) coal combustion.
March 2001 5-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 Mar et al. (2000) applied factor analysis to evaluate the relationship between PM
2 composition (and gaseous pollutants) in Phoenix. In addition to daily averages of PM25 elements
3 from XRF analysis, they included in their analyses organic and elemental carbon in PM2 5 and
4 gaseous species emitted by combustion sources (CO, NO2, and SO2). They identified five factors
5 classified as (1) motor vehicles, (2) resuspended soil, (3) vegetative burning, (4) local SO2, and
6 (5) regional sulfate.
7 Laden et al. (2000) applied specific rotation factor analysis to particulate matter
8 composition (XRF) data from six eastern cities (Ferris et al., 1979). Fine particulate matter was
9 regressed on the recentered scores to determine the daily source contributions. Three main
10 sources were identified: (1) resuspended soil (Si), (2) motor vehicle (Pb), and (3) coal
11 combustion (Se).
12 Source apportionment or receptor modeling has been applied to the personal exposure data
13 to understand the relationship between personal and ambient sources of particulate matter.
14 Application of source apportionment to ambient, indoor, and personal PM composition data is
15 especially useful in sorting out the effects of particle size and composition. If a sufficient
16 number of samples are analyzed with sufficient compositional detail, it is possible to use
17 statistical techniques to derive source category signatures, identify indoor and outdoor source
18 categories and estimate their contribution to indoor and personal PM.
19 Positive Matrix Factorization (PMF) has been applied to the PTEAM database by
20 Yakovleva et al. (1999). The authors utilize mass and XRF elemental composition data from
21 indoor and outdoor PM2 5 and personal, indoor, and outdoor PM10 samples. PMF is an advance
22 over ordinary factor analysis because it allows measurements below the quantifiable limit to be
23 used by weighting them by their uncertainty. This effectively increases the number of species
24 that can be used in the model. The factors used by the authors correspond to general source
25 categories of PM, such as outdoor soil, resuspended indoor soil, indoor soil, personal activities,
26 sea-salt, motor vehicles, nonferrous metal smelters, and secondary sulfates. PMF, by identifying
27 not only the various source factors but also apportioning them among the different monitor
28 locations (personal, indoor, and outdoor), was able to quantify an estimate of the contribution of
29 resuspended indoor dust to the personal cloud (15% from indoor soil and 30% from resuspended
30 indoor soil). Factor scores for these items then were used in a regression analysis to estimate
31 personal exposures (Yakovleva et al., 1999).
March 2001 5-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 The most important contributors to PM10 personal exposure were indoor soil, resuspended
2 indoor soil, and personal activities; these accounted for approximately 60% of the mass
3 (Yakovleva et al., 1999). Collectively, they include personal cloud PM, smoking, cooking, and
4 vacuuming. For both PM2 5 and PM10, secondary sulfate and nonferrous metal operations
5 accounted for another 25% of PM mass. Motor vehicle exhausts, especially starting a vehicle
6 inside of an attached garage, accounted for another 10% of PM mass. The authors caution that
7 these results may not apply to other geographic areas, seasons of the year, or weather conditions.
8 Simultaneous measurement of personal (PM10) and outdoor measurements (PM2 5 and
9 PM10) were evaluated as a three-way problem with PMF, which allowed for differentiation of
10 source categories based on their variation in time and type of sample, as well as their variation in
11 composition. By use of this technique, it was possible to identify three sources of coarse-mode,
12 soil-type PM. One was associated with ambient soil, one was associated with indoor soil
13 dispersed throughout the house, and one was associated with soil resulting from the personal
14 activity of the subject.
15 Two other source apportionment models have been applied to ambient measurement data
16 and can be used for the personal exposure studies. The effective variance weighted Chemical
17 Mass Balance (CMB) receptor model (Watson et al., 1984, 1990, 1991) solves a set of linear
18 equations that incorporate the uncertainty in the sample and source composition. CMB requires
19 the composition of each potential source of PM and the uncertainty for the sources and ambient
20 measurements. Source apportionment with CMB can be conducted on individual samples,
21 however, composition of each of the sources of PM must be known. An additional source
22 apportionment model, UNMIX (Henry et al., 1994) is a multivariate source apportionment
23 model. UNMIX is similar to PMF, but does not use explicitly the measurement uncertainties.
24 Because measurement uncertainties are not used, only species above the detection limit are
25 evaluated in the model. UNMIX provides the number of sources and source contributions and
26 requires a similar number of observations as PMF.
27 The Yakovleva et al. (1999) study demonstrates that source apportionment techniques also
28 could be very useful in determining parameters needed for exposure models and for determining
29 exposure to ambient-generated PM. Exposure information, similar to that obtained in the
30 PTEAM study, but including other PM components useful for definition of other source
31 categories (e.g., elemental [EC] and organic carbon [OC]; organic tracers for elemental carbon
March 2001 5-66 DRAFT-DO NOT QUOTE OR CITE
-------
1 from diesel vehicle exhaust, gasoline vehicle exhaust, and wood combustion; nitrate; Na; Mg and
2 other heavy metal tracers; and, also, gas-phase pollutants) would be useful as demonstrated in the
3 use of EC/OC and gas-phase pollutants by Mar et al. (2000).
4
5 5.4.3.4 Correlations of Particulate Matter with Other Pollutants
6 Several epidemiological studies have included the gaseous pollutants CO, NO2, SO2, and
7 O3 along with PM10 or PM25 in the analysis of the statistical association of health responses with
8 pollutants. In a recent study, the personal exposure to O3 and NO2 were determined, as well as
9 that to PM25 and PM25_10 for a cohort 15 elderly subjects in Baltimore, MD. Sarnat et al. (2000)
10 conclude that the potential for confounding of PM25 by O3, NO2, or PM 25_10 appears to be
11 limited, because, despite significant correlations observed among ambient pollutant
12 concentrations, the correlations among personal exposures were low. Spearman correlations for
13 14 subjects in summer and 14 subjects in winter are given in Table 5-9 for relationships between
14 personal PM2 5 and ambient concentrations of PM2 5, PM2 5_10, O3, and NO2. In contrast to ambient
15 concentrations, neither personal exposure to total PM2 5 nor PM25 ambient origin was correlated
16 significantly with personal exposures to the co-pollutants, PM25.10, nonambient PM25, O3, NO2,
17 and SO2. Personal-ambient associations for PM2 5_10, O3, NO2, and SO2 were similarly weak and
18 insignificant. It should be noted that measured personal exposures to O3, NO2, and SO2 were
19 below their respective LOD for 70% of the samples.
20 A newly developed Roll-Around System (RAS) was used to evaluate the hourly
21 relationship between gaseous pollutants (CO, O3, NO2, SO2, and VOCs) and PM (Chang et al.,
22 2000). Exposures were characterized over a 15-day period for the summer and winter in
23 Baltimore, based on scripted activities to simulate activities performed by older adults (65+ years
24 of age). Spearman rank correlations were reported for PM2 5, O3, CO, and toluene for both the
25 summer and winter and the correlations are given for each microenvironment in Table 5-10:
26 indoor residence, indoor other, outdoor near roadway, outdoor away from road, and in vehicle.
27 No significant relationships (p < 0.05) were found between hourly PM2 5 and O3. Significant
28 relationships were found between hourly PM25 and CO: indoor residence, winter; indoor other,
29 summer and winter; and outdoor away from roadway, summer. Significant relationships also
30 were found between hourly PM25 and toluene: indoor residence, winter; indoor other, winter;
31 and in vehicle, winter. The significant relationships between CO and PM2 5 in the winter may be
March 2001 5-67 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-9. CORRELATIONS BETWEEN PERSONAL PM2 5 AND AMBIENT
POLLUTANT CONCENTRATIONS1
Personal PM2 5
vs. Ambient:
SUMMER Subject
SA1
SA2
SA5
SB1
SB2
SB3
SB4
SB5
SB6
SCI
SC2
SC3
SC4
SC5
WINTER WAI
WA2
WA4
WAS
WB1
WB2
WB3
WB4
WC1
WC2
WC3
WC4
WC5
WC6
Median Summer
Median Winter
PM25
0.55
0.85
0.89
0.65
-0.21
0.82
0.73
0.73
0.53
0.95
0.78
0.85
0.78
0.55
0.22
-0.38
-0.18
0.22
0.80
0.62
0.55
-0.12
0.74
0.79
0.28
0.19
0.81
0.01
0.76
0.25
03
0.15
0.31
0.18
0.40
-0.62
0.55
0.62
0.45
0.15
0.78
0.68
0.78
0.66
0.51
-0.18
-0.07
0.67
-0.43
-0.84
-0.32
-0.45
-0.01
-0.62
-0.88
-0.42
-0.84
-0.62
-0.03
0.48
-0.43
NO2
0.38
0.66
0.82
-0.15
0.81
-0.14
-0.34
-0.42
-0.38
0.66
0.36
0.73
0.59
0.32
-0.26
-0.36
-0.22
0.61
0.77
0.59
0.62
0.34
-0.15
0.17
0.03
0.50
0.08
0.65
0.37
0.26
PM2,.10
-0.12
0.57
0.64
0.38
0.15
-0.04
-0.12
0.23
0.12
0.65
0.51
0.68
0.70
0.43
-0.05
-0.70
-0.29
0.50
0.41
0.09
0.04
-0.10
0.44
0.77
0.57
0.45
0.81
0.37
0.41
0.39
Personal PM
of Ambient Origin vs
03
0.27
0.21
0.33
0.89
0.26
0.52
0.45
0.36
-0.03
0.83
0.66
0.69
0.50
0.34
-0.78
-0.15
-0.33
-0.72
-0.87
-0.76
-0.77
-0.80
-0.64
-0.87
-0.77
-0.72
-0.76
-0.75
0.41
-0.76
NO2
0.71
0.64
0.81
-0.74
0.08
-0.20
-0.29
-0.48
-0.57
0.63
0.65
0.71
0.50
0.33
-0.04
-0.15
0.20
-0.09
0.53
0.59
0.56
0.68
0.02
0.25
0.30
0.22
0.05
0.19
0.42
0.21
2.5
. Ambient:
PM2,.10
0.15
0.68
0.79
-0.03
0.33
0.00
-0.14
0.33
0.32
0.57
0.76
0.80
0.51
0.27
-0.24
0.02
0.00
0.40
0.66
0.59
0.60
0.48
0.69
0.71
-0.45
0.67
0.42
-0.45
0.33
0.45
'Correlations represent Spearman's rvalues; italicized values indicate significance at the a = 0.05 level.
Source: Sarnat et al. (2000).
March 2001
5-68
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-10. CORRELATIONS BETWEEN HOURLY PERSONAL PM2 5 AND
GASEOUS POLLUTANTS
Indoor
Residence
PM2.5vs.
Summer
Winter
PM2.5vs.
Summer
Winter
PM2.svs.
Summer
Winter
N
03
35
56
CO
41
59
Toluene
46
66
rs
0.29
0.05
0.25
0.43a
0.23
0.38a
Indoor Other
N
16
37
19
39
21
47
rs
-0.14
-0.06
0.59a
0.62a
-0.14
0.44a
Outdoor Near
Roadway
N
10
11
13
13
14
17
rs
0.05
-0.28
0.14
0.37
0.26
0.40
Outdoor Away
from road
N
12
7
12
8
14
8
rs
0.45
0.04
0.62
0.41
0.02
0.48
In Vehicle
N
37
34
46
37
48
42
rs
0.21
-0.10
0.23
0.10
0.12
0.43a
aCorrelations represent Spearman's rvalues; italicized values indicate significance at the a, = 0.05 level.
Source: Chang et al. (2000).
1 caused by reduced air-exchange rates that could allow them to accumulate (Chang et al., 2000).
2 Although no significant correlation was found between in vehicle PM2 5 and CO, toluene, which
3 is a significant component of vehicle exhaust (Conner et al., 1995), was correlated significantly
4 to PM2 5 in the winter.
5 Carrer et al. (1998) present data on the correlations among personal and
6 microenvironmental PM10 exposures and concentrations and selected environmental chemicals
7 that were monitored simultaneously (using a method that was not described). These chemicals
8 were nitrogen oxides (NOX), carbon monoxide (CO), and total volatile organic compounds
9 (TVOC), benzene, toluene, xylene, and formaldehyde. The Kendall i correlation coefficient was
10 used; only results significant at p < 0.05 are mentioned here. Significant associations were found
11 only between the following pairs of substances (u shown in parentheses): personal PM10 (24 h)
12 andNOx (0.34), CO (0.34), TVOC (0.18), toluene (0.19), and xylene (0.26); office PM10 andNOx
13 (0.31); home PM10 andNOx (0.24), CO (0.24), toulene (0.17), and xylene (0.25). Surprisingly,
March 2001
5-69
DRAFT-DO NOT QUOTE OR CITE
-------
1 because most of the chemical substances are associated with motor vehicular emissions, there
2 was no significant correlation between "commuting PM10" and any of the substances (Carrer
3 etal., 1998).
4
5
6 5.5. SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA
7 5.5.1 Introduction
8 Atmospheric PM contains a number of chemical constituents that may be of significance
9 with respect to the human exposure and health effects. These constituents may be either
10 components of the ambient particles or bound to the surface of particles. They may be elements,
11 inorganic species, or organic compounds. A limited number of studies have collected data on
12 concentrations of elements, acidic aerosols, and polycyclic aromatic hydrocarbons (PAHs) in
13 ambient, personal, and microenvironmental PM samples. But, there have not been extensive
14 analyses of the constituents of PM in personal or microenvironmental samples. Data from
15 relevant studies are summarized in this section. The summary does not address bacteria,
16 bioaerosols, viruses, or fungi (e.g., Owen et al., 1992; Ren et al., 1999).
17
18 5.5.2 Monitoring Studies That Address Particulate Matter Constituents
19 A limited number of studies have measured the constituents of PM in personal or
20 microenvironmental samples. Relevant studies published in recent years are summarized in
21 Tables 5-11 and 5-12 for personal exposure measurements of PM and microenvironmental
22 samples, respectively. Studies that measured both personal and microenvironmental samples are
23 included in Table 5-11.
24 The largest database on personal, microenvironmental, and outdoor measurements of PM
25 elemental concentrations is the PTEAM study (Ozkaynak et al., 1996b). The results are
26 highlighted in the table and discussed below. The table shows that a number of studies have
27 measured aerosol acidity, sulfate, ammonia, and nitrate concentrations. Also, a number of
28 studies have measured PAHs, both indoors and outdoors. Other than the PAHs, there is little
29 data on organic constituents of PM.
30
March 2001 5-70 DRAFT-DO NOT QUOTE OR CITE
-------
.
to
o
o
TABLE 5-11. STUDIES THAT HAVE MEASURED PARTICULATE MATTER CONSTITUENTS IN PERSONAL
EXPOSURE SAMPLES
PM Constituent Study Name/Reference
Study Location
Population Size/No, of Samples
Summary of Results
Elements
PTEAM/Ozkaynak et al. (1996b) Riverside, CA
178 adults
As and Pb
Mn
Acid Aerosol
Constituents
NHEXAS/Clayton et al. (1999b) EPA Region 5
Pellizzarietal. (1998, 1999)
Clayton etal. (1999a),
Crump (2000)
Samat et al. (2000)
Braueretal. (1989)
Suhetal. (1992)
Toronto
Baltimore, MD
Boston, MA
Uniontown, PA
167 samples
925 personal samples
20 adults
24 children for 2 days
Outdoor air was the major source for most elements
indoors, providing 70 to 100% of the observed indoor
concentrations for 12 of the 15 elements. Correlation
coefficients for central monitoring site versus outdoor at
the residences were 0.98 for sulfur and 0.5 to 0.9 for other
elements (except copper).
Personal As and Pb levels higher than indoor or outdoor
levels. No community ambient site for comparison.
Mean PM2 5 Mn higher outdoors than indoors. But PM2 5
Mn concentrations higher at two fixed locations than at
participants' homes.
High correlations between personal and ambient sulfate
measurements in summer and winter.
Personal exposures to aerosol strong acidity slightly lower
than concentrations measured at stationary site.
Personal exposures to H+ and SO4"2 lower than outdoor
levels, but higher than indoor microenvironmental levels;
personal NH4+ and NO3" higher than indoor or outdoor
fe
H
I
o
o
o
H
O
O PAHs
H
W
O
H
W
Suh et al. (1 993a,b) State College, PA 47 children
Suh etal. (1994)
Waldman and Liang (1993), Georgia and New Hospital, daycares
Waldman et al. ( 1 990) Jersey
Zmirou et al. (2000) Grenoble, France 38 adults
ICVClb.
Results similar to Uniontown, PA, study.
Results indicate strong neutralization of acidity indoors.
Indoor sulfate levels were 70 to 100% of outdoor levels.
Indoor ammonia levels 5- to 50-times higher than
outdoors. Indoors, acid aerosols were largely neutralized.
Ambient air concentrations close to traffic emissions were
1 . 1 - to 3 . 5 - times higher than personal exposure
concentrations.
-------
.
to
o
o
TABLE 5-12. STUDIES THAT HAVE MEASURED PARTICULATE MATTER CONSTITUENTS IN
MICROENVIRONMENTAL SAMPLES
PM Constituent Study Name/Reference
Study Location
Population Size/No, of Samples
Summary of Results
-n
H
1
O
2|
0
H PAHs and
O phthalates
C
O
— ]
W
O
o
H
W
Chuang et al. (1999) Durham, NC 24 homes
Dubowsky et al. (1999) Boston, MA 3 buildings
Sheldon et al. ( 1 993b,c) Placerville and 280 homes
Ro Seville, CA
PTEAM/Ozkaynak et al. Riverside, CA 120 homes
(1996b),
Sheldon etal. (1993a)
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.
-------
1 5.5.3 Key Findings
2 5.5.3.1 Correlations of Personal and Indoor Concentrations with Ambient Concentrations
3 of Particulate Matter Constituents
4 The elemental composition of PM in personal samples was measured in the PTEAM study,
5 the first probability-based study of personal exposure to particles. A number of important
6 observations, made from the PTEAM data collected in Riverside, CA, are summarized by
7 Ozkaynak et al. (1996b). Population-weighted daytime personal exposures averaged
8 150 ± 9 //g/m3, compared to concurrent indoor and outdoor concentrations of 95 ± 6 //g/m3. The
9 personal exposure measurements suggested that there was a "personal cloud" of particles
10 associated with personal activities. Daytime personal exposures to 14 of the 15 elements
11 measured in the samples were considerably greater than concurrent indoor or outdoor
12 concentrations, with sulfur being the only exception.
13 The PTEAM data also showed good agreement between the concentrations of the elements
14 measured outdoors at the backyard of the residences with the concentrations measured at the
15 central site in the community. The agreement was excellent for sulfur. Although the particle and
16 element mass concentrations were higher in personal samples than for indoor or outdoor samples,
17 a nonlinear mass-balance method showed that the penetration factor was nearly 1 for all particles
18 and elements.
19 Similarly to the PTEAM results, recent measurements of element concentrations in
20 NHEXAS showed elevated concentrations of As and Pb in personal samples relative to indoor
21 and outdoor samples (Clayton et al., 1999b). The elevated concentrations of As and Pb were
22 consistent with elevated levels of PM in personal samples (median particle exposure of
23 101 //g/m3), compared to indoor concentrations (34.4 //g/m3). There was a strong association
24 between personal and indoor concentrations and indoor and outdoor concentrations for both As
25 and Pb. However, there were no central site ambient measurements for comparison to the
26 outdoor or indoor measurements at the residences.
27 Manganese (Mn) concentrations were measured in PM2 5 samples collected in Toronto
28 (Crump, 2000). The mean PM2 5 Mn concentrations were higher outdoors than indoors. But the
29 outdoor concentrations measured at the participant's homes were lower than those measured at
30 two fixed locations. Crump (2000) suggested that the difference in the concentrations may have
March 2001 5-73 DRAFT-DO NOT QUOTE OR CITE
-------
1 been because the fixed locations were likely closer to high-traffic areas than were the
2 participant's homes.
3 Studies of acidic aerosols and gases typically measure strong acidity (H+), SO42", NH4+, and
4 NO3". The relationship between the concentrations of these ions and the relationship between
5 indoor and outdoor concentrations have been addressed in a number of studies during which
6 personal samples, microenvironmental, and outdoor samples have been collected, as shown in
7 Tables 5-11 and 5-12. Key findings from these studies include those shown below.
8 • Acid aerosol concentrations measured at the residences in the Uniontown, PA, study were
9 significantly different from those measured at a fixed ambient site located 16 km from the
10 community. But, Leaderer et al. (1999) reported that the regional ambient air monitoring
11 site in Vinton, VA, provided a reasonable estimate of indoor and outdoor sulfate
12 measurements during the summer at homes without tobacco combustion.
13 • Approximately 75% of the fine aerosol indoors during the summer was associated with
14 outdoor sources based on I/O sulfate ratios measured in the Leaderer et al. (1999) study.
15 • Personal exposures to strong acidity (H+) were lower than corresponding outdoor levels
16 measured in studies by Brauer et al. (1989, 1990) and Suh et al. (1992). But the personal
17 exposure levels measured by Suh et al. (1992) were higher than the indoor
18 microenvironmental levels.
19 • Personal exposures to NH4+, and NO3" were reported by Suh et al. (1992) to be lower than
20 either indoor or outdoor levels.
21 • Personal exposures to SO42" were also lower than corresponding outdoor levels, but
22 higher than the indoor microenvironmental levels (Suh et al., 1992; 1993a,b), as shown in
23 Table 5-13.
24 The fact that the personal and indoor H+ concentrations were substantially lower than
25 outdoor concentrations suggests that a large fraction of aerosol strong acidity is neutralized by
26 ammonia. Ammonia is emitted in relatively high concentrations in exhaled breath and sweat.
27 The difference between indoor and outdoor H+ concentrations in the Suh et al. (1992, 1993a,b)
28 studies was also much higher than the difference for indoor and outdoor SO42", indicative of
29 neutralization of the H+. Results of the Suh et al. (1992, 1993a,b) studies also showed substantial
30 interpersonal variability of H+ concentrations that could not be explained by variation in outdoor
31 concentrations.
March 2001 5-74 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-13. SUMMARY STATISTICS FOR PERSONAL, INDOOR, AND
OUTDOOR CONCENTRATIONS OF SELECTED AEROSOL COMPONENTS IN
TWO PENNSYLVANIA COMMUNITIES
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
Sample Site
(In/Out)a
53/71
254/71
56/75
259/75
214/76
314/155
28/74
230/74
163/75
91/46
91/44
91/46
Indoor (12 h)
GM ± GSDb
2.1 ±2.7
3.2 ±2.3
61. 8 ±2.5
96.7 ±2.5
69.1 ±2.6
154.7 ±2.8
4.2 ±4.3
11.2±3.1
9.1 ±3.5
87.8 ±2.1
157.2 ±2.8
13.7 ±2.5
Concentration
(nmol m"3)
Outdoor (24 h)
GM ± GSDb
1.4 ±2.1
1.4 ±2.1
109.4 ±2.4
109.4 ±2.4
91.0 ±2.5
104.4 ±2.3
82.5 ±2.6
82.5 ±2.6
72.4 ±2.9
124.9 ±1.9
139.4 ±2.1
76.6 ±2.7
Personal (12 h)
GM ± GSDb
—
71.5 ±2.4
—
18.4 ±3.0
110.3 ±1.8
167.0 ± 2.0
42.8 ±2.2
aIn/Out = Indoor sample site/outdoor sample site.
bGM ± GSD = Geometric mean ± geometric standard deviation.
°A/C Homes = Homes that had air-conditioning (A/C); this does not imply that it was on during the entire
sampling period.
Non-A/C = Homes without air conditioning.
dThe sample size (n) for the personal monitoring = 209.
6n = 174 for personal monitoring.
Source: Suhetal. (1992, 1993a,b).
1 Similar results for ammonia were reported by Waldman and Liang (1993). They reported
2 that levels of ammonia in institutional settings that they monitored were 10- to 50- times higher
3 than outdoors, and that acid aerosols were largely neutralized. Leaderer et al. (1999) reported
4 that ammonia concentrations during both winter and summer in residences were an order of
5 magnitude higher indoors than outdoors, consistent with results of other studies and the presence
6 of sources of ammonia indoors.
March 2001
5-75
DRAFT-DO NOT QUOTE OR CITE
-------
1 Sulfate aerosols appear to penetrate indoors effectively. Waldman et al. (1990) reported
2 I/O ratios of 0.7 to 0.9 in two nursing care facilities and a day-care center. Sulfate I/O ratios were
3 measured for three particle size fractions in 12 residences in Birmingham, England, by Jones
4 et al. (2000). The sulfate I/O ratios were 0.7 to 0.9 for PM < 1.1 //m, 0.6 to 0.8 for PM 1.1 to
5 2.1 //m, and 0.7 to 0.8 for PM 2.1 to 10 ym. Suh et al. (1993b) reported that personal and
6 outdoor sulfate concentrations were highly correlated, as depicted in Figure 5-10.
600 r
100 200 300 400 500
Outdoor Sulfate (nmoles/m3)
600
Figure 5-10. Personal versus outdoor SO4 in State College, PA. Open circles represent
children living in air conditioned homes; the solid line is the 1:1 line.
Source: Suhetal. (1993b).
March 2001
5-76
DRAFT-DO NOT QUOTE OR CITE
-------
1 Indoor/outdoor relationships were measured for a number of PM25 components and related
2 species in Lindon, UT, during January and February of 1997 by Patterson and Eatough (2000).
3 Outdoor samples were collected at the Utah State Air Quality monitoring site. Indoor samples
4 were collected in the adjacent Lindon Elementary School. The infiltration factors, Cai/Cao, given
5 by the slope of the regression lines (Table 5-14), were low (0.27 for sulfate and 0.12 for PM25),
6 possibly because of removal of particles in the air heating and ventilation system. The authors
7 concluded that the data indicate that indoor PM2 5 mass may not always be a good indicator of
8 exposure to ambient combustion material caused by the influence of indoor sources of particles.
9 However, ambient sulfate, SO2, nitrate, soot, and total particulate number displayed strong
10 correlations with indoor exposure. Ambient PM2 5 mass was not a good indicator of indoor PM2 5
11 mass exposure.
12
13
TABLE 5-14. STATISTICAL CORRELATION OF OUTDOOR (x) VERSUS INDOOR
(y) CONCENTRATION FOR MEASURED SPECIES
(Units are nmol m3, except for soot and metals, which are
and absorption units m"3, respectively.)"
Species
SO2 All Samples
SO2 Day Samples
SO2 Night Samples
Sulfate All Samples
Sulfate Day Samples
Sulfate Night Samples
Nitrate All Samples
Nitrate Day Samples
Nitrate Night Samples
Soot Day Samples
Soot Night Samples
Total Acidity All Samples
Metals All Samples
Slope
0.0272 ± 0.0023
0.0233 ± 0.0037
0.0297 ± 0.0029
0.267 ± 0.024
0.261 ±0.034
0.282 ±0.035
0.0639 ±0.0096
0.097 ±0.0096
0.047 ±0.011
0.43 ± 0.25
0.33 ±0.13
0.04 ±0.73
0.10±0.30
Intercept
0.34±0.13
0.75 ±0.26
0.099 ±0.075
-0.14 ±0.48
0.40 ±0.66
-0.84 ±0.68
0.9± 1.5
-0.4 ± 1.4
1.5 ± 1.8
3.5 ± 1.7
0.00 ±0.55
0.42 ±0.23
0.0014 ±0.0042
9
r
0.73
0.62
0.82
0.70
0.71
0.70
0.54
0.88
0.44
0.43
0.69
0.00
0.01
Average
Outdoors
38
56
20
16
16
16
134
126
139
6
4
0.2
0.0042
aLindon Elementary School, Lindon, UT, January and February 1997.
Source: Patterson and Eatough (2000).
March 2001 5-77 DRAFT-DO NOT QUOTE OR CITE
-------
1 Oglesby et al. (2000) conducted a study to evaluate the validity of fixed-site fine particle
2 concentration measurements as exposure surrogates for air pollution epidemiology. Using 48-h
3 EXPOLIS data from Basel, Switzerland, they investigated the personal exposure/outdoor
4 concentration relationships for four indicator groups: (1) PM25 mass, (2) sulfur and potassium
5 for regional air pollution, (3) lead and bromine for traffic-related particles, and (4) calcium for
6 crustal particles. The authors reported that personal exposures to PM2 5 mass were not correlated
7 to corresponding home outdoor levels (n = 44, r = 0.07). In the study group reporting neither
8 relevant indoor sources nor relevant activities, personal exposures and home outdoor levels of
9 sulfur were highly correlated (n = 40, r = 0.85). Oglesby et al. (2000) concluded that
10
11 "for regional air pollution, fixed-site fine particle levels are valid exposure surrogates. For source-
12 specific exposures, however, fixed-site data are probably not the optimal measure. Still, in air pollution
13 epidemiology, ambient PM2 5 levels may be more appropriate exposure estimates than total personal
14 PM25 exposure, since the latter reflects a mixture of indoor and outdoor sources."
15
16 PAHs have been measured in studies by EPA and the California Air Resources Board.
17 PAH results from a probability sample of 125 homes in Riverside are discussed in reports by
18 Sheldon et al. (1992a,b) and Ozkaynak et al. (1996b). Data for two sequential 12-h samples were
19 reported for PAHs by ring size (3 to 7) and for individual phthalates. The results are summarized
20 below.
21 • The particulate-phase 5- to 7-ring species had lower relative concentrations than the more
22 volatile 3- to 4-ring species.
23 • The 12-h indoor/outdoor ratios for the 5- to 7-ring species ranged from 1.1 to 1.4 during
24 the day and from 0.64 to 0.85 during the night (Sheldon et al., 1993a).
25 • An indoor air model used to calculate indoor "source strengths" for the PAHs showed
26 that smoking had the strongest effect on indoor concentrations.
27 Results from a larger PAH probability study in 280 homes in Placerville and Roseville
28 (Sheldon et al., 1993b,c) were similar to the 125-home study. The higher-ring, particle-bound
29 PAH's had lower indoor and outdoor concentrations than the lower-ring species. For most
30 PAHs, the I/O ratio was greater than 1 for smoking and smoking/fireplace homes and less than
31 1 for fireplace-only, wood stove, wood stove/gas heat, gas heat, and "no source" homes.
March 2001 5-78 DRAFT-DO NOT QUOTE OR CITE
-------
1 A study of PAHs in indoor and outdoor air was conducted in 14 inner-city and 10 rural
2 low-income homes near Durham, NC, in two seasons (winter and summer) in 1995 (Chuang
3 etal., 1999). Fine-particle-bound PAH concentrations measured with a real-time monitor were
4 usually higher indoors than outdoors (2.47 ± 1.90 versus 0.53 ± 0.58 //g/m3). Higher indoor
5 levels were seen in smoker's homes compared with nonsmoker's homes, and higher outdoor and
6 indoor PAH levels were seen in urban areas compared with rural areas.
7 In a study reported by Dubowsky et al. (1999), the weekday indoor PAH concentrations
8 attributable to traffic (indoor source contributions were removed) were 39 ± 25 ng/m3 in a
9 dormitory that had a high air exchange rate because of open windows and doors, 26 ± 25 ng/m3
10 in an apartment, and 9 ± 6 ng/m3 in a suburban home. The study showed that both
11 outdoor—especially motor vehicular traffic—and indoor sources contributed to indoor PAH
12 concentrations.
13
14 5.5.4 Factors Affecting Correlations Between Ambient Measurements and
15 Personal or Microenvironmental Measurements of Particulate Matter
16 Constituents
17 The primary factors affecting correlations between personal exposure and ambient air PM
18 measurements have been discussed in Section 4.3.2. These include air-exchange rates, particle
19 penetration factors, decay rates and removal mechanisms, indoor air chemistry, and indoor
20 sources. The importance of these factors varies for different PM constituents. For acid aerosols,
21 indoor air chemistry is particularly important as indicated by the discussion of the neutralization
22 of the acidity by ammonia, which is present at higher concentrations indoors because of the
23 presence of indoor sources. For SVOCs, including PAHs and phthalates, the presence of indoor
24 sources will impact substantially the correlation between indoor and ambient concentrations
25 (Ozkaynak et al., 1996b). Penetration factors for PM will impact correlations between indoors
26 and outdoors for most elements, except Pb, which may have significant indoor sources in older
27 homes. Indoor air chemistry, decay rates, and removal mechanisms may affect soot and organic
28 carbon. These factors must be fully evaluated when attempting to correlate ambient, personal,
29 and indoor PM concentrations.
30
31
March 2001 5-79 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.5.5 Limitations of Available Data
2 The previous discussion demonstrates that there is very limited data available that can be
3 used to compare personal, microenvironmental, and ambient air concentrations of PM
4 constituents. Because of resource limitations, PM constituents have not been measured in many
5 studies of PM exposure. Although there is some data on acid aerosols, the comparisons between
6 the personal and indoor data generally have been with outdoor measurements at the participant's
7 residences, not with community ambient air measurement sites. The relationship between
8 personal exposure and indoor levels of acid aerosols is not clear because of the limited database.
9 The exception is sulfate, for which there appears to be a strong correlation between indoor and
10 ambient concentrations.
11 With the exception of PAHs, there are practically no data available to relate personal or
12 indoor concentrations with outdoor or ambient site concentrations of SVOCs, which may be
13 generated from a variety of combustion and industrial sources. The relationship between
14 exposure and ambient concentrations of particles from specific sources, such as diesel engines,
15 has not been determined.
16 Although there is an increasing amount of research being performed to measure PM
17 constituents in different PM size fractions, the current data are inadequate to adequately assess
18 the relationship between indoor and ambient concentrations of most PM constituents.
19
20
21 5.6. IMPLICATIONS OF USING AMBIENT PARTICULATE MATTER
22 CONCENTRATIONS IN EPIDEMIOLOGIC STUDIES OF
23 PARTICULATE MATTER HEALTH EFFECTS
24 In this section, the exposure issues that relate to the interpretation of the findings from
25 epidemiologic studies of PM health effects are examined. This section examines the errors that
26 may be associated with using ambient PM concentrations in epidemiologic analyses of PM health
27 effects. First, implications of associations found between personal exposure and ambient PM
28 concentrations are reviewed. This is discussed separately in the context of either community
29 time-series studies or long-term, cross-sectional studies of chronic effects. Next, the role of
30 compositional and spatial differences in PM concentrations are discussed and how these may
31 influence the interpretation of findings from PM epidemiology. Finally, using statistical
March 2001 5-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 methods, an evaluation of the influence of exposure measurement errors on PM epidemiology
2 studies is presented.
3
4 5.6.1 Potential Sources of Error Resulting from Using Ambient Particulate
5 Matter Concentrations in Epidemiologic Analyses
6 Measurement studies of personal exposures to PM are still few and limited in spatial,
7 temporal, and demographic coverage. Consequently, with the exception of a few longitudinal
8 panel studies, most epidemiologic studies of PM health effects rely on ambient community
9 monitoring data giving 24-h average PM concentration measurements. Moreover, because of
10 limited sampling for PM2 5, many of these epidemiologic studies had to use available PM10 or in
11 some instances had to rely on historic data on other PM measures or indicators, such as TSP,
12 SO4=, TP15, RSP, COH, KM, etc. A critical question often raised in the interpretation of results
13 from acute or chronic epidemiologic community-based studies of PM is whether the use of
14 ambient stationary site PM concentration data influences or biases the findings from these
15 studies. Because the health outcomes are measured on individuals, the epidemiologists might
16 prefer to use personal exposure measurements (total, ambient, or nonambient) instead of
17 surrogates, such as ambient PM concentration measurements collected at one or more ambient
18 monitoring sites in the community. Use of ambient concentrations could lead to
19 misclassification of individual exposures and to errors in the epidemiologic analysis of pollution
20 and health data depending on the pollutant and on the mobility and lifestyles of the population
21 studied. Ambient monitoring stations can be some distance away from the individuals and can
22 represent only a fraction of all likely outdoor microenvironments that individuals come in contact
23 with during the course of their daily lives. Furthermore, most individuals are quite mobile and
24 move through multiple microenvironments (e.g., home, school, office, commuting, shopping,
25 etc.) and engage in diverse personal activities at home (e.g, cooking, gardening, cleaning,
26 smoking). Some of these microenvironments and activities may have different sources of PM
27 and result in distinctly different concentrations of PM than that monitored by the fixed-site
28 ambient monitors. Consequently, exposures of some individuals will be classified incorrectly if
29 only ambient monitoring data are used to estimate individual level exposures to PM. Thus, bias
30 or loss of precision in the epidemiologic analysis may result from improper assessment of
31 exposures using data routinely collected by the neighborhood monitoring stations.
March 2001 5-81 DRAFT-DO NOT QUOTE OR CITE
-------
1 Because individuals are exposed to particles in a multitude of indoor and outdoor
2 microenvironments during the course of a day, concern over error introduced in the estimation of
3 PM risk coefficients using ambient, as opposed to personal, PM measurements has received
4 considerable attention recently from exposure analysts, epidemiologists, and biostatisticians.
5 Some exposure analysts contend that, for community time-series epidemiology to yield
6 information on the statistical association of a pollutant with a health response, there must be an
7 association between personal exposure to a pollutant and the ambient concentration of that
8 pollutant because people tend to spend around 90% time indoors and are exposed to both indoor
9 and outdoor-generated PM (cf. Wallace, 2000b; Brown and Paxton, 1998; Ebelt et al., 2000).
10 Consequently, numerous findings reported in the epidemiologic literature on significant
11 associations between ambient PM concentrations and various morbidity and mortality health
12 indices, in spite of the low correlations between ambient PM and concentrations and measures of
13 personal exposure, has been described by some exposure analysts as an exposure paradox
14 (Lachenmyer and Hidy, 2000, Wilson et al., 2000).
15 To resolve the so-called exposure paradox several types of analyses need to be considered.
16 The first type of analysis has to examine the correlations between ambient PM concentrations
17 and personal exposures that are relevant to most of the existing PM epidemiology studies using
18 either pooled, daily-average, or longitudinal exposure data. The second approach has to study the
19 degree of correlations between the two key components of personal PM exposures (i.e.,
20 exposures caused by ambient-generated PM and exposures caused by nonambient PM) with
21 ambient or outdoor PM concentrations, for each of the three types of exposure study design.
22 In addition, several factors influencing either the exposure or health response characterization of
23 the subjects have to be addressed. These include such factors as
24 • spatial variability of PM components,
25 • health or sensitivity status of subjects,
26 • variations of PM with other co-pollutants,
27 • formal evaluation of exposure errors in the analysis of health data, and
28 • how the results may depend on the variations in the design of the epidemiologic study.
29 To facilitate the discussion of these topics, a brief review of concepts pertinent to exposure
30 analysis issues in epidemiology is presented.
31
March 2001 5-82 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.6.2 Associations Between Personal Exposures and Ambient Particulate
2 Matter Concentrations
3 As defined earlier in Sections 5.3 and 5.4, personal exposures to PM result from an
4 individual's exposures to PM in many different types of microenvironments (e.g., outdoors near
5 home, outdoors away from home, indoors at home, indoors at office or school, commuting,
6 restaurants, malls, other public places, etc.). Total personal exposures (Et) that occur in these
7 indoor and outdoor microenvironments can be classified as those resulting from PM of outdoor
8 origin (Eag ) and those primarily generated by indoor sources and personal activities (Enonag =
9 Eig+Epact). The associations between personal exposures and ambient PM concentrations that
10 have been reported from various personal exposure monitoring studies under three broad
11 categories of study design: (1) longitudinal, (2) daily-average, or (3) pooled exposure studies are
12 summarized below.
13 In the previous Sections 5.4.3.1.2 and 5.4.3.1.3, some of the recent studies conducted
14 primarily in the United States, involving children, elderly, and subjects with COPD were
15 reviewed, and they indicated that both intra- and interindividual variability in the relationships
16 between personal exposures and ambient PM concentrations were observed. A variety of
17 different physical, chemical, and personal or behavioral factors were identified by the original
18 investigators that seem to influence the magnitude and the strength of the associations reported.
19 Clearly, for cohort studies in which individual daily health response are obtained,
20 individual longitudinal PM personal exposure data (including ambient-generated and nonambient
21 components) provide the appropriate indicators. In this case, health responses of each individual
22 can be associated with the total personal exposure, the ambient-generated exposure, or the
23 nonambient exposure of each individual. Also, the relationships of personal exposure indicators
24 with ambient concentration can be investigated. In the case of community time-series
25 epidemiology, however, it is not feasible to obtain experimental measurements of personal
26 exposure for the millions of people over time periods of years that are needed to investigate the
27 relationship between air pollution and infrequent health responses such as deaths or even hospital
28 admissions. The epidemiologist must work with the aggregate number of health responses
29 occurring each day and a measure of the ambient concentration that is presumed to be
30 representative of the entire community. The relationship of PM exposures of the potentially
31 susceptible groups to monitored ambient PM concentrations depends on their activity pattern and
March 2001 5-83 DRAFT-DO NOT QUOTE OR CITE
-------
1 level, residential building and HVAC factors (which influence the infiltration factor), status of
2 exposure to ETS, amount of cooking or cleaning indoors, and seasonal factors, among others.
3 Average personal exposures of these special subgroups to ambient-generated PM are correlated
4 well with ambient PM concentrations regardless of individual variation in the absence of major
5 microenvironmental sources.
6 There seem to be clear differences in the relationships of ambient (Eag) and nonambient
7 (Enonag) exposure with ambient concentration (Ca). Various researchers have shown that Enonag is
8 independent of Ca, but that Eag is a function of Ca. Wilson et al. (2000) explains the difference
9 based on different temporal patterns that effect PM concentrations. "Concentrations of ambient
10 PM are driven by meteorology and by changes in the emission rates and locations of emission
11 sources, while concentrations of nonambient PM are driven by the daily activities of people."
12 Ott et al. (2000) also discuss the reasons for assuming that Enonag is independent of Eag and
13 Ca. They show that the nonambient component of total personal exposure is uncorrelated with
14 the outdoor concentration data. Ott et al. (2000) show the Enonag is similar for three population-
15 based exposure studies, including two large probability-based studies, the PTEAM study
16 conducted in Riverside (Clayton et al., 1993; Thomas et al., 1993; Ozkaynak et al., 1996a,b) and
17 a study in Toronto (Pelizzarri et al., 1999; Clayton et al., 1999a), as well as a nonprobability-
18 based study, conducted in Phillipsburg (Lioy et al., 1990). Based on these three studies, they
19 conclude that Eno and the distribution of(Enonag)it can be treated as constant from city to city.
20 Dominici et al. (2000) examined a larger database consisting of five different PM exposure
21 studies and concluded that Eno can be treated as relatively constant from city to city.
22 If (Enonag)t were constant, this would imply that it would have a zero correlation with (Ca)t.
23 However, this hypothesis of constant (Enonag)it has not been established fully because only a few
24 studies have obtained the data needed to estimate (Enonag)it. Although Enonag is independent of
25 Ca, it may not be independent of a. Sarnat et al. (2000) show that Eno goes up as the
26 ventilation rate (and a) goes down. Lachenmeyer and Hidy (2000) also show, by comparing
27 winter and summer regression equations, that as the slope (a) goes down, the intercept (Enonag)
28 goes up.
29 Mage et al. (1999) assume that the PM10 concentration component from indoor sources,
30 such as smoking, cooking, cleaning, burning candles, and so on, is not correlated with the
March 2001 5-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 outdoor concentration. They indicate that this lack of correlation is expected, because people are
2 unaware of ambient concentrations and do not necessarily change their smoking or cooking
3 activities as outdoor PM10 concentrations vary, an assumption supported by other empirical
4 analyses of personal exposure data. For the PTEAM data set, Mage et al. (1999) have shown that
5 E; and Ca have r near zero (R2 = 0.005). Wilson et al. (2000) have shown the Cai and Q also
have r near zero (R = 0.03). Figure 5-11 shows the relationship of estimated (Enonag)it and En,
ionag
7 with Ca (calculated by EPA from PTEAM and THEES data).
8 Based on these results it is reasonable to assume that ordinarily Enonag has no relationship
9 with Ca. Therefore, in linear nonthreshold models of PM health effects, Enonag is not expected to
10 contribute to the relative risk determined in a regression of health responses on Ca. Furthermore,
11 in time-series analysis of pooled or daily heath data, it is expected that Eag rather than Et will have
12 the stronger association with Ca.
13
14 5.6.3 Role of Compositional Differences in Exposure Characterization for
15 Epidemiology
16 The majority of the available data on PM exposures and relationships with ambient PM
17 have come from a few large-scale studies, such as PTEAM, or longitudinal studies on selected
18 populations, mostly the elderly. Consequently, for most analyses, exposure scientists and
19 statisticians had to rely on PM10 or PM2 5 mass data, instead of elemental or chemical
20 compositional information on individual or microenvironmental samples. In a few cases,
21 researchers have examined the factors influencing indoor outdoor ratios or penetration and
22 deposition coefficients using elemental mass data on personal, indoor, and outdoor PM data (e.g.,
23 Ozkaynak et al. 1996a,b; Yakovleva et al. 1999). These results have been informative in terms
24 of understanding relative infiltration of different classes of particle sizes and sources into
25 residences (e.g., fossil fuel combustion, mobile source emissions, soil-derived, etc.). Clearly, in
26 the accumulation-mode, particles associated with stationary or mobile combustion sources have
27 greater potential for penetration into homes and other microenvironments than do crustal
28 material. The chemical composition of even these broad categories of source classes may have
29 distinct composition and relative toxicity. Moreover, when particles and reactive gases are
30 present indoors in the presence of other pollutants or household chemicals, they may react to
31 form additional or different compounds and particles with yet unknown physical, chemical, and
March 2001 5-85 DRAFT-DO NOT QUOTE OR CITE
-------
Cl)
200
E 150-
C/5 O)
O 13-
Q. _
IN "= 10°
_ 9^
co !Q
m /—
Q- o
-50
(a)
* *
A *
* * *
**
**»* ^ **** *
» * *
+*
50
100
150
200
250
60-
o _-
£ c
0) O>
Q- J2
co
50-
40-
30-
O
Q.
X
LLJ 20-
(b)
50 100 150
Ambient Concentration, pg/m3
200
Figure 5-11. Plots of nonambient exposure to PM10, (a) daytime individual values from
PTEAM data and (b) daily-average values from THEES data.
Source: Data taken from (a) Clayton et al. (1993) and (b) Lioy et al. (1990).
1 toxic composition (Isukapalli et al. 2000). Thus, if indoor-generated and outdoor-generated PM
2 were responsible for different types of health effects, or had significantly different toxicities on a
3 per unit mass basis, it would be then be important that Eag and Enonag should be separated and
4 treated as different species, much like the current separation of PM10 into PM25 and PM10.2 5.
5 These complexities in personal exposure profiles may introduce nonlinearities and other
March 2001
5-86
DRAFT-DO NOT QUOTE OR CITE
-------
1 statistical challenges in the selection and fitting of concentration-response models.
2 Unfortunately, PM health effects models have not yet been able to meaningfully consider such
3 complexities. The relationships of toxicity to the chemical and physical properties of PM are
4 discussed in Chapter 7.
5 It is important also to note that individuals spend time in places other than their homes and
6 outdoors. Many of the interpretations reported in the published literature on factors influencing
7 personal PM10 exposures, as well as in this chapter, come from the PTEAM study. The PTEAM
8 study was conducted 10 years ago in one geographic location in California, during one season,
9 and most residences had very high and relatively uniform air-exchange rates. Nonhome indoor
10 microenvironments were not monitored directly during the PTEAM study. Commuting
11 exposures from traffic or exposures in a variety of different public places or office buildings
12 could not be assessed directly. Nonresidential buildings may have lower or higher ambient
13 infiltration rates depending on the use and type of the mechanical ventilation systems employed.
14 Because the source and chemical composition of particulate matter effecting personal exposures
15 in different microenvironments vary by season, day-of-the-week, and time of day, it is likely that
16 some degree of misclassification of exposures to PM toxic agents of concern will be introduced
17 when health effects models use only daily-average mass measures such as PM10 or PM2 5.
18 Because of the paucity of currently available data on many of these factors, it is impossible to
19 ascertain at this point the magnitude and severity of these more complex exposure
20 missclassification problems in the interpretation of results from PM epidemiology.
21
22 5.6.4 Role of Spatial Variability in Exposure Characterization for
23 Epidemiology
24 Chapter 3 (Section 3.2.3) and Chapter 5 (Section 5.3) present information on the spatial
25 variability of PM mass and chemical components at fixed-site ambient monitors; for purposes of
26 this chapter, this spatial variability is called an "ambient gradient." Any gradient that may exist
27 between a fixed-site monitor and the outdoor ^e near where people live, work, and play,
28 obviously affects the concentration profile actually experienced by people as they go about their
29 daily lives.
30 However, the evidence so far indicates that PM concentrations, especially fine PM (mass
31 and sulfate), generally are distributed uniformly in most metropolitan areas. This reduces the
March 2001 5-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 potential for exposure misclassification because of outdoor spatial gradients when a limited
2 number of ambient PM monitors are used to represent population average ambient exposures in
3 time-series or cross-sectional epidemiologic studies of PM. This topic is further discussed below
4 in Section 5.6.5. However, as discussed earlier, the same assumption is not necessarily true for
5 different components of PM, because source-specific and other spatially nonuniform pollutant
6 emissions could alter the spatial profile of individual PM components in a community.
7 For example, particulate and gaseous pollutants emitted from motor vehicles tend to be higher
8 near roadways and inside cars. Likewise, acidic and organic PM species may be location- and
9 time-dependent. Furthermore, human activities are complex, and if outdoor PM constituent
10 concentration profiles are either spatially or temporally variable, it is likely that exposure
11 misclassification errors could be introduced in the analysis of PM air pollution and health data.
12
13 5.6.5 Analysis of Exposure Measurement Error Issues in Particulate Matter
14 Epidemiology
15 The effects of exposure misclassification on relative risk estimates of disease using
16 classical 2x2 contingency design (i.e., exposed/nonexposed versus diseased/nondiseased) have
17 been studied extensively in the epidemiologic literature. It has been shown that the magnitude of
18 the exposure-disease association (e.g., relative risk) because of either misclassification of
19 exposure or disease alone (i.e., nondifferential misclassification) biases the effect results toward
20 the null, and differential misclassification (i.e, different magnitudes of disease misclassification
21 in exposed and nonexposed populations) can bias the effect measure toward or away from the
22 null value relative to the true measure of association (Shy et al., 1978; Gladen and Rogan, 1979;
23 Copeland et al., 1977; Ozkaynak et al., 1986). However, the extension of these results from
24 contingency analysis design to multivariate (e.g., log-linear regression, Poissson, logit) models
25 typically used in recent PM epidemiology has been more complicated. Recently, researchers
26 have developed a framework for analyzing measurement errors typically encountered in the
27 analysis of time-series mortality and morbidity effects from exposures to ambient PM (cf. Zeger
28 et al., 2000; Dominici et al., 2000; Samet et al., 2000). Some analysis in the context of cross-
29 sectional epidemiology have also been conducted (e.g. Navidi et al., 1999).
30 The appropriateness of using ambient PM concentration as an exposure metric in the
31 context of epidemiologic analysis of health effects associated with exposure to PM recently has
March 2001 5-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 been examined by a number of investigators (cf. Zeger et al., 2000; Dominici et al., 2000; Navidi
2 et al., 1999; Ozkaynak and Spengler, 1996). In the following section, the error analysis model
3 framework developed in Zeger et al. (2000) will be discussed in the context of time-series
4 epidemiology. After which, issues and implications of exposure errors to findings from long-
5 term/chronic or cross-sectional epidemiology will be discussed briefly.
6
7 5.6.5.1 Analysis of Exposure Measurement Errors in Time-Series Studies
8 Zeger et al. (2000) provide a useful framework for analyzing exposure error in community
9 time-series epidemiology. This framework, coupled with results from recent exposure studies,
10 makes it possible to clarify some important questions regarding relationships among the three
1 1 aspects of personal exposure (1) total personal, (2) personal caused by ambient PM, and
12 (3) personal resulting from nonambient PM and ambient concentration. Consider the regression
13 of a health response (i.e., mortality rate on day t, Yt, against the ambient concentration of PM on
14 day t, Ct). In analyzing pollution-level data on mortality and air pollution, log-linear regressions
15 of the form:
(5-11)
16
17 are fit, where Y; is the expected mortality rate; s(t) is an arbitrary but smooth function of time,
18 introduced to control for the confounding of longer trends and seasonality; Ct, is the average of
19 multiple monitor measurements of ambient pollution measurement for day t; and ut are other
20 possible confounders such as temperature and dew point on the same or previous day. Each
21 coefficient, P, in Equation 5-11 gives the expected change in the health response, Y, because of a
22 unit change in its corresponding variable.
23 However, instead of Equation 5-11, Zeger et al. (2000) suggest that the analyst would like
24 to know the corresponding relationship for personal exposure rather than ambient concentration,
25
Yt = exp[s(0 + EtpE + i*,AJ . (5-12)
26 Zeger et al. (2000) do not differentiate among the three aspects of personal or community
27 exposure. To understand the error in P caused by using ambient concentrations instead of
March 2001 5-89 DRAFT-DO NOT QUOTE OR CITE
-------
1 personal exposure in the regression analysis, it is necessary to examine the relationship between
2 Pc, based on a unit change in the ambient concentration, C, and PE, based on a unit change in one
3 of the three aspects of personal exposure, E. In considering the consequences for Pc, as an
4 estimate of PE, of having a measure of ambient pollution Ct, rather than actual personal exposure
5 Eit, it is convenient to express the desired pollution measurement, Eit, as Ct plus three error terms:
6
-Ct). (5-13)
8
9 Here Et represents the daily, community-average personal exposure. The first term,
10 (Eit - Et ), is the error resulting from having only aggregated or community-averaged exposure
1 1 rather than individual-level exposure data. The second term, (Et - Ct ), is the difference
12 between the average personal exposure and the true ambient pollutant level, and the third term,
13 (C*t -Ct), represents the difference between the true and the measured ambient concentration.
14 In the evaluation of these error terms, two types of measurement error often are considered
15 in the context of epidemiology. The classical error model assumes that measurement error,
16 (Ct-Et), depends on ambient measurements [simply referred to as Ct here instead of (Ca),]. The
17 Berkson error model assumes that the measurement error is dependent on the true value or the
18 personal exposure (Et). The regression coefficient (Pc), estimated from the health effects model
19 in the Berkson error case, gives an unbiased estimate of PE. In the classical error case, Pc is a
20 biased estimate of PE, and the degree of bias depends on the correlation between the
21 measurement error and Ct . The measurement error analysis of Zeger et al. (2000) includes three
22 components: (1) an individual's deviation from the risk-weighted average personal exposure;
23 (2) the difference between the average personal exposure and the true ambient level; and (3) the
24 difference between the measured and the true ambient levels, which include the spatial variation
25 of outdoor PM and instrument sampling error. Zeger et al. (2000) conclude that the first and
26 third components are of the Berkson type and, therefore, are likely to have smaller effects on the
27 relative risk estimates for PM. However, the second component can be a source of substantial
28 bias if, for example, there are short-term associations of the contributions of indoor sources with
29 ambient concentrations. However, recent analysis of PTEAM data (Mage et al., 2000) and
30 theoretical considerations (Ott et al., 2000) indicate that it is unlikely that nonambient exposures
March 2001 5-90 DRAFT-DO NOT QUOTE OR CITE
-------
1 will be correlated with the ambient concentration. Therefore, this type of bias is unlikely.
2 However, if the community average exposure to ambient PM is less than the ambient
3 concentration, the risk regression coefficient, Pc, will be biased low. According to Carrol (1995),
4 pc = a PE, where Pc is the percentage increase in risk because of a unit increase in ambient
5 concentration, and PE is the estimated percentage increase in risk because of a unit increase in the
6 community-average personal exposure to ambient PM. Both Zeger et al. (2000) and Dominici
7 et al. (2000) examine the nature of error with this second component. Both of these analyses
8 conclude that the error introduced because of measured differences between the average personal
9 exposure and ambient levels can bias the regression coefficients. In both cases they find the Pc is
10 close to a PE.
11 This framework analysis demonstrates the importance of the daily community-average
12 exposure, Et, in community time-series epidemiology. It is Et, not the random, pooled values of
13 Eit, that need to have a statistically significant correlation with Ct for proper interpretation of
14 community time-series epidemiology studies based on ambient monitoring data, as discussed
15 further in Wilson et al. (2000) and Mage et al. (1999).
16 A critical assumption in the above analysis is that the risk varies linearly with C or E (i.e.,
17 Pc and PE are constant). This assumption does not permit a threshold (a concentration below
18 which there is no effect). It also includes the assumption that the appropriate metric for
19 determination of a health response is the 24-h average PM mass concentration. Zeger et al.
20 (2000) show that the likely consequence of using ambient concentrations instead of the risk-
21 weighted average personal exposure measures is to underestimate the pollution effects.
22 According to Zeger et al. (2000) the largest biases in inferences about the mortality-personal
23 exposure relative risk will occur because of more complex errors between ambient concentration
24 and daily-average personal exposure measures. It is important to note that both the Zeger et al.
25 (2000) and the Domini ci et al. (2000) error analyses used personal PM10 data from the PTEAM
26 study data. However, effects of measurement error estimates may differ by particle size and
27 composition. It is possible that PM25, ultrafine particle measures, or another component of PM,
28 may better reflect personal exposures to PM of outdoor origin. Finally, the seasonal or temporal
29 variations in the measurement errors and correlations between different PM concentration
30 measures and co-pollutants (e.g. SO2, CO, NO2, O3) could influence the error analysis results
31 reported by the investigators cited above.
March 2001 5-91 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.6.5.2 Analysis of Exposure Measurement Errors in Long-Term Epidemiology Studies
2 The Six Cities (Dockery et al., 1993) and ACS (Pope et al., 1995) studies have played an
3 important role in assessing the health effects from long-term exposures to particulate pollution.
4 Even though these studies often have been considered as chronic epidemiologic studies, it is not
5 easy to differentiate the role of historic exposures from those of recent exposures on chronic
6 disease mortality. In the Six Cities study, fine particles and sulfates were measured at the
7 community level, and the final analysis of the database used six city-wide average ambient
8 concentration measurements. This limitation also applies to the ACS study but has less impact
9 because of the larger number of cities considered in that study. In a HEI-sponsored reanalysis of
10 the Six Cities and the ACS data sets, Krewski et al. (2000) attempted to examine some of the
11 exposure misclassification issues either analytically or through sensitivity analysis of the
12 aerometric and health data. The HEI reanalysis project also addressed exposure measurement
13 error issues related to the Six Cities study. For example, the inability to account for exposures
14 prior to the enrollment of the cohort, hampered accurate interpretation of the relative risk
15 estimates in terms of acute versus chronic causes. Although the results seem to suggest past
16 exposures are more strongly associated with mortality than recent exposures, the measurement
17 error for long-term averages could be higher, thus influencing these interpretations. For example,
18 Krewski et al. (2000), using the individual mobility data available for the Six Cities cohort,
19 analyzed the mover and nonmover groups separately. The relative risk of fine particle effects on
20 all-cause mortality was shown to be higher for the nonmover group than for the mover group,
21 suggesting the possibility of higher exposure misclassification biases for the movers. The issue
22 of using selected ambient monitors in the epidemiologic analyses also was investigated by the
23 ACS and Six Cities studies reanalysis team. Krewski et al.(2000) presented the sensitivity of
24 results to choices made in selecting stationary or mobile-source-oriented monitors. For the ACS
25 study, reanalysis of the sulfate data using only those monitors designated as residential or urban,
26 and excluding sites designated as industrial, agricultural, or mobile did not change the risk
27 estimates appreciably. On the other hand, application of spatial analytic methods designed to
28 control confounding at larger geographic scales (i.e., between cities) caused changes in the
29 particle and sulfate risk coefficients. Spatial adjustment may account for differences in pollution
30 mix or PM composition, but many other cohort-dependent risk factors will vary across regions or
31 cities in the United States. Therefore, it is difficult to interpret these findings solely in terms of
March 2001 5-92 DRAFT-DO NOT QUOTE OR CITE
-------
1 spatial differences in pollution composition or relative PM toxicity until further research is
2 concluded.
3 Another study that has examined the influence of measurement errors in air pollution
4 exposure and health effects assessments is the one reported by Navidi et al. (1999). This study
5 developed techniques to incorporate exposure measurement errors encountered in long-term air
6 pollution health effects studies and tested them on the data from the University of Southern
7 California Children's Health Study conducted in 12 communities in California. These
8 investigators developed separate error analysis models for direct (i.e., personal sampling) and
9 indirect (i.e., microenvironmental) personal exposure assessment methods. These models were
10 generic to most air pollutants, but a specific application was performed using a simulated data set
11 for studying ozone health effects on lung function decline in children. Because the assumptions
12 made in their microenvironmental simulation modeling framework were similar to those made in
13 estimating personal PM exposures, it is useful to consider the conclusions from Navidi et. al.
14 (1999). According to Navidi et al. (1999), neither the microenvironmental nor the personal
15 sampler method produces reliable estimates of the exposure-response slope (for O3) when
16 measurement error is uncorrected. Because of nondifferential measurement error, the bias was
17 toward zero under the assumptions made in Navidi et al. (1999) but could be away from zero if
18 the measurement error was correlated with the health response. A simulation analysis indicated
19 that the standard error of the estimate of a health effect increases as the errors in exposure
20 assessment increase (Navidi et al., 1999). According to Navidi et al. (1999), when a fraction of
21 the ambient level in a microenvironment is estimated with a standard error of 30%, the standard
22 error of the estimate is 50% higher than it would be if the true exposures were known. It appears
23 that errors in estimating ambient PM indoor/ambient PM outdoor ratios have much more
24 influence on the accuracy of the microenvironmental approach than do errors in estimating time
25 spent in these microenvironments.
26
27 5.6.5.3 Conclusions from Analysis of Exposure Measurement Errors on Particulate Matter
28 Epidemiology
29 Personal exposures to PM are influenced by a number of factors and sources of PM located
30 in both indoor and outdoor microenvironments. However, PM resulting from ambient sources
31 does penetrate into indoor environments, such as residences, offices, public buildings, etc., in
March 2001 5-93 DRAFT-DO NOT QUOTE OR CITE
-------
1 which individuals spend a large portion of their daily lives. The correlations between total
2 personal exposures and ambient or outdoor PM concentrations can vary depending on the relative
3 contributions of indoor PM sources to total personal exposures. Panel studies of both adult and
4 young subjects have shown that, in fact, individual correlations of personal exposures with
5 ambient PM concentrations could vary person to person, and even day to day, depending on the
6 specific activities of each person. Separation of PM exposures into two components,
7 ambient-generated PM and nonambient PM, would reduce uncertainties in the analysis and
8 interpretation of PM health effects data. Nevertheless, because ambient-generated PM is an
9 integral component of total personal exposures to PM, statistical analysis of cohort-average
10 exposures are strongly correlated with ambient PM concentrations when the size of the
11 underlying population studied is large. Using the PTEAM study data, analysis of exposure
12 measurement errors, in the context of time-series epidemiology, also has shown that errors or
13 uncertainties introduced by using surrogate exposure variables, such as ambient PM
14 concentrations, could lead to biases in the estimation of health risk coefficients. These then
15 would need to be corrected by suitable calibration of the PM health risk coefficients.
16 Correlations between the PM exposure variables and other covariates (e.g., gaseous
17 co-pollutants, weather variables, etc.) also could influence the degree of bias in the estimated PM
18 regression coefficients. However, most time-series regression models employ seasonal or
19 temporal detrending of the variables, thus reducing the magnitude of this cross-correlation
20 problem (Ozkaynak and Spengler 1996).
21 Ordinarily, exposure measurement errors are not expected to influence the interpretation of
22 findings from either the cross-sectional or time-series epidemiologic studies that have used
23 ambient concentration data if they include sufficient adjustments for seasonality and key
24 confounders. Clearly, there is no question that better estimates of exposures to components of
25 PM of health concern are beneficial. Composition of PM may vary in different geographic
26 locations and different exposure microenvironments. Compositional and spatial variations could
27 lead to further errors in using ambient PM measures as surrogates for exposures to PM. Even
28 though the spatial variability of PM (PM2 5 in particular) mass concentrations in urban
29 environments seems to be small, the same conclusions drawn above regarding the influence of
30 measurement errors may not necessarily hold for all of the PM toxic components. Again, the
31 expectation based on statistical modeling considerations is that these exposure measurement
March 2001 5-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 errors or uncertainties will most likely reduce the statistical power of the PM health effects
2 analysis, making it difficult to detect a true underlying association between the correct exposure
3 metric and the health outcome studied. However, until more data on exposures to toxic agents of
4 PM become available, existing studies on PM exposure measurement errors must be relied on;
5 these indicate that use of ambient PM concentrations as a surrogate for exposures is not expected
6 to change the principal conclusions from PM epidemiologic studies, utilizing community average
7 health and pollution data.
8
9
10 5.7 SUMMARY OF KEY FINDINGS AND LIMITATIONS
11 Exposure Definitions and Components
12 • Personal exposure (E) to PM mass or its constituents results when individuals come in contact
13 with particulate pollutant concentrations (C) in locations or microenvironments (fj,e) that they
14 frequent during a specific period of time. Various PM exposure metrics can be defined
15 according to its source (i.e., ambient, nonambient) and the microenvironment where exposure
16 occurs.
17 • Personal exposure to PM results from an individual's exposure to PM in many different types
18 of microenvironments (e.g., outdoors near home, outdoors away from home, indoors at home,
19 indoors at office or school, commuting, restaurants, malls, other public places, etc.). Thus, total
20 daily exposure to PM for a single individual (Et) can be expressed as the sum of various
21 microenvironmental exposures that the person encounters during the course of a day.
22 • In a given (j,e, particles may originate from a wide variety of sources. In an indoor
23 microenvironment, PM may be generated from within as a result of PM generating activities
24 (e.g., cooking, cleaning, smoking, resuspending PM from PM resulting from both indoor and
25 outdoor sources that had settled out), from outside (outdoor PM entering through cracks and
26 openings in the structure), and from the chemical interaction of pollutants from outdoor air with
27 indoor-generated pollutants.
28 • The total daily exposure to PM for a single individual (Et) also can be expressed as the sum of
29 contributions of ambient-generated (Eag) and nonambient-generated (Enonag) PM (i.e.,
30 E = Eag+ Enonag). Enonag, in turn, is composed of PM generated by indoor sources (Eig ) and PM
31 generated by personal activities (Epact) (i.e., Enonag = Eig + Epact). Eag is composed of exposures to
March 2001 5-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 ambient PM concentrations while outdoors, 2., Cahta, and ambient PM that has infiltrated
t
2 indoors, £ CBl.Af,. while indoors (i.e., Eag = £ CaMa + £ Ca,.Af,.).
/ t t
3 • Exposure models are useful tools for examining the importance of sources, microenvironments,
4 and physical and behavioral factors that influence personal exposures to PM. However,
5 development and evaluation of population exposure models for PM and its components has
6 been limited. Improved modeling methodologies and new model input data are needed.
7
8 Factors Affecting Concentrations and Exposures to Particulate Matter
9 • Concentrations of PM indoors are affected by several factors and mechanisms: ambient
10 concentrations outdoors; air exchange rates; particle penetration factors; particle production
11 from indoor sources and indoor air chemistry; and indoor particle decay rates and removal
12 mechanisms caused by physical processes or resulting from mechanical filtration, ventilation or
13 air-conditioning devices.
14 • Average personal exposures to PM mass and its constituents are influenced by
15 microenvironmental PM concentrations and by how much time is spent by each individual in
16 these various indoor and outdoor microenvironments. Nationwide, individuals, on average,
17 spend nearly 90% of their time indoors (at home and in other indoor locations) and about 6% of
18 their time outdoors.
19 • The relative size of personal exposure to ambient-generated PM relative to nonambient-
20 generated PM depends on the ambient concentration, the infiltration rate of outdoor PM into
21 indoor microenvironments, the amount of PM generated indoors (e.g., ETS, cooking and
22 cleaning emissions), and the amount of PM generated by personal activity sources. Infiltration
23 rates primarily depend on air-exchange rate, size-dependent particle penetration across the
24 building membrane, and size-dependent removal rates. All of these factors vary over time and
25 across subjects and building types.
26 • The relationship between PM exposure and health outcome could depend on the concentration,
27 composition, and toxicity of the PM originating from different sources. Application of source
28 apportionment techniques to ambient, indoor, and personal PM composition data have
29 identified the following general source categories of importance: outside soil, resuspended
30 indoor soil, indoor soil, personal activities, sea-salt, motor vehicles, nonferrous metal smelters,
31 and secondary sulfates.
March 2001 5-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 • There have been only a limited number of studies that have measured the physical and
2 chemical constituents of PM in personal or microenvironmental samples. Available data on
3 PM constituents indicate that
4 S personal and indoor sulfate measurements often are correlated highly with outdoor and
5 ambient sulfate concentration measurements;
6 S for acid aerosols, indoor air chemistry is particularly important because of the
7 neutralization of the acidity by ammonia, which is present at higher concentrations
8 indoors because of the presence of indoor sources of ammonia;
9 S for SVOCs, including PAHs and phthalates, the presence of indoor sources will
10 substantially impact the relation between indoor and ambient concentrations;
11 S penetration and decay rates are a functions of size and will cause variations in the
12 attenuation factors as a function of particle size; infiltration rates will be higher for PMl
13 and PM25 than for PM10, PM10_2 5 or ultrafme particles; and
14 S Indoor air chemistry may increase indoor concentrations of organic PM.
15 • Even though there is an increasing amount of research being performed to measure PM
16 constituents in different PM size fractions, with few exceptions (i.e., sulfur or sulfates), the
17 current data are inadequate to adequately assess the relationship between personal, indoor, and
18 ambient concentrations of most PM constituents.
19
20 Correlations Between Personal Exposures, Indoor, Outdoor, and Ambient Measurements
21 • Most of the available personal data on PM measurements and information on the relationships
22 between personal and ambient PM come from a few large-scale studies, such as the PTEAM
23 study, or the longitudinal panel studies, which have been conducted on selected populations,
24 such as the elderly.
25 • Panel and cohort studies that have measured PM exposures and concentrations typically have
26 reported their results in terms of three types of correlations: (1) longitudinal, (2) pooled, and
27 (3) daily-average correlations between personal and ambient or outdoor PM.
28 • The type of correlation analysis performed can have a substantial effect on the resulting
29 correlation coefficient. Low correlations with ambient concentrations could result when people
30 with very different nonambient exposures are pooled, even though temporally, their individual
31 personal exposures may be correlated highly with ambient concentrations.
March 2001 5-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Recent studies conducted by EPA of the elderly subjects living in a retirement facility in
2 Baltimore and a group of elderly living in Fresno produced higher correlation coefficients
3 between personal and ambient PM for daily-average correlations compared to longitudinal
4 correlations. This supports earlier analyses showing the daily-average correlations are higher
5 than pooled correlations.
6 • Longitudinal and pooled correlations between personal exposure and ambient or outdoor PM
7 concentrations reported by various investigators varied considerably among the different
8 studies and in each study between the study subjects. Most studies report longitudinal
9 correlation coefficients that range from close to zero to near one, indicating that individual's
10 activities and residence type may have a significant effect on total personal exposures to PM.
11 • Longitudinal studies that measured sulfate found high correlations between personal and
12 ambient sulfate.
13 • In general, probability-based population studies tend to show low pooled correlations because
14 of the high differences in levels of nonambient PM generating activities from one subject to
15 another. In contrast, the absence of indoor sources for the populations in several of the
16 longitudinal panel studies resulted in high correlations between personal exposure and ambient
17 PM within subjects over time for these populations. But even for these studies, correlations
18 varied by individual depending on their activities and on the microenvironments that they
19 occupied.
20
21 Potential Sources of Error Resulting from Using Ambient Particulate Matter
22 Concentrations in Epidemiologic Analyses
23 • There is, as yet, no clear consensus among exposure analysts as to how well ambiently
24 measured PM concentrations represent a surrogate for personal exposure to total PM or to
25 ambient-generated PM.
26 • Measurement studies of personal exposures to PM are still few and limited in spatial, temporal,
27 and demographic coverage. Consequently, with the exception of a few longitudinal panel
28 studies, most epidemiologic studies on PM health effects have relied on daily-average PM
29 concentration measurements obtained from ambient community monitoring data as a surrogate
30 for the exposure variable.
March 2001 5-98 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Because individuals are exposed to particles in a multitude of indoor and outdoor
2 microenvironments during the course of a day, concerns over error introduced in the estimation
3 of PM risk coefficients using ambient, as opposed to personal PM measurements, have been
4 raised.
5 • Total personal exposures to PM could vary from person to person, and even day to day,
6 depending on the specific activities of each person. Separation of PM exposures into two
7 components, ambient-generated PM and nonambient-generated PM, would reduce potential
8 uncertainties in the analysis and interpretation of PM health effects data.
9 • Available data indicate that PM mass concentrations, especially fine PM, typically are
10 distributed uniformly in most metropolitan areas, thus reducing the potential for exposure
11 misclassification because of spatial variability when a limited number of ambient PM monitors
12 are used to represent population average ambient exposures in community time-series or
13 long-term, cross-sectional epidemiologic studies of PM.
14 • Even though the spatial variability of PM (in particular, PM25) mass concentrations in urban
15 environments seems to be small, the same conclusions drawn above regarding the influence of
16 measurement errors may not necessarily hold for all of the PM components.
17 • There are important differences in the relationship of ambient PM concentrations (Ca) with
18 exposures to ambient PM (Eag), and with exposures to nonambient PM (Enonag). Various
19 researchers have shown that Eag is a function of Ca, and that concentrations of ambient PM are
20 driven by meteorology, by changes in source emission rates, and in locations of emission
21 sources relative to the measurement site. However, Enonag is independent of Ca, because
22 concentrations of nonambient PM are driven by the daily activities of people.
23 • Because personal exposures also include a contribution from ambient concentrations, the
24 correlation between daily-average personal exposure and the daily-average ambient
25 concentration increases as the number of subjects measured daily increases. An application of
26 a Random Component Superposition (RCS) model has shown that the contributions of ambient
27 PM10 and indoor-generated PM10 to community mean exposure can be decoupled in modeling
28 urban population exposure distributions.
29 • If linear nonthreshold models are assumed in time-series analysis of daily-average ambient PM
30 concentrations and community health data, Enonag is not expected to contribute to the relative
31 risk estimates determined by regression of health responses on Ca.
March 2001 5-99 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Using the PTEAM study data, analysis of exposure measurement errors in the context of
2 time-series epidemiology has shown that errors or uncertainties introduced by using surrogate
3 exposure variables, such as ambient PM concentrations, could lead to biases in the estimation
4 of health risk coefficients.
5 • Because sources and chemical composition of particulate matter affecting personal exposures in
6 different microenvironments vary, by season, day-of-the-week, and time of day, it is likely that
7 some degree of misclassification of exposures to PM toxic agents of concern will be introduced
8 when health effects models use only daily-average mass measures such as PM10 or PM2 5.
9 Because of the paucity of currently available data on many of these factors, it is impossible to
10 ascertain at this point the significance of these more complex exposure misclassification
11 problems in the interpretation of results from PM epidemiology.
12 • Exposure measurement errors may depend on particle size and composition. PM2 5 better
13 reflects personal exposure to PM of outdoor origin than PM10. It is possible that various
14 ultrafine particle measures, or other components of PM may be better exposure indicators for
15 epidemiologic studies.
16 • Seasonal or temporal variations in the measurement errors and their correlations between
17 different PM concentration measures and co-pollutants (e.g., SO2, CO, NO2, O3) could
18 influence the error analysis results but not likely the interpretation of current findings.
19 • Ordinarily, PM exposure measurement errors are not expected to influence the interpretation of
20 findings from either the community time-series or long-term epidemiologic studies that have
21 used ambient concentration data if they include sufficient adjustments for seasonality and key
22 personal and geographic confounders.
23 • To reduce exposure misclassification errors in PM epidemiology, conducting new cohort
24 studies of sensitive populations with better real-time techniques for exposure monitoring and
25 further speciation of indoor-generated, ambient, and personal PM mass are essential.
26 • Based on statistical modeling considerations, it is expected that existing PM exposure
27 measurement errors or uncertainties most likely will reduce the statistical power of the PM
28 health effects analysis, thus making it difficult to detect a true underlying association between
29 the correct exposure metric and the health outcome studied.
March 2001 5-100 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Currently available studies on PM exposure measurement errors indicate that use of ambient
2 PM concentrations as a surrogate for personal exposures is not expected to change the key
3 conclusions derived from most of the recent epidemiologic studies on PM health effects.
4
March 2001 5-101 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Abt, E.; Suh, H. H.; Allen, G.; Koutrakis, P. (2000a) Characterization of indoor particle sources: a study conducted
3 in the metropolitan Boston area. Environ. Health Perspect. 108: 35-44.
4 Abt, E.; Suh, H. H.; Catalano, P.; Koutrakis, P. (2000b) The relative contribution of outdoor and indoor particle
5 sources to indoor concentrations. Environ. Sci. Technol. 34: 3579-3587.
6 Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
7 monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
8 Technol. 19:911-918.
9 Al-Raheem, M. Y. A.; El-Desouky, M; Mage, D. T.; Kollander, M.; Wallace, L. A.; Spengler, J. D. (2000)
10 Personal exposures to PM and gaseous air pollutants at the end of the 1991 Kuwait oil fires. Presented at:
11 PM2000: paniculate matter and health—the scientific basis for regulatory decision-making, specialty
12 conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
13 Aim, S.; Jantunen, M. J.; Vartiainen, M. (1999) Urban commuter exposure to particle matter and carbon monoxide
14 inside an automobile. J. Exposure Anal. Environ. Epidemiol. 9: 237-244.
15 Anuszewski, J.; Larson, T. V.; Koenig, J. Q. (1998) Simultaneous indoor and outdoor particle light-scattering
16 measurements at nine homes using a portable nephelometer. J. Exposure Anal. Environ. Epidemiol.
17 8:483-493.
18 Baek, S.-O.; Kim, Y.-S.; Perry, R. (1997) Indoor air quality in homes, offices and restaurants in Korean urban
19 areas—indoor/outdoor relationships. Atmos. Environ. 31: 529-544.
20 Bahadori, T.; Suh, H. H.; Rojas-Bracho, L.; Koutrakis, P. (2001) Personal exposure to paniculate matter of
21 individuals with chronic obstructive pulmonary disease (COPD). J. Expos. Anal. Environ. Epidemiol.:
22 Accepted.
23 Beyea, J.; Hatch, M. (1999) Geographic exposure modeling: a valuable extension of geographic information
24 systems for use in environmental epidemiology. Environ. Health Perspect. Suppl. 107(1): 181-190.
25 Bohadana, A. B.; Massin, N.; Wild, P.; Toamain, J.-P.; Engel, S.; Goutet, P. (2000) Symptoms, airway
26 responsiveness, and exposure to dust in beech and oak wood workers. Occup. Environ. Med. 57: 268-273.
27 Brauer, M.; 't Mannetje, A. (1998) Restaurant smoking restrictions and environmental tobacco smoke exposure.
28 Am. J. Public Health 88: 1834-1836.
29 Brauer, M.; Koutrakis, P.; Spengler, J. D. (1989) Personal exposures to acidic aerosols and gases. Environ. Sci.
30 Technol. 23: 1408-1412.
31 Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1990) Indoor and outdoor concentrations of acidic aerosols
32 and gases. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and climate,
33 volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
34 International Conference on Indoor Air Quality and Climate, Inc.; pp. 447-452.
35 Brauer, M.; Bartlett, K.; Regalado-Pineda, J.; Perez-Padilla, R. (1996) Assessment of paniculate concentrations
36 from domestic biomass combustion in rural Mexico. Environ. Sci. Technol. 30: 104-109.
37 Brauer, M.; Hirtle, R. D.; Hall, A. C.; Yip, T. R. (1999) Monitoring personal fine particle exposure with a particle
38 counter. J. Exposure Anal. Environ. Epidemiol. 9: 228-236.
39 Brown, K. G.; Paxton, M. B. (1998) Predictability of personal exposure to airborne paniculate matter from ambient
40 concentrations in four communities. In: Vostal, J. J., ed. Health effects of paniculate matter in ambient air.
41 Proceedings of an international conference; 1997; Prague, Czech Republic. Pittsburgh, PA: Air & Waste
42 Management Association; pp. 333-350. (A&WMA publication VIP-80).
43 Burke, J. M.; Zufall, M. J.; Ozkaynak, H. A.; Xue, J.; Zidek, J. V. (2000) Predicting population exposures to PM:
44 the importance of microenvironmental concentrations and human activities. Presented at: PM2000:
45 paniculate matter and health~the scientific basis for regulatory decision-making, specialty conference &
46 exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
47 Burke, J.; Zufall, M.; Ozkaynak, H. (2001) A population exposure model for paniculate matter: case study results
48 for PM2.5 in Philadelphia, PA. J. Expos. Anal. Environ. Epidemiol.: submitted.
49 Burton, R. M.; Suh, H. H.; Koutrakis, P. (1996) Spatial variation in paniculate concentrations within metropolitan
50 Philadelphia. Environ. Sci. Technol. 30: 400-407.
51 Buzorius, G.; Ha'meri, K.; Pekkanen, J.; Kulmala, M. (1999) Spatial variation of aerosol number concentration in
52 Helsinki city. Atmos. Environ. 33: 553-565.
53 Byrne, M. A.; Lange, C.; Goddard, A. J. H.; Roed, J. (1992) Indoor aerosol deposition studies using
54 neutron-activatable tracers. J. Aerosol Sci. 23(suppl. 1): S543-S546.
March 2001 5-102 DRAFT-DO NOT QUOTE OR CITE
-------
1 Calder, K. L. (1957) A numerical analysis of the protection afforded by buildings against BW aerosol attack.
2 Fort Derrick, MD: Office of the Deputy Commander for Scientific Activities; BWL tech. study no. 2.
3 Carrer, P.; Alcini, D.; Cavallo, D.; Bollini, D.; Ghalandar, R.; Lovato, L.; Ploia, E.; Vercelli, F.; Visigalli, F.;
4 Maroni, M. (1998) Personal exposure to PM10 of subjects living in Milan. In: Vostal, J. J., ed. Health effects
5 of paniculate matter in ambient air. Proceedings of an international conference; 1997; Prague, Czech
6 Republic. Pittsburgh, PA: Air & Waste Management Association; pp. 359-365. (A&WMA publication
7 VIP-80).
8 Carroll, R. J.; Ruppert, D.; Stefanski, L. A. (1995) Measurement error in nonlinear models. London, United
9 Kingdom: Chapman & Hall. (Cox, D. R.; Hinkley, D. V.; Keiding, N.; Reid, N.; Rubin, D. B.; Silverman,
10 B. W., eds. Monographs on statistics and applied probability: v. 63).
11 Cha, Q.; Chen, Y.-Z.; Du, Y.-X. (1996) Influence of indoor air pollution on quality of outdoor air in city of
12 Guangzhou, China. In: Indoor air '96: the 7th international conference on indoor air quality, volume I; July;
13 Nagoya, Japan. Tokyo, Japan: Institute of Public Health; pp. 1131-1135.
14 Chang, L.-T.; Koutrakis, P.; Catalano, P. J.; Suh, H. H. (2000) Hourly personal exposures to fine particles and
15 gaseous pollutants - results from Baltimore, Maryland. J. Air Waste Manage. Assoc. 50: 1223-1235.
16 Chuang, J. C.; Callahan, P. J.; Lyu, C. W.; Wilson, N. K. (1999) Polycyclic aromatic hydrocarbon exposures of
17 children in low-income families. J. Exposure Anal. Environ. Epidemiol. 2: 85-98.
18 Clayton, C. A.; Perritt, R. L.; Pellizzari, E. D.; Thomas, K. W.; Whitmore, R. W.; Wallace, L. A.; Ozkaynak, H.;
19 Spengler, J. D. (1993) Particle total exposure assessment methodology (PTEAM) study: distributions of
20 aerosol and elemental concentrations in personal, indoor, and outdoor air samples in a southern California
21 community. J. Exposure Anal. Environ. Epidemiol. 3: 227-250.
22 Clayton, C. A.; Pellizzari, E. D.; Rodes, C. E.; Mason, R. E.; Piper, L. L. (1999a) Estimating distributions of
23 long-term paniculate matter and manganese exposures for residents of Toronto, Canada. Atmos. Environ.
24 33:2515-2526.
25 Clayton, C. A.; Pellizzari, E. D.; Whitmore, R. W.; Perritt, R. L., Quackenboss, J. J. (1999b) National human
26 exposure assessment survey (NHEXAS): distributions and associations of lead, arsenic, and volatile organic
27 compounds in EPA Region 5. J. Exposure Anal. Environ. Epidemiol. 9: 381-392.
28 Cohen, B. S.; Li, W.; Xiong, J. Q.; Lippmann, M. (2000) Detecting H+ in ultrafine ambient aerosol using iron
29 nano-film detectors and scanning probe microscopy. Appl. Occup. Environ. Hyg. 15: 80-89.
30 Conner, T. L.; Lonneman, W. A.; Seila, R. L. (1995) Transportation-related volatile hydrocarbon source profiles
31 measured in Atlanta. J. Air Waste Manage. Assoc. 45: 383-394.
32 Copeland, K. T.; Checkoway, H.; McMichael, A. J.; Holbrook, R. H. (1977) Bias due to misclassification in the
33 estimation of relative risk. Am. J. Epidemiol. 105: 488-495.
34 Crump, K. S. (2000) Manganese exposures in Toronto during use of the gasoline additive, methylcyclopentadienyl
35 manganese tricarbonyl. J. Exposure Anal. Environ. Epidemiol. 10: 227-239.
36 Daisey, J. M.; Mahanama, K. R. R.; Hodgson, A. T. (1998) Toxic volatile organic compounds in simulated
37 enviromental tobacco smoke: emission factors for exposure assessment. J. Exposure Anal. Environ.
38 Epidemiol. 8:313-334.
39 Dockery, D. W.; Spengler, J. D. (1981) Indoor-outdoor relationships of respirable sulfates and particles.
40 Atmos. Environ. 15: 335-343.
41 Dockery, D. W.; Pope, C. A., Ill; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
42 (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med.
43 329: 1753-1759.
44 Dominici, F.; Zeger, S. L.; Samet, J. (2000) A measurement error model for time-series studies of air pollution and
45 mortality. Biostatistics 1: 157-175.
46 Donham, K. J.; Cumro, D.; Reynolds, S. J.; Merchant, J. A. (2000) Dose-response relationships between
47 occupational aerosol exposures and cross-shift declines of lung function in poultry workers:
48 recommendations for exposure limits. J. Occup. Environ. Med. 42: 260-269.
49 Duan, N. (1982) Models for human exposure to air pollution. Environ. Int. 8: 305-309.
50 Duan, N. (1991) Stochastic microenvironment models for air pollution exposure. J. Exp. Anal. Environ. Epidemiol.
51 1:235-257.
52 Dubowsky, S. D.; Wallace, L. A.; Buckley, T. J. (1999) The contribution of traffic to indoor concentrations of
53 polycyclic aromatic hydrocarbons. J. Exposure Anal. Environ. Epidemiol. 9: 312-321.
March 2001 5-103 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ebelt, S. T.; Petkau, A. J.; Vedal, S.; Fisher, T. V.; Brauer, M. (2000) Exposure of chronic obstructive pulmonary
2 disease patients to paniculate matter: relationships between personal and ambient air concentrations. J. Air
3 Waste Manage. Assoc. 50: 1081-1094.
4 Esmen, N. A.; Hall, T. A. (2000) Theoretical investigation of the interrelationship between stationary and personal
5 sampling in exposure estimation. Appl. Occup. Environ. Hyg. 15: 114-119.
6 Evans, G. F.; Highsmith, R. V.; Sheldon, L. S.; Suggs, J. C.; Williams, R. W.; Zweidinger, R. B.; Creason, J. P.;
7 Walsh, D.; Rodes, C. E.; Lawless, P. A. (2000) The 1999 Fresno paniculate matter exposure studies:
8 comparison of community, outdoor, and residential PM mass measurements. J. Air Waste Manage. Assoc.
9 50: 1887-1896.
10 Ferris, B. G., Jr.; Speizer, F. E.; Spengler, J. D.; Dockery, D.; Bishop, Y. M. M.; Wolfson, M.; Humble, C. (1979)
11 Effects of sulfur oxides and respirable particles on human health: methodology and demography of
12 populations in study. Am. Rev. Respir. Dis. 120: 767-779.
13 Fogh, C. L.; Byrne, M. A.; Roed, J.; Goddard, A. J. H. (1997) Size specific indoor aerosol deposition measurements
14 and derived I/O concentrations ratios. Atmos. Environ. 31: 2193-2203.
15 Freijer, J. L; Bloemen, H. J. T. (2000) Modeling relationships between indoor and outdoor air quality. J. Air Waste
16 Manage. Assoc. 50: 292-300.
17 Frey, H. C.; Rhodes, D. S. (1996) Characterizing, simulating and analyzing variability and uncertainty:
18 an illustration of methods using an air toxic emissions example. Hum. Ecol. Risk Assess. 2: 762-797'.
19 Gladen, B.; Rogan, W. J. (1979) Misclassification and the design of environmental studies. Am. J. Epidemiol.
20 109:607-616.
21 Glen, G.; Lakkadi, Y.; Tippett, J. A.; del Valle-Torres, M. (1997) Development of NERL/CHAD: the National
22 Exposure Research Laboratory consolidated human activity database. Research Triangle Park, NC: U.S.
23 Environmental Protection Agency, Office of Research and Development; contract no. 68-D5-0049.
24 Goldstein, B. R.; Tardiff, G.; Hoffnagle, G.; Kester, R. (1992) Valdez health study: summary report. Ankorage, AK:
25 Alyeska Pipeline Service Company.
26 Hadwen, G. E.; McCarthy, J. F.; Womble, S. E.; Girman, J. R.; Brightman, H. S. (1997) Volatile organic compound
27 concentrations in 41 office buildings in the continental United States. In: Woods, J. E.; Grimsrud, D. T.;
28 Boschi, N., eds. Proceedings of healthy buildings/indoor air quality international conference; September;
29 Bethesda, MD. Washington, DC: Healthy Buildings/IAQ. [Healthy Buildings - International Conference;
30 v. 2, pp. 465-470.]
31 Hartwell, T. D.; Clayton, C. A.; Michie, R. M., Jr.; Whitmore, R. W.; Zelon, H. S.; Whitehurst, D. A.; Akland,
32 G. G. (1984) Study of carbon monoxide exposures of residents of Washington, D.C. Research Triangle Park,
33 NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
34 EPA-600/D-84-177. Available from: NTIS, Springfield, VA; PB84-230051.
35 Hayes, S. R.; Marshall, J. R. (1999) Designing optimal strategies to attain the new U.S. paniculate matter standards:
36 some initial concepts. J. Air Waste Manage. Assoc. 49: PM192-PM198.
37 Henry, R. C.; Lewis, C. W.; Collins, J. F. (1994) Vehicle-related source compositions from ambient data:
38 the GRACE/SAFER method. Environ. Sci. Technol. 28: 823-832.
39 Houseman, E. A.; Ryan, L.; Levy; Richardson, D.; Spenger, J. J. (2001) Agri. Biol. Environ. Stat: In Press.
40 Howard-Reed, C.; Rea, A. W.; Zufall, M. J.; Burke, J. M.; Williams, R. W.; Suggs, J. C.; Sheldon, L. S.; Walsh, D.;
41 Kwock, R. (2000) Use of a continuous nephelometer to measure personal exposure to particles during the
42 U.S. Environmental Protection Agency Baltimore and Fresno panel studies. J. Air Waste Manage. Assoc.
43 50: 1125-1132.
44 Institute for Social Research. (1997) Child development supplement to the panel study of income dynamics.
45 Ann Arbor, MI: University of Michigan. Available: www.isr.umich.edu/src/child-development/home.html
46 [2001, January 31].
47 Isukapalli, S. S.; Georgopoulos, P. G. (2000) Modeling of outdoor/indoor relationships of gas phase pollutants and
48 paniculate matter. Presented at: PM2000: paniculate matter and health—the scientific basis for regulatory
49 decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste
50 Management Association.
51 Jang, M.; Kamens, R. M. (1999) Newly characterized products and composition of secondary aerosols from the
52 reaction of a-pinene with ozone. Atmos. Environ. 33: 459-474.
53 Janssen, N. (1998) Personal exposure to airborne particles: validity of outdoor concentrations as a measure of
54 exposure in time series studies [dissertation]. Wageningen, Netherlands: Landbouwuniversiteit Wageningen.
March 2001 5-104 DRAFT-DO NOT QUOTE OR CITE
-------
1 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Branekreef, B. (1995) A relation between personal and ambient.
2 Epidemiology 6(suppl.): S45.
3 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Branekreef, B. (1997) Childhood exposure to PM10: relation between
4 personal, classroom, and outdoor concentrations. Occup. Environ. Med. 54: 888-894.
5 Janssen, N. A. H.; Hoek, G.; Branekreef, B.; Harssema, H.; Mensink, I.; Zuidhof, A. (1998a) Personal sampling of
6 particles in adults: relation among personal, indoor, and outdoor air concentrations. Am. J. Epidemiol.
7 147: 537-547.
8 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Branekreef, B. (1998b) Personal sampling of airborne particles:
9 method performance and data quality. J. Exposure Anal. Environ. Epidemiol. 8: 37-49.
10 Janssen, N. A. H.; Hoek, G.; Branekreef, B.; Harssema, H. (1999a) Mass concentration and elemental composition
11 of PM10 in classrooms. Occup. Environ. Med. 56: 482-487.
12 Janssen, L. H. J. M.; Buringh, E.; van der Meulen, A.; van den Hout, K. D. (1999b) A method to estimate the
13 distribution of various fractions of PM10 in ambient air in the Netherlands. Atmos. Environ. 33: 3325-3334.
14 Janssen, N. A. H.; De Hartog, J. J.; Hoek, G.; Branekreef, B.; Lanki, T.; Timonen, K. L.; Pekkanen, J. (2000)
15 Personal exposure to fine paniculate matter in elderly subjects: relation between personal, indoor, and
16 outdoor concentrations. J. Air Waste Manage. Assoc. 50: 1133-1143.
17 Jantunen, M. J.; Hanninen, O.; Katsouyanni, K.; Knoppel, H.; Kuenzli, N.; Lebret, E.; Maroni, M.; Saarela, K.;
18 Sram, R.; Zmirou, D. (1998) Air pollution exposure in European cities: the EXPOLIS study. J. Exposure
19 Anal. Environ. Epidemiol. 8: 495-518.
20 Jantunen, M. J.; Kousa, A.; Hanninen, O.; Koistinen, K.; Sram, R.; Lanki, T.; Lioy, P. J. (2000) Long term
21 population PM exposures correlate well with outdoor air levels - short term individual exposures do not.
22 Presented at: PM2000: paniculate matter and health—the scientific basis for regulatory decision-making,
23 specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management
24 Association.
25 Jenkins, R. A.; Palausky, M. A.; Counts, R. W.; Guerin, M. R.; Dindal, A. B.; Bayne, C. K. (1996a) Determination
26 of personal exposure of non-smokers to environmental tobacco smoke in the United States. Lung Cancer
27 14(suppl. 1): S195-S213.
28 Jenkins, R. A.; Palausky, A.; Counts, R. W.; Bayne, C. K.; Dindal, A. B.; Guerin, M. R. (1996) Exposure to
29 environmental tobacco smoke in sixteen cities in the United States as determined by personal breathing zone
30 air sampling. J. Exposure Anal. Environ. Epidemiol. 6: 473-502.
31 Jensen, S. S. (1999) A geographic approach to modeling human exposure to traffic air pollution using GIS.
32 Roskilde, Denmark: National Environmental Research Institute.
33 Johnson, T. (1984) A study of personal exposure to carbon monoxide in Denver, Colorado. Research Triangle Park,
34 NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
35 EPA-600/4-84-014.
36 Johnson, T. (1989) Human activity patterns in Cincinnati, Ohio [final report]. Palo Alto, CA: Electric Power
37 Research Institute; project no. RP940-06; report no. EN-6204.
3 8 Johnson, T. R. (1995) Recent advances in the estimation of population exposure to mobile source pollutants.
39 J. Exposure Anal. Environ. Epidemiol. 5: 551-571.
40 Johnson, T.; Long, T.; Ollison, W. (2000) Prediction of hourly microenvironmental concentrations of fine particles
41 based on measurements obtained from the Baltimore scripted activity study. J. Exposure Anal. Environ.
42 Epidemiol. 10:403-411.
43 Jones, N. C.; Thornton, C. A.; Mark, D.; Harrison, R. M. (2000) Indoor/outdoor relationships of paniculate matter
44 in domestic homes with roadside, urban and rural locations. Atmos. Environ. 34: 2603-2612.
45 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
46 segregated mass concentration data and ultrafine particle number concentrations in urban areas.
47 Atmos. Environ. 33: 2907-2913.
48 Kleeman, M. J.; Cass, G. R. (1998) Source contributions to the size and composition distribution of urban
49 paniculate air pollution. Atmos. Environ. 32: 2803-2816.
50 Klepeis, N. E. (1999) An introduction to the indirect exposure assessment approach: modeling human exposure
51 using microenvironmental measurements and the recent national human activity pattern survey. Environ.
52 Health Perspect. Suppl. 107(2): 365-374.
53
March 2001 5-105 DRAFT-DO NOT QUOTE OR CITE
-------
1 Klepeis, N. E.; Ott, W. R.; Switzer, P. (1994) A Total Human Exposure Model (THEM) for Respirable Suspended
2 Particles (RSP). Presented at: 87th annual meeting & exhibition of the Air & Waste Management
3 Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no. 94-
4 WA75A.03. Available from: NTIS, Springfield, VA; PB94-197415.
5 Klepeis, N. E.; Ott, W. R.; Switzer, P. (1995) Modeling the time series of respirable suspended particles and carbon
6 monoxide from multiple smokers: validation in two public smoking lounges. Presented at: 88th annual
7 meeting and exhibition of the Air & Waste Management Association; June; San Antonio, TX. Pittsburgh,
8 PA: Air & Waste Management Association; paper no. 95-MPH.03.
9 Klepeis, N. E.; Ott, W. R.; Switzer, P. (1996) A multiple-smoker model for predicting indoor air quality in public
10 lounges. Environ. Sci. Technol. 30: 2813-2820.
11 Klepeis, N. E.; Switzer, P.; Ott, W. R. (2000) Estimating exposure to environmental tobacco smoke particles using
12 24-hour time-activity recall diaries. Presented at: PM2000: paniculate matter and health—the scientific basis
13 for regulatory decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA:
14 Air & Waste Management Association.
15 Koontz, M. D.; Rector, H. E. (1995) Estimation of distributions for residential air exchange rates: final report.
16 Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics;
17 contract nos. 68-D9-0166 and 68-D3-0013.
18 Koontz, M. D.; Niang, L. L. (1998) California population indoor exposure model (CPIEM): version 1.4F. Final
19 report A933-157: user's guide. Sacramento, CA: California Air Resources Board; GEOMET report no.
20 IE-2632.
21 Koutrakis, P.; Briggs, S. L. K.; Leaderer, B. P. (1992) Source apportionment of indoor aerosols in Suffolk and
22 Onondaga Counties, New York. Environ. Sci. Technol. 26: 521-527.
23 Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J.; Jerrett, M.; Abrahamowicz, M.;
24 White, W. H. (2000) Reanalysis of the Harvard Six Cities study and the American Cancer Society study of
25 paniculate air pollution and mortality: a special report of the Institute's Particle Epidemiology Reanalysis
26 Project. Cambridge, MA: Health Effects Institute.
27 Lachenmyer, C.; Hidy, G. M. (2000) Urban measurements of outdoor-indoor PM2 5 concentrations and personal
28 exposure in the deep south. Part I. Pilot study of mass concentrations for nonsmoking subjects. Aerosol Sci.
29 Technol. 32: 34-51.
30 Laden, F.; Neas, L. M.; Dockery, D. W.; Schwartz, J. (2000) Association of fine paniculate matter from different
31 sources with daily mortality in six U.S. cities. Environ. Health Perspect. 108: 941-947.
32 Lai, A. C. K.; Nazaroff, W. W. (2000) Modeling indoor particle deposition from turbulent flow onto smooth
33 surfaces. J. Aerosol Sci. 31: 463-476.
34 Leaderer, B. P.; Naeher, L.; Jankun, T.; Balenger, K.; Holford, T. R.; Toth, C.; Sullivan, J.; Wolf son, J. M.;
35 Koutrakis, P. (1999) Indoor, outdoor, and regional summer and winter concentrations of PM10, PM25, SO42",
36 H+, NH4+, NO3", NH3, and nitrous acid in homes with and without kerosene space heaters. Environ. Health
37 Perspect. 107:223-231.
38 Lee, S. C.; Chang, M. (1999) Indoor air quality investigations at five classrooms. Indoor Air 9: 134-138.
39 Lessler, J. T.; Kalsbeek, W. D. (1992) Nonsampling error in surveys. New York, NY: John Wiley & Sons, Inc.;
40 p. 7.
41 Liao, D.; Creason, J.; Shy, C.; Williams, R.; Watts, R.; Zweidinger, R. (1999) Daily variation of paniculate air
42 pollution and poor cardiac autonomic control in the elderly. Environ. Health Perspect. 107: 521-525.
43 Ligocki, M. P.; Liu, H. I. H.; Cass, G. R.; John, W. (1990) Measurements of particle deposition rates inside
44 Southern California museums. Aerosol Sci. Technol. 13: 85-101.
45 Lillquist, D. R.; Lee, J. S.; Ramsay, J. R.; Boucher, K. M.; Walton, Z. L.; Lyon, J. L. (1998) A comparison of
46 indoor/outdoor PM10 concentrations measured at three hospitals and a centrally located monitor in Utah.
47 Appl. Occup. Environ. Hyg. 13: 409-415.
48 Linn, W. S.; Gong, H., Jr.; Clark, K. W.; Anderson, K. R. (1999) Day-to-day paniculate exposures and health
49 changes in Los Angeles area residents with severe lung disease. J. Air Waste Manage. Assoc.
50 49:PM108-PM115.
51 Lioy, P. J.; Waldman, J. M.; Buckley, T.; Butler, J.; Pietarinen, C. (1990) The personal, indoor and outdoor
52 concentrations of PM-10 measured in an industrial community during the winter. Atmos. Environ. Part B
53 24:57-66.
54 Lioy, P. J.; Wainman, T.; Zhang, J.; Goldsmith, S. (1999) Typical household vacuum cleaners: the collection
55 efficiency and emissions characteristics for fine particles. J. Air Waste Manage. Assoc. 49: 200-206.
March 2001 5-106 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lippmann, M; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
2 mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report no.
3 95.
4 Long, C. M.; Suh, H. H.; Koutrakis, P. (2000a) Characterization of indoor particle sources using continuous mass
5 and size monitors. J. Air Waste Manage. Assoc. 50: 1236-1250.
6 Long, C. M.; Suh, H. H.; Koutrakis, P. (2000b) Using time- and size-resolved paniculate data to explore infiltration
7 and deposition behavior. Environ. Sci. Technol.: accepted.
8 Lurmann, F. W.; Korc, M. E. (1994) User's guide to the regional human exposure (REHEX) model. San Francisco,
9 CA: Bay Area Air Quality Management District; draft report no. STI-93150-1414-DR.
10 Macintosh, D. L.; Xue, J.; Ozkaynak, H.; Spengler, J. D.; Ryan, P. B. (1995) A population-based exposure model
11 for benzene. J. Expos. Anal. Environ. Epidemiol. 5: 375-403.
12 Mage, D. T. (1985) Concepts of human exposure assessment for airborne paniculate matter. Environ. Int.
13 11:407-412.
14 Mage, D. T. (1998) How much PM of ambient origin are people exposed to? In: Measurement of toxic and related
15 air pollutants. Volume II: proceedings of a specialty conference; September; Gary, NC. Pittsburgh, PA: Air &
16 Waste Management Association; pp. 666-681. (A&WMA publication VIP-85).
17 Mage, D. T.; Ott, W. R. (1996) Accounting for nonuniform mixing and human exposure in indoor environments.
18 In: Tichenor, B. A., ed. Characterizing sources of indoor air pollution and related sink effects. Philadelphia,
19 PA: American Society for Testing and Materials; pp. 263-278.
20 Mage, D.; Wilson, W.; Hasselblad, V.; Grant, L. (1999) Assessment of human exposure to ambient paniculate
21 matter. J. Air Waste Manage. Assoc. 49: 174-185.
22 Mar, T. F.; Norris, G. A.; Koenig, J. Q.; Larson, T. V. (2000) Associations between air pollution and mortality in
23 Phoenix, 1995-1997. Environ. Health Perspect. 108: 347-353.
24 McBride, S. J.; Ferro, A. R.; Ott, W. R; Switzer, P.; Hildemann, L. M. (1999) Investigations of the proximity effect
25 for pollutants in the indoor environment. J. Exposure Anal. Environ. Epidemiol. 9: 602-621.
26 McCurdy, T. (1995) Estimating human exposure to selected motor vehicle pollutants using the NEM series of
27 models: lessons to be learned. J. Exposure Anal. Environ. Epidemiol. 5: 533-550.
28 McCurdy, T. (1997) Modeling the dose profile in human exposure assessments: ozone as an example. Rev. Toxicol.
29 1:3-23.
30 McCurdy, T. (2000) Conceptual basis for multi-route intake dose modeling using an energy expenditure approach.
31 J. Exposure Anal. Environ. Epidemiol. 10: 86-97.
32 Mihlan, G. J.; Todd, L. A.; Truong, K. N. (2000) Assessment of occupational exposure patterns by
33 frequency-domain analysis of time series data. Appl. Occup. Environ. Hyg. 15: 120-130.
34 Miller, S. L.; Branoff, S.; Nazaroff, W. W. (1998) Exposure to toxic air contaminants in environmental tobacco
3 5 smoke: an assessment for California based on personal monitoring data. J. Exposure Anal. Environ.
36 Epidemiol. 8:287-311.
37 Monn, Ch.; Carabias, V.; Junker, M.; Waeber, R.; Karrer, M.; Wanner, H. U. (1997) Small-scale spatial variability
38 of paniculate matter <10 um (PM10) and nitrogen dioxide. Atmos. Environ. 31: 2243-2247.
39 Murray, D. M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
40 parametric distributions by season and climatic region. Risk Anal. 15: 459-465.
41 National Air Pollution Control Administration. (1969) Air quality criteria for paniculate matter. Washington, DC:
42 U.S. Department of Health, Education, and Welfare, Public Health Service; NAPCA publication no. AP-49.
43 Available from: NTIS, Springfield, VA; PB-190251/BA.
44 National Research Council. (1991) Human exposure assessment for airborne pollutants: advances and opportunities.
45 Washington, DC: National Academy of Sciences.
46 National Research Council. (1998) Research priorities for airborne paniculate matter. I. Immediate priorities and a
47 long-range research portfolio. Washington, DC: National Academy Press.
48 National Research Council. (1999) Research priorities for airborne paniculate matter. II. Evaluating research
49 progress and updating the portfolio. Washington, DC: National Academy Press.
50 Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
51 effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
52 Nelson, P. R.; Conrad, F. W.; Kelly, S. P.; Maiolo, K. C.; Richardson, J. D.; Ogden, M. W. (1998) Composition of
53 environmental tobacco smoke (ETS) from international cigarettes part II: nine country follow-up. Environ.
54 Int. 24:251-257.
March 2001 5-107 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nieuwenhuijsen, M. J.; Noderer, K. S.; Schenker, M. B.; Vallyathan, V.; Olenchock, S. (1999) Personal exposure to
2 dust, endotoxin and crystalline silica in California agriculture. Ann. Occup. Hyg. 43: 35-42.
3 Oglesby, L.; Ktinzli, N.; Roosli, M.; Braun-Fahrlander, C.; Mathys, P.; Stern, W.; Jantunen, M.; Kousa, A. (2000)
4 Validity of ambient levels of fine particles as surrogate for personal exposure to outdoor air
5 pollution—results of the European EXPOLIS-EAS study (Swiss Center Basel). J. Air Waste Manage. Assoc.
6 50: 1251-1261.
7 Ott, W. R. (1984) Exposure estimates based on computer generated activity patterns. J. Toxicol. Clin. Toxicol.
8 21:97-128.
9 Ott, W. R. (1999) Mathematical models for predicting indoor air quality from smoking activity. Environ. Health
10 Perspect. 107(suppl. 2): 375-381.
11 Ott, W.; Thomas, J.; Mage, D.; Wallace, L. (1988) Validation of the simulation of human activity and pollutant
12 exposure (SHAPE) model using paired days from the Denver, CO, carbon monoxide field study. Atmos.
13 Environ. 22:2101-2113.
14 Ott, W. R.; Mage, D. T.; Thomas, J. (1992) Comparison of microenvironmental CO concentrations in two cities for
15 human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 2: 249-267.
16 Ott, W. R.; Klepeis, N. E.; Switzer, P. (1995) Modeling environmental tobacco smoke in the home using transfer
17 functions. Presented at: 88th annual meeting and exhibition of the Air & Waste Management Association;
18 June; San Antonio, TX; paper no. 95-WP84B.03. Research Triangle Park, NC: U.S. Environmental
19 Protection Agency, Atmospheric Research And Exposure Assessment Laboratory; report no.
20 EPA/600/A-95/079. Available from: NTIS, Springfield, VA; PB95-225488.
21 Ott, W.; Switzer, P.; Robinson, J. (1996) Particle concentrations inside a tavern before and after prohibition of
22 smoking: evaluating the performance of an indoor air quality model. J. Air Waste Manage. Assoc.
23 46: 1120-1134.
24 Ott, W.; Wallace, L.; Mage, D. (2000) Predicting paniculate (PM10) personal exposure distributions using a random
25 component superposition (RCS) statistical model. J. Air Waste Manage. Assoc. 50: 1390-1406.
26 Owen, M. K.; Ensor, D. S.; Sparks, L. E. (1992) Airborne particle sizes and sources found in indoor air.
27 Atmos. Environ. Part A 26: 2149-2162.
28 Ozkaynak, H.; Spengler, J. (1996) The role of outdoor paniculate matter in assessing total human exposure.
29 In: Wilson, R.; Spengler, J., eds. Particles in our air: concentrations and health effects. Cambridge, MA:
30 Harvard University Press.
31 Ozkaynak, H.; Thurston, G. D. (1987) Associations between 1980 U.S. mortality rates and alternative measures of
32 airborne particle concentration. Risk Anal. 7: 449-461.
33 Ozkaynak, H.; Ryan, P. B.; Spengler, J. D.; Laird, N. M. (1986) Bias due to misclassification of personal exposures
34 in epidemiologic studies of indoor and outdoor air pollution. In: Berglund, B.; Berglund, U.; Lindvall, T.;
35 Spengler, J.; Sundell, J., eds. Indoor air quality: papers from the third international conference on indoor air
36 quality and climate; August 1984; Stockholm, Sweden. Environ. Int. 12: 389-393.
37 Ozkaynak, H.; Spengler, J. D.; Ludwig, J. F.; Butler, D. A.; Clayton, C. A.; Pellizzari, E.; Wiener, R. W. (1990)
3 8 Personal exposure to paniculate matter: findings from the Particle Total Exposure Assessment Methodology
39 (PTEAM) prepilot study. In: Indoor air '90: precedings of the 5th international conference on indoor air
40 quality and climate, volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON,
41 Canada: International Conference on Indoor Air Quality and Climate, Inc.; pp. 571-576.
42 Ozkaynak, H.; Spengler, J. D.; Xue, J.; Koutrakis, P.; Pellizzari, E. D.; Wallace, L. (1993) Sources and factors
43 influencing personal and indoor exposures to particles, elements and nicotine: findings from the particle
44 TEAM pilot study. In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma,
45 H., eds. Indoor air '93: proceedings of the 6th international conference on indoor air quality and climate, v. 3,
46 combustion products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93;
47 pp. 457-462.
48 Ozkaynak, H.; Xue, J.; Spengler, J.; Wallace, L.; Pellizzari, E.; Jenkins, P. (1996) Personal exposures to airborne
49 particles and metals: results from the particle TEAM study in Riverside, California. J. Exp. Anal. Environ.
50 Epidemiol. 6: 57-78.
51 Ozkaynak, H.; Xue, J.; Weker, R.; Butler, D.; Koutrakis, P.; Spengler, J. (1996b) The particle TEAM (PTEAM)
52 study: analysis of the data: final report, volume III. Research Triangle Park, NC: U.S. Environmental
53 Protection Agency, Atmospheric Research and Exposure Assessment Laboratory; report no.
54 EPA/600/R-95/098. Available from: NTIS, Springfield, VA; PB97-102495.
March 2001 5-108 DRAFT-DO NOT QUOTE OR CITE
-------
1 Patterson, E.; Eatough, D. J. (2000) Indoor/outdoor relationships for ambient PM2 5 and associated pollutants:
2 epidemiological implications in Lindon, Utah. J. Air Waste Manage. Assoc. 50: 103-110.
3 Pellizzari, E. D., Mason, R. E.; Clayton, C. A.; Thomas, K. W.; Cooper, S.; Piper. L.; Rodes, C.; Goldberg, M.;
4 Roberds, J.; Michael, L. (1998) Manganese exposure study (Toronto). Final report. Research Triangle Park,
5 NC: Research Triangle Institute, Analytical and Chemical Sciences; report no. RTI/6312/02-01 F.
6 Pellizzari, E. D.; Clayton, C. A.; Rodes, C. E.; Mason, R. E.; Piper, L. L.; Fort, B.; Pfeifer, G.; Lynam, D. (1999)
7 Paniculate matter and manganese exposures in Toronto, Canada. Atmos. Environ. 33: 721-734.
8 Phillips, K.; Howard, D. A.; Browne, D.; Lewsley, J. M. (1994) Assessment of personal exposures to environmental
9 tobacco smoke in British nonsmokers. Environ. Int. 20: 693-712.
10 Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G. (1996) Assessment of air quality in Stockholm by personal
11 monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Scand. J.
12 Work Environ. Health 22(suppl. 1): 1-24.
13 Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G.; Huici, A. (1997a) Assessment of air quality in Barcelona by
14 personal monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke.
15 Environ. Int. 23: 173-196.
16 Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1997b) Assessment of air quality in Turin by personal
17 monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Environ. Int.
18 23:851-871.
19 Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G. (1998a) Assessment of air quality in Paris by personal
20 monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Environ. Int.
21 24:405-425.
22 Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1998b) Assessment of environmental tobacco smoke and
23 respirable suspended particle exposures for nonsmokers in Lisbon by personal monitoring. Environ. Int.
24 24:301-324.
25 Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1999) Assessment of environmental tobacco smoke and
26 respirable suspended particle exposures for nonsmokers in Basel by personal monitoring. Atmos. Environ.
27 33: 1889-1904.
28 Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
29 (1995) Paniculate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
30 Respir. Crit. Care Med. 151: 669-674.
31 Pope, C. A., Ill; Hill, R. W.; Villegas, G. M. (1999) Paniculate air pollution and daily mortality on Utah's Wasatch
32 Front. Environ. Health Perspect: 107: 567-573.
33 Praml, G.; Schierl, R. (2000) Dust exposure in Munich public transportation: a comprehensive 4-year survey in
34 buses and trams. Int. Arch. Occup. Environ. Health 73: 209-214.
35 Quintana, P. J. E.; Samimi, B. S.; Kleinman, M. T.; Liu, L.-J.; Soto, K.; Warner, G. Y.; Bufalino, C.; Valencia, J.;
36 Francis, D.; Hovell, M. H.; Delfino, R. J. (2000) Evaluation of a real-time passive personal particle monitor
37 in fixed site residential indoor and ambient measurements. J. Exposure Anal. Environ. Epidemiol.
38 10:437-445.
39 Ramachandran, G.; Vincent, J. H. (1999) A Bayesian approach to retrospective exposure assessment. Appl. Occup.
40 Environ. Hyg. 14: 547-557.
41 Rea, A. W.; Zufall, M. J.; Williams, R. W.; Howard-Reed, C.; Sheldon, L. S. (2001) The influence of human
42 activity patterns on personal PM exposure: a comparative analysis of filter-based and continuous particle
43 measurements. J. Air Waste Manage. Assoc.: In Press.
44 Ren, P.; Jankun, T. M.; Leaderer, B. P. (1999) Comparisons of seasonal fungal prevalence in indoor and outdoor air
45 and in house dusts of dwellings in one northeast American county. J. Exposure Anal. Environ. Epidemiol.
46 9:560-568.
47 Robinson, J. P.; Wiley, J. A.; Piazza, T.; Garrett, K.; Cirksena, K. (1989) Activity patterns of California residents
48 and their implications for potential exposure to pollution. Sacramento, CA: California Air Resources Board
49 report no. CARB-A6-177-33.
50 Robinson, J. P.; Switzer, P.; Ott, W. (1994) Microenvironmental factors related to Californians' potential exposure
51 to environmental tobacco smoke (ETS). Research Triangle Park, NC: U.S. Environmental Protection Agency,
52 Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/R-94/116. Available from:
53 NTIS, Springfield, VA; PB94-214814.
54
March 2001 5-109 DRAFT-DO NOT QUOTE OR CITE
-------
1 Rodes, C.; Sheldon, L.; Whitaker, D.; Clayton, A.; Fitzgerald, K.; Flanagan, J.; DiGenova, F.; Hering, S.;
2 Frazier, C. (1998) Measuring concentrations of selected air pollutants inside California vehicles [final
3 report]. Sacramento, CA: California Environmental Protection Agency, Air Resources Board; contract no.
4 95-339.
5 Rojas-Bracho, L.; Suh, H. H.; Koutrakis, P. (2000) Relationships among personal, indoor, and outdoor fine and
6 coarse particle concentrations for individuals with COPD. J. Exposure Anal. Environ. Epidemiol.
7 10:294-306.
8 Roorda-Knape, M. C.; Janssen, N. A. H.; De Hartog, J. J.; Van Vliet, P. H. N.; Harssema, H.; Brunekreef, B. (1998)
9 Air pollution from traffic in city districts near major motorways. Atmos. Environ. 32: 1921-1930.
10 Ryan, P. B. (1991) An overview of human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 1: 453-474.
11 Santos-Burgoa, C.; Rojas-Bracho, L.; Rosas-Perez, L; Ramirez-Sanchez, A.; Sanchez-Rico, G.;
12 Mejia-Hernandez, S. (1998) Modelaje de exposition a particulas en poblacion general y riesgo de
13 enfermedad respiratoria [Particle exposure modeling in the general population and risk of respiratory
14 disease]. Gac. Med. Mex. 134: 407-417.
15 Sarnat, J. A.; Koutrakis, P.; Suh, H. H. (2000) Assessing the relationship between personal paniculate and gaseous
16 exposures of senior citizens living in Baltimore, MD. J. Air Waste Manage. Assoc. 50: 1184-1198.
17 Schwartz, J.; Levin, R. (1999) Drinking water turbidity and health. Epidemiology 10: 86-90.
18 Sexton, K.; Ryan, P. B. (1988) Assessment of human exposure to air pollution: methods, measurements, and
19 models. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile, and public health.
20 Washington, DC: National Academy Press; pp. 207-238.
21 Sexton, K.; Spengler, J. D.; Treitman, R. D. (1984) Personal exposure to respirable particles: a case study in
22 Waterbury, Vermont. Atmos. Environ. 18: 1385-1398.
23 Sheldon, L.; Clayton, A.; Jones, B.; Keever, J.; Perritt, R.; Smith, D.; Whitaker, D.; Whitmore, R. (1992a) Indoor
24 pollutant concentrations and exposures. Sacramento, CA: California Air Resources Board; report no.
25 ARB/R-92/486.
26 Sheldon, L.; Clayton, A.; Keever, J. Perritt, R.; Whitaker, D. (1992b) PTEAM: monitoring of phthalates and PAHs
27 in indoor and outdoor air samples in Riverside, California, volume II [final report]. Sacramento, CA:
28 California Environmental Protection Agency, State Air Resources Board, Research Division; contract no.
29 A933-144. Available from: NTIS, Springfield, VA; PB93-205649/XAB.
30 Sheldon, L.; Clayton, A.; Perritt, R.; Whitaker, D. A.; Keever, J. (1993a) Indoor concentrations of polycyclic
31 aromatic hydrocarbons in California residences and their relationship to combustion source use. In: Jantunen,
32 M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings
33 of the 6th international conference on indoor air quality and climate. Volume 3: combustion products, risk
34 assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 29-34.
3 5 Sheldon, L.; Whitaker, D.; Keever, J.; Clayton, A.; Perritt, R. (1993b) Phthalates and PAHs in indoor and outdoor
36 air in a southern California community. In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.;
37 Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings of the 6th international conference on indoor
38 air quality and climate. Volume 3: combustion products, risk assessment, policies; July; Helsinki, Finland.
39 Helsinki, Finland: Indoor Air '93; pp. 109-114.
40 Sheldon, L.; Clayton, A.; Keever, J.; Perritt, R.; Whitaker, D. (1993c) Indoor concentrations of polycyclic aromatic
41 hydrocarbons in California residences. Sacramento, CA: Air Resources Board; ARE contract no. A033-132.
42 Shy, C. M.; Kleinbaum, D. G.; Morgenstern, H. (1978) The effect of misclassification of exposure status in
43 epidemiological studies of air pollution health effects. Bull. N. Y. Acad. Med. 54: 1155-1165.
44 Siegmann, K.; Scherrer, L.; Siegmann, H. C. (1999) Physical and chemical properties of airborne nanoscale
45 particles and how to measure the impact on human health. J. Mol. Struct. (Theochem) 458: 191-201.
46 Sinclair, J. D.; Psota-Kelty, L. A.; Weschler, C. J. (1988) Indoor/outdoor ratios and indoor surface accumulations of
47 ionic substances at Newark, New Jersey. Atmos. Environ. 22: 461-469.
48 Sinclair, J. D.; Psota-Kelty, L. A.; Weschler, C. J.; Shields, H. C. (1990) Measurement and modeling of airborne
49 concentrations and indoor surface accumulation rates of ionic substances at Neenah, Wisconsin. Atmos.
50 Environ. Part A 24: 627-638.
51 Spengler, J. D.; Treitman, R. D.; Tosteson, T. D.; Mage, D. T.; Soczek, M. L. (1985) Personal exposures to
52 respirable particulates and implications for air pollution epidemiology. Environ. Sci. Technol. 19: 700-707.
53
54
March 2001 5-110 DRAFT-DO NOT QUOTE OR CITE
-------
1 Spengler, J. D.; Ozkaynak, H.; Ludwig, J.; Allen, G.; Pellizzari, E. D.; Wiener, R. (1989) Personal exposures to
2 paniculate matter: instruments and methodologies P-TEAM. In: Measurement of toxic and related air
3 pollutants: proceedings of the 1989 U.S. EPA/A&WMA international symposium; May; Raleigh, NC.
4 Pittsburgh, PA: Air & Waste Management Association; pp. 449-463. (A&WMA publication VTP-13).
5 Spier, C. E.; Little, D. E.; Trim, S. C.; Johnson, T. R.; Linn, W. S.; Hackney, J. D. (1992) Activity patterns in
6 elementary and high school students exposed to oxidant pollution. J. Exposure Anal. Environ. Epidemiol.
7 2:277-293.
8 Suggs, J. C.; Burton, R. M. (1983) Spatial characteristics of inhalable particles in the Philadelphia metropolitan
9 area. J. AirPollut. Control Assoc. 33: 688-691.
10 Suh, H. H.; Spengler, J. D.; Koutrakis, P. (1992) Personal exposures to acid aerosols and ammonia. Environ. Sci.
11 Technol. 26:2507-2517.
12 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993a) Indoor and outdoor acid aerosol and gas concentrations.
13 In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air
14 '93: proceedings of the 6th international conference on indoor air quality and climate. Volume 3: combustion
15 products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 257-262.
16 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993b) Validation of personal exposure models for sulfate and aerosol
17 strong acidity. J. Air Waste Manage. Assoc. 43: 845-850.
18 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1994) The relationship between airborne acidity and ammonia in indoor
19 environments. J. Exposure Anal. Environ. Epidemiol. 4: 1-23.
20 Tamura, K.; Ando, M.; Sagai, M.; Matsumoto, Y. (1996a) Estimation of levels of personal exposure to suspended
21 paniculate matter and nitrogen dioxide in Tokyo. Environ. Sci. (Tokyo) 4: 37-51.
22 Tamura, K.; Ando, M.; Miyazaki, T.; Funasaka, K.; Kuroda, K. (1996b) Indoor concentrations and personal
23 exposure to paniculate matter at roadside houses. In: Indoor air '96: the 7th international conference on
24 indoor air quality, volume I; July; Nagoya, Japan. Tokyo, Japan: Institute of Public Health; pp. 599-602.
25 Teschke, K.; Marion, S. A.; Vaughan, T. L.; Morgan, M. S.; Camp, J. (1999) Exposures to wood dust in U.S.
26 industries and occupations, 1979 to 1997. Am. J. Ind. Med. 35: 581-589.
27 Thatcher, T. L.; Layton, D. W. (1995) Deposition, resuspension, and penetration of particles within a residence.
28 Atmos. Environ. 29: 1487-1497.
29 Thomas, K. W.; Pellizzari, E. D.; Clayton, C. A.; Whitaker, D. A.; Shores, R. C.; Spengler, J.; Ozkaynak, H.;
30 Froehlich, S. E.; Wallace, L. A. (1993) Particle total exposure assessment methodology (PTEAM) 1990
31 study: method performance and data quality for personal, indoor, and outdoor monitoring. J. Exposure Anal.
32 Environ. Epidemiol. 3:203-226.
33 Tsang, A. M.; Klepeis, N. E. (1996) Descriptive statistics tables from a detailed analysis of the national human
34 activity pattern survey (NHAPS) data. Las Vegas, NV: U.S. Environmental Protection Agency, Office of
3 5 Research and Development; report no. EPA/600/R-96/148.
36 U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
37 Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
3 8 Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
39 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
40 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
41 U.S. Environmental Protection Agency. (1998) Particulate matter research needs for human health risk assessment
42 to support future reviews of the national ambient air quality standards for paniculate matter. Research
43 Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
44 U.S. Environmental Protection Agency. (2000) Aerometric Information Retrieval System (AIRS). Research
45 Triangle Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/
46 [2000, October 6].
47 Vakeva, M.; Hameri, K.; Kulmala, M.; Lahdes, R.; Ruuskanen, J.; Laitinen, T. (1999) Street level versus rooftop
48 concentrations of submicron aerosol particles and gaseous pollutants in an urban street canyon.
49 Atmos. Environ. 33: 1385-1397.
50 Van Vliet, P.; Knape, M.; De Hartog, J.; Janssen, N.; Harssema, H.; Brunekreef, B. (1997) Motor vehicle exhaust
51 and chronic respiratory symptoms in children living near freeways. Environ. Res. 74: 122-132.
52 Vette, A. F.; Rea, A. W.; Lawless, P. A.; Rodes, C. E.; Evans, G.; Highsmith, V R.; Sheldon, L. (2001)
53 Characterization of indoor-outdoor aerosol concentration relationships during the Fresno PM exposure
54 studies. Aerosol Sci. Technol.: accepted.
March 2001 5-111 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of
2 submicron particle exposure. Environ. Health Perspect. 108: 1139-1145.
3 Waldman, J. M; Liang, C. S. K. (1993) Penetration and neutralization of acid aerosols at institutional settings.
4 In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air
5 '93: proceedings of the 6th international conference on indoor air quality and climate. Volume 3: combustion
6 products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 263-268.
7 Waldman, J. M.; Liang, C. S. K.; Menon, P. (1990) Indoor exposures to acidic aerosols at child and elderly care
8 facilities. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and climate,
9 volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
10 International Conference on Indoor Air Quality and Climate, Inc.; pp. 607-612.
11 Wallace, L. (1996) Indoor particles: a review. J. Air Waste Manage. Assoc. 46: 98-126.
12 Wallace, L. (2000a) Real-time monitoring of particles, PAH, and CO in an occupied townhouse. Appl. Occup.
13 Environ. Hyg. 15: 39-47.
14 Wallace, L. (2000b) Correlations of personal exposure to particles with outdoor air measurements: a review of
15 recent studies. Aerosol Sci. Technol. 32: 15-25.
16 Wallace, L.; Quackenboss, J.; Rodes, C. (1997) Continuous measurements of particles, PAH, and CO in an
17 occupied townhouse in Reston, VA. In: Measurement of toxic and related air pollutants. Volume II:
18 proceedings of a specialty conference; April-May; Research Triangle Park, NC. Pittsburgh, PA: Air & Waste
19 Management Association, pp. 860-871. (A&WMA publication VIP-74).
20 Watson, J. G.; Cooper, J. A.; Huntzicker, J. J. (1984) The effective variance weighting for least squares calculations
21 applied to the mass balance receptor model. Atmos. Environ. 18: 1347-1355.
22 Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
23 (1990) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
24 Watson, J. G.; Chow, J. C.; Pace, T. G. (1991) Chemical mass balance. In: Hopke, P. K., ed. Receptor modeling for
25 air quality management. New York, NY: Elsevier; pp. 83-116. (Data handling in science and technology—
26 volume 7).
27 Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
28 Environ. 33:2301-2312.
29 White, W. H. (1998) Statistical considerations in the interpretation of size-resolved paniculate mass data. J. Air
30 Waste Manage. Assoc. 48: 454-458.
31 Wiener, R. W. (1988) Measurement and evaluation of personal exposure to aerosols. In: Measurement of toxic and
32 related air pollutants: proceedings of the 1988 EPA/APCA international symposium; May; Research Triangle
33 Park, NC. Pittsburgh, PA: Air Pollution Control Association; pp. 84-88. (APCA publication VTP-10).
34 Wiener, R. W. (1989) Particle total exposure assessment methodology~an overview of planning and
3 5 accomplishments. In: Measurement of toxic and related air pollutants: proceedings of the 1989 U.S.
36 EPA/A&WMA international symposium; May; Raleigh, NC. Pittsburgh, PA: Air & Waste Management
37 Association; pp. 442-448. (A&WMA publication VTP-13).
38 Wiener, R. W.; Wallace, L.; Pahl, D.; Pellizzari, E.; Whittaker, D.; Spengler, J.; Ozkaynak, H. (1990) Review of the
3 9 particle TEAM 9 home field study. In: Measurement of toxic and related air pollutants: proceedings of the
40 1990 EPA/A&WMA international symposium, v. 2; May; Raleigh, NC. Pittsburgh, PA: Air & Waste
41 Management Association; pp. 452-460. (A&WMA publication VIP-17).
42 Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991a) Study of children's activity
43 patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
44 Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
45 (2000a) The 1998 Baltimore Paniculate Matter Epidemiology-Exposure Study: part 1. Comparison of
46 ambient, residential outdoor, indoor and apartment paniculate matter monitoring. J. Exposure Anal. Environ.
47 Epidemiol. 10: 518-532.
48 Williams, R.; Suggs, J.; Creason, J.; Rodes, C.; Lawless, P.; Kwok, R.; Zweidinger, R.; Sheldon, L. (2000b) The
49 1998 Baltimore paniculate matter epidemiology-exposure study: part 2. Personal exposure assessment
50 associated with an elderly study population. J. Exposure Anal. Environ. Epidemiol.: 10: 533-543.
51 Williams, R.; Creason, J.; Zweidinger, R.; Watts, R.; Sheldon, L.; Shy, C. (2000c) Indoor, outdoor, and personal
52 exposure monitoring of paniculate air pollution: the Baltimore elderly epidemiology-exposure pilot study.
53 Atmos. Environ. 34: 4193-4204.
54 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
55 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
March 2001 5-112 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wilson, W. E.; Mage, D. T.; Grant, L. D. (2000) Estimating separately personal exposure to ambient and
2 nonambient paniculate matter for epidemiology and risk assessment: why and how. J. Air Waste Manage.
3 Assoc. 50: 1167-1183.
4 Yakovleva, E.; Hopke, P. K.; Wallace, L. (1999) Receptor modeling assessment of particle total exposure
5 assessment methodology data. Environ. Sci. Technol. 33: 3645-3652.
6 Zeger, S. L.; Thomas, D.; Dominici, F.; Samet, J. M.; Schwartz, J.; Dockery, D.; Cohen, A. (2000) Exposure
7 measurement error in time-series studies of air pollution: concepts and consequences. Environ. Health
8 Perspect. 108: 419-426.
9 Zmirou, D.; Masclet, P.; Boudet, C.; Dor, F.; Dechenaux, J. (2000) Personal exposure to atmospheric polycyclic
10 aromatic hydrocarbons in a general adult population and lung cancer risk assessment. J. Occup. Environ.
11 Med. 42: 121-126.
12
March 2001 5-113 DRAFT-DO NOT QUOTE OR CITE
------- |