v>EPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/P-93/004aF
July 1996
Air Quality Criteria for
Ozone and Related
Photochemical Oxidants
Volume I of III
-------
EPA/600/P-93/004aF
July 1996
Air Quality Criteria for Ozone
and Related Photochemical Oxidants
Volume I of
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Printed on Recycled Paper
-------
Disclaimer
This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
-------
Preface
In 1971, the U.S. Environmental Protection Agency (EPA) promulgated National
Ambient Air Quality Standards (NAAQS) to protect the public health and welfare from
adverse effects of photochemical oxidants. In 1979, the chemical designation of the standards
was changed from photochemical oxidants to ozone (O3). This document focuses primarily
on the scientific air quality criteria for O3 and, to a lesser extent, on those for other
photochemical oxidants such as hydrogen peroxide and the peroxyacyl nitrates.
The EPA promulgates the NAAQS on the basis of scientific information contained
in air quality criteria issued under Section 108 of the Clean Air Act. The previous O3 criteria
document, Air Quality Criteria for Ozone and Other Photochemical Oxidants, was released in
August 1986 and a supplement, Summary of Selected New Information on Effects of Ozone on
Health and Vegetation, was released in January 1992. These documents were the basis for a
March 1993 decision by EPA that revision of the existing 1-h NAAQS for O3 was not
appropriate at that time. That decision, however, did not take into account some of the newer
scientific data that became available after completion of the 1986 criteria document. The
purpose of this revised air quality criteria document for O3 and related photochemical
oxidants is to critically evaluate and assess the latest scientific data associated with exposure
to the concentrations of these pollutants found in ambient air. Emphasis is placed on the
presentation of health and environmental effects data; however, other scientific data are
presented and evaluated in order to provide a better understanding of the nature, sources,
distribution, measurement, and concentrations of O3 and related photochemical oxidants and
their precursors in the environment. Although the document is not intended to be an
exhaustive literature review, it is intended to cover all pertinent literature available through
1995.
This document was prepared and peer reviewed by experts from various state and
Federal governmental offices, academia, and private industry and reviewed in several public
meetings by the Clean Air Scientific Advisory Committee. The National Center for
Environmental Assessment (formerly the Environmental Criteria and Assessment Office) of
EPA's Office of Research and Development acknowledges with appreciation the contributions
provided by these authors and reviewers as well as the diligence of its staff and contractors in
the preparation of this document at the request of the Office of Air Quality Planning and
Standards.
-------
-------
Air Quality Criteria for Ozone
and Related Photochemical Oxidants
Table of Contents
Volume I
1. Executive Summary 1-1
2. Introduction 2-1
3. Tropospheric Ozone and Its Precursors 3-1
4. Environmental Concentrations, Patterns, and Exposure Estimates 4-1
Appendix A: Abbreviations and Acronyms A-l
Volume II
5. Environmental Effects of Ozone and Related Photochemical
Oxidants 5-1
Appendix A: Abbreviations and Acronyms A-l
Appendix B: Colloquial and Latin Names B-l
Volume III
6. Toxicological Effects of Ozone and Related Photochemical Oxidants ... 6-1
7. Human Health Effects of Ozone and Related Photochemical Oxidants ... 7-1
8. Extrapolation of Animal Toxicological Data to Humans 8-1
9. Integrative Summary of Ozone Health Effects 9-1
Appendix A: Abbreviations and Acronyms A-l
l-v
-------
-------
Table of Contents
List of Tables
List of Figures
Authors, Contributors, and Reviewers
U.S. Environmental Protection Agency Science Advisory Board,
Clean Air Scientific Advisory Committee
U.S. Environmental Protection Agency Project Team for Development
of Air Quality Criteria for Ozone and Related Photochemical Oxidants
Page
I-xiii
I-xviii
I-xxv
I-xxxi
I-xxxiii
1. EXECUTIVE SUMMARY
1.1 INTRODUCTION
1.2 LEGISLATIVE AND REGULATORY BACKGROUND ..
1.3 TROPOSPHERIC OZONE AND ITS PRECURSORS
1.4 ENVIRONMENTAL CONCENTRATIONS, PATTERNS,
AND EXPOSURE ESTIMATES
1.5 ENVIRONMENTAL EFFECTS OF OZONE AND
RELATED PHOTOCHEMICAL OXIDANTS
1.6 TOXICOLOGICAL EFFECTS OF OZONE AND
RELATED PHOTOCHEMICAL OXIDANTS
1.7 HUMAN HEALTH EFFECTS OF OZONE AND
RELATED PHOTOCHEMICAL OXIDANTS
1.8 EXTRAPOLATION OF ANIMAL TOXICOLOGICAL
DATA TO HUMANS
1.9 INTEGRATIVE SUMMARY OF OZONE HEALTH
EFFECTS
2. INTRODUCTION
2.1 LEGISLATIVE BACKGROUND
2.2 REGULATORY BACKGROUND
2.3 SUMMARY OF MAJOR SCIENTIFIC TOPICS
PRESENTED
2.3.1 Air Chemistry
2.3.2 Air Quality
2.3.3 Environmental Effects
2.3.4 Health Effects
2.4 ORGANIZATION AND CONTENT OF THE DOCUMENT
REFERENCES
3. TROPOSPHERIC OZONE AND ITS PRECURSORS
3.1 INTRODUCTION
1-1
1-1
1-1
1-2
1-9
1-12
1-18
1-22
1-27
1-28
2-1
2-2
2-2
2-5
2-5
2-5
2-5
2-6
2-6
2-8
3-1
3-1
l-vii
-------
Table of Contents (cont'd)
Pagi
3.2 TROPOSPHERIC OZONE CHEMISTRY 3-2
3.2.1 Background Information 3-2
3.2.2 Structure of the Atmosphere 3-3
3.2.2.1 Vertical and Horizontal Mixing in
the Atmosphere 3-4
3.2.2.2 Formation of Stratospheric Ozone 3-4
3.2.3 Background Ozone in the Troposphere 3-6
3.2.3.1 Tropospheric Hydroxyl Radicals 3-7
3.2.3.2 Tropospheric Nitrogen Oxides Chemistry 3-8
3.2.3.3 The Methane Oxidation Cycle 3-10
3.2.3.4 Cloud Processes in the Methane-Dominated
Troposphere 3-15
3.2.4 Photochemistry of the Polluted Atmosphere 3-16
3.2.4.1 Tropospheric Loss Processes of
Volatile Organic Compounds 3-17
3.2.4.2 Chemical Formation of Ozone in Polluted
Air 3-30
3.2.4.3 Hydrocarbon Reactivity with Respect to
Ozone Formation 3-34
3.2.5 Photochemical Production of Aerosols 3-38
3.2.5.1 Phase Distributions of Organic Compounds ... 3-38
3.2.5.2 Acid Deposition 3-40
3.3 METEOROLOGICAL PROCESSES INFLUENCING OZONE
FORMATION AND TRANSPORT 3-42
3.3.1 Meteorological Processes 3-42
3.3.1.1 Surface Energy Budgets 3-42
3.3.1.2 Planetary Boundary Layer 3-43
3.3.1.3 Cloud Venting 3-46
3.3.1.4 Stratospheric-Tropospheric Ozone
Exchange 3-47
3.3.2 Meteorological Parameters 3-48
3.3.2.1 Sunlight 3-48
3.3.2.2 Temperature 3-49
3.3.2.3 Wind Speed 3-54
3.3.2.4 Air Mass Characteristics 3-56
3.3.3 Normalization of Trends 3-58
3.4 PRECURSORS OF OZONE AND OTHER OXIDANTS 3-59
3.4.1 Sources and Emissions of Precursors 3-59
3.4.1.1 Introduction 3-59
3.4.1.2 Nitrogen Oxides 3-60
3.4.1.3 Volatile Organic Compounds 3-70
l-viii
-------
Table of Contents (cont'd)
Page
3.4.1.4 Relationship of Summertime Precursor
Emissions and Ozone Production 3-79
3.4.2 Concentrations of Precursor Substances in Ambient
Air 3-80
3.4.2.1 Nonmethane Organic Compounds 3-81
3.4.2.2 Nitrogen Oxides 3-84
3.4.2.3 Ratios of Concentrations of Nonmethane
Organic Compounds and Nitrogen Oxides .... 3-85
3.4.3 Source Apportionment and Reconciliation 3-86
3.4.3.1 Source Apportionment 3-86
3.4.3.2 Source Reconciliation 3-89
3.5 ANALYTICAL METHODS FOR OXIDANTS AND
THEIR PRECURSORS 3-90
3.5.1 Sampling and Analysis of Ozone and Other
Oxidants 3-90
3.5.1.1 Ozone 3-90
3.5.1.2 Peroxyacetyl Nitrate and Its Homologues 3-101
3.5.1.3 Gaseous Hydrogen Peroxide 3-105
3.5.2 Sampling and Analysis of Volatile Organic
Compounds 3-107
3.5.2.1 Introduction 3-107
3.5.2.2 Nonmethane Hydrocarbons 3-108
3.5.2.3 Carbonyl Species 3-114
3.5.2.4 Polar Volatile Organic Compounds 3-116
3.5.3 Sampling and Analysis of Nitrogen Oxides 3-117
3.5.3.1 Introduction 3-117
3.5.3.2 Measurement of Nitric Oxide 3-118
3.5.3.3 Measurements for Nitrogen Dioxide 3-120
3.5.3.4 Calibration Methods 3-126
3.6 OZONE AIR QUALITY MODELS 3-127
3.6.1 Definitions, Description, and Uses 3-128
3.6.1.1 Grid-Based Models 3-129
3.6.1.2 Trajectory Models 3-131
3.6.2 Model Components 3-133
3.6.2.1 Emissions Inventory 3-133
3.6.2.2 Meteorological Input to Air Quality
Models 3-135
3.6.2.3 Chemical Mechanisms 3-139
3.6.2.4 Deposition Processes 3-140
3.6.2.5 Boundary and Initial Conditions 3-143
3.6.2.6 Numerical Methods 3-143
l-ix
-------
Table of Contents (cont'd)
Page
3.6.3 Urban and Regional Ozone Air Quality Models 3-144
3.6.3.1 The Urban Airshed Model 3-148
3.6.3.2 The Regional Oxidant Model 3-151
3.6.3.3 The Regional Acid Deposition Model 3-154
3.6.4 Evaluation of Model Performance 3-156
3.6.4.1 Model Performance Evaluation
Procedures 3-157
3.6.4.2 Performance Evaluation of Ozone Air
Quality Models 3-159
3.6.4.3 Database Limitations 3-160
3.6.5 Use of Ozone Air Quality Models for Evaluating
Control Strategies 3-162
3.6.6 Conclusions 3-163
3.7 SUMMARY AND CONCLUSIONS 3-165
3.7.1 Tropospheric Ozone Chemistry 3-165
3.7.1.1 Ozone in the Unpolluted Atmosphere 3-165
3.7.1.2 Ozone Formation in the Polluted
Troposphere 3-166
3.7.2 Meteorological Processes Influencing Ozone
Formation and Transport 3-168
3.7.2.1 Meteorological Processes 3-168
3.7.2.2 Meteorological Parameters 3-168
3.7.2.3 Normalization of Trends 3-169
3.7.3 Precursors 3-169
3.7.3.1 Volatile Organic Compound Emissions 3-169
3.7.3.2 Nitrogen Oxides Emissions 3-169
3.7.3.3 Concentrations of Volatile Organic
Compounds in Ambient Air 3-170
3.7.3.4 Concentrations of Nitrogen Oxides in
Ambient Air 3-170
3.7.3.5 Ratios of Concentrations of Nonmethane
Organic Compounds to Nitrogen Oxides 3-171
3.7.3.6 Source Apportionment and Reconciliation .... 3-171
3.7.4 Analytical Methods for Oxidants and Their
Precursors 3-172
3.7.4.1 Oxidants 3-172
3.7.4.2 Volatile Organic Compounds 3-173
3.7.4.3 Oxides of Nitrogen 3-174
3.7.5 Ozone Air Quality Models 3-174
3.7.5.1 Definitions, Descriptions, and Uses 3-174
3.7.5.2 Model Components 3-175
3.7.5.3 Evaluation of Model Performance 3-176
l-x
-------
Table of Contents (cont'd)
Page
3.7.5.4 Use of Ozone Air Quality Model for
Evaluating Control Strategies 3-176
3.7.5.5 Conclusions 3-176
REFERENCES 3-177
4. ENVIRONMENTAL CONCENTRATIONS, PATTERNS, AND
EXPOSURE ESTIMATES 4-1
4.1 INTRODUCTION 4-1
4.1.1 Characterizing Ambient Ozone Concentrations 4-2
4.1.2 The Identification and Use of Existing Ambient
Ozone Data 4-4
4.2 TRENDS IN AMBIENT OZONE CONCENTRATIONS 4-6
4.3 SURFACE OZONE CONCENTRATIONS 4-14
4.3.1 Introduction 4-14
4.3.2 Urban Area Concentrations 4-15
4.3.3 Nonurban Area Concentrations 4-27
4.3.3.1 Sites That Experience Low Maximum Hourly
Average Concentrations 4-27
4.3.3.2 Urban-Influenced Nonurban Areas 4-36
4.4 DIURNAL VARIATIONS IN OZONE CONCENTRATIONS . . . 4-46
4.4.1 Introduction 4-46
4.4.2 Urban Area Diurnal Patterns 4-47
4.4.3 Nonurban Area Diurnal Patterns 4-51
4.5 SEASONAL PATTERNS IN OZONE CONCENTRATIONS . . . 4-55
4.5.1 Urban Area Seasonal Patterns 4-55
4.5.2 Nonurban Area Seasonal Patterns 4-57
4.5.3 Seasonal Pattern Comparisons with Sites
Experiencing Low Exposures 4-61
4.6 SPATIAL VARIATIONS IN OZONE CONCENTRATIONS .... 4-62
4.6.1 Urban-Nonurban Area Concentration Differences 4-62
4.6.2 Concentrations Experienced at High-Elevation Sites .... 4-62
4.6.3 Other Spatial Variations in Ozone Concentrations 4-65
4.7 INDOOR OZONE CONCENTRATIONS 4-72
4.8 ESTIMATING EXPOSURE TO OZONE 4-73
4.8.1 Introduction 4-73
4.8.2 Fixed-Site Monitoring Information Used To Estimate
Population and Vegetation Exposure 4-76
4.8.3 Personal Monitors 4-77
4.8.4 Population Exposure Models 4-78
4.8.5 Concentration and Exposures Used in Research
Experiments 4-80
l-xi
-------
Table of Contents (cont'd)
Page
4.9 CONCENTRATIONS OF PEROXYACETYL NITRATES IN
AMBIENT ATMOSPHERES 4-81
4.9.1 Introduction 4-81
4.9.2 Urban Area Peroxyacetyl Nitrate Concentrations 4-82
4.9.3 Concentration of Peroxyacetyl Nitrate and
Peroxypropionyl Nitrate in Rural Areas 4-83
4.10 CONCENTRATION AND PATTERNS OF HYDROGEN
PEROXIDE IN THE AMBIENT ATMOSPHERE 4-86
4.11 CO-OCCURRENCE OF OZONE 4-88
4.11.1 Introduction 4-88
4.11.2 Nitrogen Oxides 4-89
4.11.3 Sulfur Dioxide 4-89
4.11.4 Acidic Sulfate Aerosols 4-90
4.11.5 Acid Precipitation 4-91
4.11.6 Acid Cloudwater 4-93
4.12 SUMMARY 4-94
REFERENCES 4-102
APPENDIX A: ABBREVIATIONS AND ACRONYMS A-l
l-xii
-------
List of Tables
Number Page
2-1 National Ambient Air Quality Standards for Ozone 2-3
3-1 Estimated Emissions of Methane, Nonmethane Organic
Compounds, Nitrous Oxide, and Nitrogen Oxides into the
Earth's Atmosphere from Biogenic and Anthropogenic
Sources 3-16
3-2 Calculated Tropospheric Lifetimes of Selected Volatile
Nonmethane Organic Compounds Due to Photolysis and
Reaction with Hydroxyl and Nitrate Radicals and with
Ozone 3-19
3-3 Calculated Incremental Reactivities of Selected Volatile
Organic Compounds as a Function of the Volatile Organic
Compound/Nitrogen Oxide Ratio for an Eight-Component
Volatile Organic Compound Mixture and Low-Dilution
Conditions 3-37
3-4 Rates of Increase of Peak Ozone with Diurnal Maximum
Temperature for Temperature Less Than 300 K and
Temperature Greater Than 300 K, Based on Measurements
for April 1 to September 30, 1988 3-52
3-5 Recent Studies Examining Trends in Ozone Data After Removal of
Variability Associated with Meteorological Factors 3-60
3-6 Source Categories Used To Inventory Nitrogen Oxides
Emissions 3-62
3-7 1991 Emission Estimates for Manmade Sources of Nitrogen
Oxides in the United States 3-63
3-8 Recent Trends in Nitrogen Oxides Emissions for Major
Manmade Source Categories 3-66
3-9 Comparison of Estimates of Nitrogen Oxides Emissions
from Manmade Sources in the United States 3-68
3-10 Annual Nitrogen Oxides Emissions from Soils by
U.S. Environmental Protection Agency Region 3-69
l-xiii
-------
List of Tables (cont'd)
Number Page
3-11 Estimated 1991 Emissions of Volatile Organic Compounds
from Manmade Sources in the United States 3-71
3-12 Recent Trends in Emissions of Volatile Organic Compounds
from Major Categories of Manmade Sources 3-73
3-13 Annual Biogenic Hydrocarbon Emission Inventory for the
United States 3-77
3-14 Annual Biogenic Hydrocarbon Emission Inventory by Month
and by U.S. Environmental Protection Agency Region for
United States Emissions 3-78
3-15 Performance Specifications for Automated Methods of Ozone
Analysis 3-92
3-16 Reference and Equivalent Methods for Ozone Designated
by the U.S. Environmental Protection Agency 3-93
3-17 List of Designated Reference and Equivalent Methods for
Ozone 3-94
3-18 Performance Specifications for Nitrogen Dioxide Automated
Methods 3-121
3-19 Comparability Test Specifications for Nitrogen Dioxide 3-121
3-20 Reference and Equivalent Methods for Nitrogen Dioxide
Designated by the U.S. Environmental Protection Agency 3-122
3-21 Grid-Based Urban and Regional Air Pollution Models: Overview
of Three-Dimensional Air Quality Models 3-145
3-22 Grid-Based Urban and Regional Air Pollution Models: Treatment
of Emissions and Spatial Resolution 3-146
3-23 Grid-Based Urban and Regional Air Pollution Models: Treatment
of Meteorological Fields, Transport, and Dispersion 3-147
3-24 Grid-Based Urban and Regional Air Pollution Models: Treatment
of Chemical Processes 3-149
l-xiv
-------
List of Tables (cont'd)
Number Page
3-25 Grid-Based Urban and Regional Air Pollution Models: Treatment
of Cloud and Deposition Processes 3-150
3-26 Regional Oxidant Model Geographical Domains 3-152
3-27 Applications of Photochemical Air Quality Models to Evaluating
Ozone 3-164
4-1 Ozone Monitoring Season by State 4-5
4-2 Summary by Forestry and Agricultural Regions for Ozone Trends
Using the W126 Exposure Parameter Accumulated on a Seasonal
Basis 4-13
4-3 The Highest Second Daily Maximum One-Hour Ozone
Concentration by Metropolitan Statistical Area for the Years
1989 to 1991 4-17
4-4 Summary of Percentiles of Hourly Average Concentrations for
the April-to-October Period 4-20
4-5 The Highest Second Daily Maximum Eight-Hour Average Ozone
Concentration by Metropolitan Statistical Area for the Years
1989 to 1991 4-22
4-6 Seasonal (April to October) Percentile Distribution
of Hourly Ozone Concentrations, Number of Hourly
Mean Ozone Occurrences Greater Than or Equal to 0.08
and Greater Than or Equal to 0.10, Seasonal Seven-Hour
Average Concentrations, W126, and SUM06 Values for Sites
Experiencing Low Hourly Average Concentrations with Data
Capture Greater Than or Equal to 75% 4-30
4-7 Seasonal (April to October) Percentile Distribution
of Hourly Ozone Concentrations, Number of Hourly
Mean Ozone Occurrences Greater Than or Equal to 0.08
and Greater Than or Equal to 0.10, Seasonal Seven-Hour
Average Concentrations, and W126 Values for Three
"Clean" National Forest Sites with Data Capture
Greater Than or Equal to 75% 4-33
I-xv
-------
List of Tables (cont'd)
Number Page
4-8 The Value of the W126 Sigmoidal Exposure Parameter
Calculated Over the Annual Period 4-34
4-9 The Value of the Ozone Season (Seven-Month) Average
of the Daily Seven-Hour (0900 to 1559 Hours)
Concentration 4-35
4-10 Summary of Percentiles, Number of Hourly Occurrences
Greater Than or Equal to 0.10 ppm, and Three-Month
SUM06 Values for Selected Rural Ozone Monitoring
Sites in 1989 (April to October) 4-37
4-11 Summary of Percentiles of Hourly Average Concentrations
for Electric Power Research Institute Sulfate Regional
Experiment Program Sites/Eastern Regional Air Quality Study
Ozone Monitoring Sites 4-38
4-12 Seven-Hour Growing Season Mean, W126 Values, and
Number of Hourly Ozone Concentrations Greater Than or Equal
to 80 ppb for Selected Eastern National Dry Deposition
Network Sites 4-40
4-13 Summary of Percentiles for National Dry Deposition
Network Monitoring Sites 4-41
4-14 Description of Mountain Cloud Chemistry Program
Sites 4-63
4-15 Seasonal (April to October) Percentiles, SUM06, SUM08,
and W126 Values for the Mountain Cloud Chemistry
Program Sites 4-64
4-16 Summary Statistics for 11 Integrated Forest Study Sites 4-67
4-17 Quarterly Maximum One-Hour Ozone Values at Sites
in and Around New Haven, Connecticut, 1976 4-69
4-18 Summary of Reported Indoor-Outdoor Ozone Ratios 4-74
4-19 Summary of Measurements of Peroxyacetyl Nitrate and
Peroxypropionyl Nitrate in Urban Areas 4-84
l-xvi
-------
List of Tables (cont'd)
Number Page
4-20 Summary of Measurements of Peroxyacetyl Nitrate and
Peroxypropionyl Nitrate in Rural Areas 4-87
l-xvii
-------
List of Figures
Number Page
3-1 The cyclic reactions of tropospheric nitrogen oxides 3-10
3-2 Atmospheric reactions in the complete oxidation
of methane 3-13
3-3 Cyclic reactions of methane oxidation to formaldehyde,
conversion of nitric oxide to nitrogen dioxide, and concomitant
formation of ozone in the atmosphere 3-15
3-4 Major steps in production of ozone in ambient air 3-31
3-5 Time-concentration profiles for selected species during
irradiations of a nitrogen oxide-propene-air mixture in an
indoor chamber with constant light intensity 3-32
3-6 Time-concentration profiles for selected species during
irradiations of a nitrogen oxide-propene-air mixture in an
outdoor chamber with diurnally varying light intensity 3-32
3-7 Surface radiation budget for short- and long-wave radiation 3-43
3-8 The number of reports of ozone concentrations greater
than or equal to 120 ppb at the 17 cities studied in
Samson and Shi (1988) 3-50
3-9 A scatter plot of maximum daily ozone concentration in Atlanta,
Georgia, and New York, New York, versus maximum daily
temperature 3-51
3-10 A scatter plot of maximum daily ozone concentration in Detroit,
Michigan, and Phoenix, Arizona, versus maximum daily
temperature 3-51
3-11 A scatter plot of maximum ozone concentration versus maximum
daily temperature for four nonurban sites 3-52
l-xviii
-------
List of Figures (cont'd)
Number Page
3-12 The frequency of 24-hour trajectory transport distance en route
to city when ozone was greater than or equal to 120 ppb in four
Southern U.S. cities, compared with the percent frequency
distribution for all 17 cities of a nationwide study,
1983 to 1985 3-55
3-13 The frequency of 24-hour trajectory transport distance en route
to city when ozone was greater than or equal to 120 ppb in four
New England cities, compared with the percent frequency
distribution for all 17 cities of a nationwide study,
1983 to 1985 3-55
3-14 The root-mean-square-difference between CLASS
observations and profiler observations as a function of height
above ground level 3-56
3-15 The root-mean-square-difference between CLASS
observations and lidar observations as a function of height
above ground level 3-57
3-16 Model of ozone levels using regression techniques 3-58
3-17 Simulated versus observed ozone levels using regression
techniques on an independent data set obtained in the summer of
1992 in Atlanta, Georgia 3-59
3-18 The 50 largest sources of nitrogen oxides (power plants) in the
United States 3-63
3-19 Nitrogen oxides emissions from manmade sources in the
10 U.S. Environmental Protection Agency regions of the
United States, 1991 3-64
3-20 Changes in nitrogen oxides emissions from manmade sources in
the United States, 10-year intervals, 1940 through 1990 3-65
3-21 Changes in nitrogen oxides emissions from stationary source
fuel combustion and transportation from 1940 through
1990 3-66
l-xix
-------
List of Figures (cont'd)
Number Page
3-22 Changes in emissions of volatile organic compounds from major
manmade sources in the United States, 10-year intervals, 1940
through 1990 3-72
3-23 Changes in emissions of volatile organic compounds from major
manmade sources, 1940 through 1990 3-73
3-24 Estimated biogenic emissions of volatile organic compounds in
the United States as a function of season 3-80
3-25 Example of Empirical Kinetic Modeling Approach diagram
for high-oxidant urban area 3-132
3-26 Regional oxidant model superdomain with modeling domains .... 3-153
4-1 National trend in the composite average of the second highest
maximum one-hour ozone concentration at both National Air
Monitoring Stations and all sites with 95% confidence intervals,
1983 to 1992 4-7
4-2 The annually averaged composite diurnal curves for the following
sites that changed from nonattainment to attainment status:
Montgomery County, Alabama; Concord, California;
Louisville, Kentucky; and Dade County, Florida; for the
period 1987 to 1990 4-10
4-3 A summary of the seasonal (January to December) averaged
composite ozone diurnal curve and integrated exposure W126
index for the Los Angeles, California, site for the period 1980
to 1991 4-11
4-4 United States map of the highest second daily maximum one-hour
average ozone concentration by Metropolitan Statistical Area,
1991 4-16
4-5 The relationship between the second highest daily maximum
hourly average ozone concentration and the maximum three-
month SUM06 value and the second highest daily maximum
eight-hour average ozone concentration and the maximum three-
month SUM06 value for specific site years at rural
agricultural sites for the 1980-to-1991 period 4-24
I-xx
-------
List of Figures (cont'd)
Number Pat
4-6 The relationship between the second highest daily maximum
hourly average ozone concentration and the maximum three-
month SUM06 value and the second highest daily maximum
eight-hour average ozone concentration and the maximum three-
month SUM06 value for specific site years at rural forested
sites for the 1980-to-1991 period 4-25
4-7 The location of National Dry Deposition Network monitoring sites
as of December 1990 4-39
4-8 The kriged 1985 to 1986 maximum seven-hour and 12-hour
average concentrations of ozone across the United States 4-44
4-9 The kriged estimates of the W126 integrated ozone exposure index
for the eastern United States for 1988 and 1989 4-45
4-10 The comparison of the seasonal diurnal patterns using 1988 data
for Jefferson County, Kentucky, and Oliver County, North
Dakota 4-48
4-11 Diurnal behavior of ozone at rural sites in the United States
in July 4-49
4-12 Percent of time hourly average concentrations greater than or equal
to 0.1 ppm occurred between 0900 and 1559 hours in comparison
to 24-hour period for all rural agricultural and forested sites with
three-month SUM06 greater than or equal to 26.4 ppm per hour . . 4-50
4-13 Percent of time hourly average concentrations greater than or equal
to 0.1 ppm occured between 0900 and 1559 hours in comparison
to 24-hour period for all non-California rural agricultural and
forested sites with three-month SUM06 greater than or equal to
26.4 ppm per hour 4-50
4-14 Diurnal pattern of one-hour ozone concentrations on July 13, 1979,
Philadelphia, Pennsylvania 4-51
4-15 Diurnal and one-month composite diurnal variations in ozone
concentrations, Washington, District of Columbia, July 1981 4-52
l-xxi
-------
List of Figures (cont'd)
Number Page
4-16 Diurnal and one-month composite diurnal variations in ozone
concentrations, St. Louis County, Missouri, September 1981 4-53
4-17 Diurnal and one-month composite diurnal variations in ozone
concentrations, Alton, Illinois, October 1981 (fourth quarter) 4-54
4-18 Composite diurnal patterns of ozone concentrations by quarter,
Alton, Illinois, 1981 4-55
4-19 Quarterly composite diurnal patterns of ozone concentrations
at selected sites representing potential for exposure of major
crops, 1981 4-56
4-20 Composite diurnal ozone pattern at a rural National Crop Loss
Assessment Network site in Argonne, Illinois, August 6 through
September 30, 1980 4-57
4-21 Composite diurnal ozone pattern at selected National Dry
Deposition Network sites 4-58
4-22 Composite diurnal pattern at Whiteface Mountain, New York, and
the Mountain Cloud Chemistry Program Shenandoah National
Park site for May to September 1987 4-59
4-23 Seasonal variations in ozone concentrations as indicated by
monthly averages and the one-hour maximum in each month at
selected sites, 1981 4-60
4-24 Seven- and 12-hour means at Whiteface Mountain and
Shenandoah National Park for May to September 1987 and
integrated exposures at Whiteface Mountain and
Shenandoah National Park for May to September 1987 4-65
4-25 Integrated exposures for three non-Mountain Cloud Chemistry
Program Shenandoah National Park sites, 1983 to 1987 4-66
4-26 Number or days in 1991 for which the maximum hourly average
ozone concentration was greater than 0.1 ppm at Chicago,
Illinois 4-70
l-xxii
-------
List of Figures (cont'd)
Number Page
4-27 Maximum one-hour ozone concentrations and average 0800 to
2000 hours strong acid concentrations for each day that
pulmonary function data were collected at Fairview Lake
camp in 1988 4-78
4-28 Maximal one-hour ozone concentrations at Fairview Lake during
the study period 4-79
4-29 The number of occurrences for each of the seven categories
described in text 4-82
4-30 The co-occurrence pattern of ozone and sulfuric acid for
July 25, 1986, at a summer camp on the north shore of Lake Erie,
Ontario, Canada 4-91
4-31 Sulfate, hydrogen ion, and ozone measured at Breadalbane Street
(Site 3) in Toronto during July and August, 1986, 1987, and
1988 4-92
l-xxiii
-------
-------
Authors, Contributors, and Reviewers
Chapter 1. Executive Summary
Principal Authors
Mr. James A. Raub—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William G. 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. Judith A. Graham—National Exposure Research Laboratory (MD-75), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly E. Tilton—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Chapter 2. Introduction
Principal Author
Mr. James A. Raub—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Chapter 3. Tropospheric Ozone and Its Precursors
Principal Authors
Dr. A. Paul Altshuller—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Roger Atkinson—Statewide Air Pollution Research Center, University of California,
900 Watkins Avenue, Riverside, CA 92521
Mr. Michael W. Holdren—Battelle, 505 King Avenue, Columbus, OH 43201
Dr. Thomas J. Kelly—Battelle, 505 King Avenue, Columbus, OH 43201-2693
I-xxv
-------
Authors, 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
Dr. Perry J. Samson—Department of Atmospheric, Oceanic, and Space Sciences, University of
Michigan, 2455 Hay ward Street, Ann Arbor, MI 48109
Dr. John H. Seinfeld—Division of Engineering and Applied Science, California Institute of
Technology, 391 South Holliston Avenue, Pasadena, CA 91125
Dr. Joseph Sickles II—National Exposure Research Laboratory (MD-75), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly E. Tilton—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Halvor (Hal) Westberg—Department of Civil and Environmental Engineering, Washington
State University, Pullman, WA 99164
Reviewers
Dr. A. Paul Altshuller—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Robert R. Arnts—National Exposure Research Laboratory (MD-84) U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. Frank M. Black—National Exposure Research Laboratory (MD-46), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph J. Bufalini—National Exposure Research Laboratory (MD-84), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Daewon Byun—National Exposure Research Laboratory (MD-80), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Jason K. S. Ching—National Exposure Research Laboratory (MD-80), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Kenneth L. Demerjian—Atmospheric Sciences Research Center (SUNY-Albany),
100 Fuller Road, Albany NY 12205
l-xxvi
-------
Authors, Contributors, and Reviewers (cont'd)
Dr. Robin L. Dennis—National Exposure Research Laboratory (MD-80), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Basil Dimitriades—National Exposure Research Laboratory (MD-75), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Marcia C. Dodge—National Exposure Research Laboratory (MD-84), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. Chris D. Geron—National Risk Management Laboratory (MD-62), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Michael W. Gery—Atmospheric Research Associates, 160 North Washington Street,
Boston, MA 02114
Dr. James M. Godowitch—National Exposure Research Laboratory (MD-80),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jimmie W. Hodgeson—National Exposure Research Laboratory (MD-84),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Harvey E. Jeffries—Department of Environmental Sciences and Engineering, School of
Public Health, CB #7400, University of North Carolina, Chapel Hill, North Carolina
27599-7400
Dr. Douglas R. Lawson—Energy and Environmental Engineering Center, Desert Research
Institute, Reno, NV 89506
Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. William A. Lonneman—National Exposure Research Laboratory (MD-84),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William A. McClenny—National Exposure Research Laboratory (MD-44),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Frank F. McElroy—National Exposure Research Laboratory (MD-77), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. Thomas B. McMullen—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
l-xxvii
-------
Authors, Contributors, and Reviewers (cont'd)
Dr. Edwin L. Meyer—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. David C. Misenheimer—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. J. David Mobley—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Will Ollison—American Petroleum Institute, 1220 L Street NW, Washington, DC 20005
Dr. Kenneth Olszyna—Tennessee Valley Authority, CEB 2A, Muscle Shoals, AL 35660
Mr. Thomas E. Pierce—National Exposure Research Laboratory (MD-80), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. Larry J. Purdue—National Exposure Research Laboratory (MD-56), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. Kenneth A. Rehme—National Exposure Research Laboratory (MD-77),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Harold G. Richter—Private consultant, 8601 Little Creek Farm Road, Chapel Hill, NC
27516
Mr. Shawn J. Roselle—National Exposure Research Laboratory (MD-80), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. Kenneth L. Schere—National Exposure Research Laboratory (MD-80),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jack H. Shreffler—National Exposure Research Laboratory (MD-75), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph Sickles II—National Exposure Research Laboratory (MD-75), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. Robert L. Seila—National Exposure Research Laboratory (MD-84), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly E. Tilton—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
l-xxviii
-------
Authors, Contributors, and Reviewers (cont'd)
Dr. Fred Vukovich—Private consultant, 7820 Harps Mill Road, Raleigh, NC 27615
Mr. Richard A. Wayland—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Chapter 4. Environmental Concentrations, Patterns, and
Exposure Estimates
Principal Authors
Dr. Allen S. Lefohn—A.S.L. & Associates, 111 Last Chance Gulch, Suite 4A,
Helena, MT 59601
Dr. A. Paul Altshuller—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Reviewers
Dr. Thomas C. Curran—Office of Air Quality Planning and Standards (MD-12),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Gary F. Evans—National Exposure Research Laboratory (MD-56), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. William G. Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Warren P. Freas—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Jon Heuss—General Motors Environmental and Energy Staff, 3044 West Grand Blvd.,
Detroit, MI 48202
Dr. Nelson Kelly—Environmental Sciences Department, General Motors Research and
Development Center, Warren, MI 48090
Dr. Paul J. Lioy—Department of Environmental and Community Medicine, UMDNJ-Robert
Wood Johnson Medical School, Piscataway, NY 08854
Mr. Thomas R. McCurdy— National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
l-xxix
-------
Authors, Contributors, and Reviewers (cont'd)
Mr. Cornelius J. Nelson—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William Parkhurst—Tennessee Valley Authority, CEB 2A, Muscle Shoals, AL 35660
Mr. Harvey M. Richmond—Office of Air Quality Planning and Standards (MD-12),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
I-XXX
-------
U.S. Environmental Protection Agency
Science Advisory Board
Clean Air Scientific Advisory Committee
Ozone Review
Chairman
Dr. George T. Wolff—General Motors Corporation, Environmental and Energy Staff,
General Motors Bldg., 12th Floor, 3044 West Grand Blvd., Detroit, MI 48202
Members
Dr. Stephen Ayres—Office of International Health Programs, Virginia Commonwealth
University, Medical College of Virginia, Box 980565, Richmond, VA 23298
Dr. Jay S. Jacobson—Boyce Thompson Institute, Tower Road, Cornell University, Ithaca, NY
14853
Dr. Joseph Mauderly—Inhalation Toxicology Research Institute, Lovelace Biomedical and
Environmental Research Institute, P.O. Box 5890, Albuquerque, NM 87185
Dr. Paulette Middleton—Science & Policy Associates, Inc., Western Office, Suite 140,
3445 Penrose Place, Boulder, CO 80301
Dr. James H. Price, Jr.—Research and Technology Section, Texas Natural Resources
Conservation Commission, P.O. Box 13087, Austin, TX 78711
Invited Scientific Advisory Board Members
Dr. Morton Lippmann—Institute of Environmental Medicine, New York University Medical
Center, Long Meadow Road, Tuxedo, NY 10987
Dr. Roger O. McClellan—Chemical Industry Institute of Toxicology, P.O. Box 12137,
Research Triangle Park, NC 27711
Consultants
Dr. Stephen D. Colome—Integrated Environmental Services, University Tower, Suite 280,
4199 Campus Drive, Irvine, CA 92715
l-xxxi
-------
U.S. Environmental Protection Agency
Science Advisory Board
Clean Air Scientific Advisory Committee
(cont'd)
Dr. A. Myrick Freeman—Department of Economics, Bowdoin College, Brunswick, ME 04011
Dr. Allan Legge—Biosphere Solutions, 1601 llth Avenue, NW, Calgary, Alberta T2N 1H1,
CANADA
Dr. William Manning—Department of Plant Pathology, University of Massachusetts, Amherst,
MA 01003
Dr. D. Warner North—Decision Focus, Inc., 650 Castro Street, Suite 300, Mountain View,
CA 94041
Dr. Frank E. Speizer—Harvard Medical School, Channing Lab, 180 Longwood Avenue,
Boston, MA 02115
Dr. George E. Taylor—Department of Environmental and Resource Sciences, 130 Fleischmann
Agriculture Bldg. 199, University of Nevada, Reno, NV 89557
Dr. Mark J. Utell—Pulmonary Disease Unit, Box 692, University of Rochester Medical
Center, 601 Elmwood Avenue, Rochester, NY 14642
Designated Federal Official
Mr. Randall C. Bond—Science Advisory Board (1400), U.S. Environmental Protection
Agency, 401 M Street, SW, Washington, DC 20460
Staff Assistant
Ms. Lori Anne Gross—Science Advisory Board (1400), U.S. Environmental Protection
Agency, 401 M Street, SW, Washington, DC 20460
l-xxxii
-------
U.S. Environmental Protection Agency
Project Team for Development of Air Quality Criteria
for Ozone and Related Photochemical Oxidants
Scientific Staff
Mr. James A. Raub—Health Scientist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. A. Paul Altshuller—Physical Scientist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William G. Ewald—Health Scientist, National Center for Environmental Assessment (MD-
52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. J.H.B. Garner—Ecologist, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Judith A. Graham—Associate Director, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Ellie R. Speh—Secretary, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly E. Tilton—Physical Scientist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Technical Support Staff
Mr. Douglas B. Fennell—Technical Information Specialist, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Mr. Allen G. Hoyt—Technical Editor and Graphic Artist, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Ms. Diane H. Ray—Technical Information Manager (Public Comments), National Center for
Environmental Assessment (MD-52), U.S. Environmental Protection Agency, Research
Triangle Park, NC 27711
l-xxxiii
-------
U.S. Environmental Protection Agency
Project Team for Development of Air Quality Criteria
for Ozone and Related Photochemical Oxidants
(cont'd)
Mr. Richard N. Wilson—Clerk, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Document Production Staff
Ms. Marianne Barrier—Graphic Artist, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Mr. John R. Barton—Document Production Coordinator, ManTech Environmental Technology,
Inc., P.O. Box 12313, Research Triangle Park, NC 27709
Ms. Lynette D. Cradle—Word Processor, ManTech Environmental Technology, Inc., P.O. Box
12313, Research Triangle Park, NC 27709
Ms. Shelia H. Elliott—Word Processor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Ms. Sandra K. Eltz—Word Processor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Ms. Jorja R. Followill—Word Processor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Ms. Sheila R. Lassiter—Word Processor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Ms. Wendy B. Lloyd—Word Processor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Ms. Carolyn T. Perry—Word Processor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Ms. Cheryl B. Thomas—Word Processor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Mr. Peter J. Winz—Technical Editor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
l-xxxiv
-------
U.S. Environmental Protection Agency
Project Team for Development of Air Quality Criteria
for Ozone and Related Photochemical Oxidants
(cont'd)
Technical Reference Staff
Mr. John A. Bennett—Bibliographic Editor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
Ms. S. Blythe Hatcher—Bibliographic Editor, Information Organizers, Inc., P.O. Box 14391,
Research Triangle Park, NC 27709
Ms. Susan L. McDonald—Bibliographic Editor, Information Organizers, Inc., P.O. Box 14391,
Research Triangle Park, NC 27709
Ms. Carol J. Rankin—Bibliographic Editor, Information Organizers, Inc., P.O. Box 14391,
Research Triangle Park, NC 27709
Ms. Deborah L. Staves—Bibliographic Editor, Information Organizers, Inc., P.O. Box 14391,
Research Triangle Park, NC 27709
Ms. Patricia R. Tierney—Bibliographic Editor, ManTech Environmental Technology, Inc.,
P.O. Box 12313, Research Triangle Park, NC 27709
I-xxxv
-------
1
Executive Summary
1.1 Introduction
Air Quality Criteria for Ozone and Related Photochemical Oxidants evaluates the
latest scientific information useful in deriving criteria that form the scientific basis for U.S.
Environmental Protection Agency (EPA) decisions regarding the National Ambient Air
Quality Standards (NAAQS) for ozone (O3). This Executive Summary concisely summarizes
key conclusions from the document, which comprises nine chapters. Following this Executive
Summary is a brief Introduction (Chapter 2) containing information on the legislative and
regulatory background for review of the O3 NAAQS, as well as a brief discussion of the
issues presented and the format for their discussion in the document. Chapter 3 provides
information on the chemistry, sources, emissions, measurement, and transport of O3 and
related photochemical oxidants and their precursors, whereas Chapter 4 covers environmental
concentrations, patterns, and exposure estimates of O3 and oxidants. Chapter 5 deals with
environmental effects, and Chapters 6, 7, and 8 discuss animal toxicological studies, human
health effects, and extrapolation of animal toxicological data to humans, respectively. The
last chapter, Chapter 9, provides an integrative, interpretative characterization of health effects
associated with exposure to O3. The following sections conform to the chapter organization
of the criteria document.
1.2 Legislative and Regulatory Background
The photochemical oxidants found in ambient air in the highest concentrations are
O3 and nitrogen dioxide (NO2). Other oxidants, such as hydrogen peroxide (H2O2) and
peroxyacyl nitrates, also have been observed, but in lower and less certain concentrations. In
1971, EPA promulgated NAAQS to protect the public health and welfare from adverse effects
of photochemical oxidants, at that time, defined on the basis of commercially available
measurement methodology. After 1971, however, O3-specific commercial analytical methods
became available, as did information on the concentrations and effects of the related non-
03 photochemical oxidants. As a result, the chemical designation of the standards was
changed in 1979 from photochemical oxidants to O3.
The EPA is required under Sections 108 and 109 of the Clean Air Act to evaluate
periodically the air quality criteria that reflect the latest scientific information relevant to
review of the O3 NAAQS. These air quality criteria are useful for indicating the kind and
extent of all identifiable effects on public health or welfare that may be expected from the
presence of O3 and related photochemical oxidants in ambient air. The last O3 criteria
1-1
-------
document was released in 1986, and a supplement was released in 1992. These documents
were the basis for a March 1993 decision by EPA that revision of the existing 1-h NAAQS
for O3 was not appropriate at that time. That decision, however, did not take into
consideration more recent scientific information that has been published since the last
literature review in early 1989. The purpose of this revised criteria document, therefore, is to
summarize the pertinent information contained in the previous O3 criteria document and to
critically evaluate and assess the more recent scientific data associated with exposure to
O3 and, to a lesser extent, to H2O2 and the peroxyacyl nitrates, particularly peroxyacetyl
nitrate (PAN). This document will be used by EPA's Office of Air Quality Planning and
Standards to provide a staff paper assessing the most significant scientific information and
presenting staff recommendations on whether revisions to the O3 NAAQS are appropriate.
1.3 Tropospheric Ozone and Its Precursors
Introduction
Ozone is found in the stratosphere, the "free" troposphere, and the planetary
boundary layer (PEL) of the earth's atmosphere. In the PEL, background O3 occurs as the
result of (1) the intrusions of stratospheric O3 into the "free" troposphere and downward
transport into the PEL, and (2) photochemical reactions of methane (CH4), carbon monoxide
(CO), and nitrogen oxides (NOX). These processes contribute to the background O3 near the
surface. The major source of O3 in the PEL is the photochemical process involving
anthropogenic and biogenic emissions of NOX with the many classes of volatile organic
compounds (VOCs).
The topics considered in this section of the document include: tropospheric
O3 chemistry; meteorological influences on O3 formation and transport; precursor VOC and
NOX emissions, ambient concentrations of VOCs and NOX, and source apportionment and
reconciliation of measured VOC ambient concentrations with emission inventories; O3 air
quality models; and analytical methods for oxidants and precursors.
Tropospheric Ozone Chemistry
Ozone occurs in the stratosphere as the result of chemical reactions initiated by
short-wavelength radiation from the sun. In the "free" troposphere, O3 occurs as the result of
incursions from the stratosphere; upward venting from the PEL (the layer next to the surface
of the earth) through certain cloud processes; and photochemical formation from precursors,
notably CH4, CO, and NOX.
The photochemical production of O3 and other oxidants found at the surface of the
earth (in the PEL, troposphere, or ambient air [used interchangeably in this summary]) is the
result of atmospheric physical processes and complex, nonlinear chemical processes involving
two classes of precursor pollutants: (1) reactive anthropogenic and biogenic VOCs and
(2) NOX. The only significant initiator of the photochemical production of O3 in the polluted
troposphere is the photolysis of NO2, yielding nitric oxide (NO) and a ground-state oxygen
atom that reacts with molecular oxygen to form O3. The O3 thus formed reacts with NO,
yielding oxygen and NO2. These cyclic reactions attain equilibrium in the absence of VOCs.
However, in the presence of VOCs, which are abundant in polluted ambient air, the
equilibrium is upset, resulting in a net increase in O3. Methane is the chief VOC found in the
free troposphere and in most "clean" areas of the PEL. The VOCs found in polluted ambient
1-2
-------
air are much more complex and more reactive than CH4, but, as with CH4, their atmospheric
oxidative degradation is initiated through attack on the VOCs by hydroxyl (OH) radicals. As
in the CH4 oxidation cycle, the conversion of NO to NO2 during the oxidation of VOCs is
accompanied by the production of O3 and the efficient regeneration of the OH radical. The
O3, PAN, and higher homologues formed in polluted atmospheres increase with the NO2/NO
concentration ratio.
At night, in the absence of photolysis of reactants, the simultaneous presence of
O3 and NO2 results in the formation of the nitrate (NO3) radical. Reactions with NO3 radicals
appear to constitute major sinks for alkenes, cresols, and several other compounds, although
the chemistry is not well characterized.
Most inorganic gas-phase processes (i.e., the nitrogen cycle and its
interrelationships with O3 production) are well understood. The chemistry of the VOCs in
ambient air is not as well understood. It is well known, however, that the chemical loss
processes of gas-phase VOCs include reaction with OH and NO3 radicals and O3, and
photolysis. Reaction with the OH radical is the only important atmospheric reaction (loss
process) for alkanes, aromatic hydrocarbons, and the higher aldehydes and ketones that lack
>C=C< bonds; and the only atmospheric reaction of alcohols and ethers. Photolysis is the
major loss process for formaldehyde and acetone. Reactions with OH and NO3 radicals and
with O3 are all important loss processes for alkenes and for carbonyls containing >C=C<
bonds.
Uncertainties in the atmospheric chemistry of the VOCs can affect quantification
of the NO-to-NO2 conversion and of O3 yields, and can present difficulties in representation
of chemical mechanisms, products, and product yields in O3 air quality models. Major
uncertainties in understanding the atmospheric chemistry of the VOCs with NOX in both urban
and rural atmospheres include chemistry of alkyl nitrate formation, mechanisms and products
of >C4 w-alkanes and branched alkanes, mechanisms and products of alkene-O3 reactions, and
mechanisms and products of aromatic hydrocarbons.
It should be noted that the atmospheric chemical processes involved in the
photooxidation of certain higher molecular weight VOCs and in the formation of O3 also can
lead to the formation of particulate-phase organic compounds. The OH radicals produced not
only can oxidize VOCs to particulate-phase organic compounds but also can react with NO2
and sulfur dioxide (SO2) to form nitric acid (HNO3) and sulfuric acid (H2SO4), respectively,
portions of which become incorporated into aerosols as particulate nitrate and sulfate.
Meteorological Influences on Ozone Formation and Transport
The surface energy (radiation) budget of the earth strongly influences the dynamics
of the PEL. The redistribution of energy through the PEL creates thermodynamic conditions
that influence vertical mixing. Growing evidence indicates that the strict use of mixing
heights in modeling is an oversimplification of the complex processes by which pollutants are
redistributed within urban areas, and that it is necessary to treat the turbulent structure of the
atmosphere directly and acknowledge the vertical variations in mixing. Energy balances
therefore require study so that more realistic simulations can be made of the structure of the
PEL.
Day-to-day variability in O3 concentrations depends heavily on day-to-day
variations in meteorological conditions, including temperature, solar radiation, and the degree
of mixing that occurs between release of a pollutant or its precursors and their arrival at a
1-3
-------
receptor; the occurrence of inversion layers (layers in which temperature increases with height
above ground level); and the transport of O3 left overnight in layers aloft and subsequent
downward mixing of that O3 to the surface.
The transport of O3 and its precursors beyond the urban scale (<50 km) to
neighboring rural and urban areas has been well documented. Episodes of high
O3 concentrations in urban areas are often associated with high concentrations of O3 in the
surroundings. Areas of O3 accumulation usually are characterized by synoptic-scale
subsidence of air in the free troposphere, resulting in development of an elevated inversion
layer; relatively low wind speeds associated with the weak horizontal pressure gradient around
a surface high pressure system; a lack of cloudiness; and high temperatures.
Ultraviolet (UV) radiation from the sun plays a key role in initiating the
photochemical processes leading to O3 formation and affects individual photolytic reaction
steps. Still, there is little empirical evidence in the literature linking day-to-day variations in
observed UV radiation levels to variations in O3 levels. An association, however, between
tropospheric O3 concentrations and temperature has been demonstrated. Empirical data from
four urban areas, for example, show an apparent upper bound on O3 concentrations that
increases with temperature. A similar qualitative relationship exists at a number of rural
locations.
The relationship between wind speed and O3 buildup varies from one part of the
country to another.
Statistical techniques (e.g., regression techniques) can be used to help identify real
trends in O3 concentrations, both intra- and interannual, by normalizing meteorological
variability.
Precursors
Volatile Organic Compound Emissions
Hundreds of VOCs, usually containing from 2 to 12 carbon atoms, are emitted by
evaporative and combustion processes from a large number of source types. Total U.S.
anthropogenic VOC emissions in 1991 were estimated at 21.0 Tg; the two largest source
categories were (1) industrial processes (10.0 Tg) and (2) transportation (7.9 Tg). Emissions
of VOCs from highway vehicles accounted for almost 75% of the transportation-related
emissions; studies have shown that the majority of these VOC emissions come from about
20% of the automobiles in service, many, perhaps most, of which are older cars that are
poorly maintained. The accuracy of VOC emission estimates is difficult to determine for
both stationary and mobile sources.
Vegetation emits significant quantities of VOCs into the atmosphere, chiefly
monoterpenes and isoprene, but also oxygenated VOCs, according to recent studies. The
most recent biogenic VOC emissions estimate for the United States showed annual emissions
of 29.1 Tg/year.
Although the biogenic VOC emission estimates exceed the anthropogenic
estimates, the biogenic emissions are more diffusely distributed than the anthropogenic
emissions, which tend to be concentrated in population centers. However, the large
uncertainties in both biogenic and anthropogenic VOC emission inventories prevent
establishing the relative contributions of these two categories.
1-4
-------
Nitrogen Oxides Emissions
Anthropogenic NOX is associated with combustion processes. The primary
pollutant emitted is NO, formed at high combustion temperatures from nitrogen and oxygen
in the air and from nitrogen in the combustion fuel. Emissions of NOX in 1991 in the United
States totaled 21.39 Tg. The two largest single NOX emission sources are electric power
generating plants and highway vehicles. Because a large proportion of anthropogenic NOX
emissions come from distinct point sources, published annual estimates are thought to be
much more reliable than VOC estimates.
Natural NOX sources include stratospheric intrusion, oceans, lightning, soil, and
wildfires. Lightning and soil emissions are the only two significant natural sources of NOX in
the United States. It is estimated that combined natural sources contribute about 2.2 Tg of
NOX to the troposphere over the continental United States; however, uncertainties in natural
NOX emission inventories are much greater than those for anthropogenic NOX emissions.
Concentrations of Volatile Organic Compounds in Ambient Air
The VOCs most frequently analyzed in ambient air are the nonmethane
hydrocarbons (NMHCs). Morning (6:00 to 9:00 a.m.) concentrations most often have been
measured because of the use of morning data in the Empirical Kinetic Modeling Approach
(EKMA) and in air quality simulation models.
Concurrent measurements of anthropogenic and biogenic NMHCs have shown that
biogenic NMHCs usually constituted much less than 10% of the total NMHCs. For example,
average isoprene concentrations ranged from 0.001 to 0.020 ppm carbon (C) and terpenes
from 0.001 to 0.030 ppm C.
Concentrations of Nitrogen Oxides in Ambient Air
Measurements of NOX made in 22 and 19 U.S. cities in 1984 and 1985,
respectively, showed median 6:00-to-9:00 a.m. NOX concentrations ranging from 0.02 to 0.08
ppm in most of these cities. Nonurban NOX concentrations, reported as average seasonal or
annual NOX, range from <0.005 to 0.015 ppm.
Ratios of Concentrations of Nonmethane Organic Compounds to Nitrogen Oxides
Ratios of 6:00-to-9:00 a.m. nonmethane organic compounds (NMOC) to NOX are
higher in southeastern and southwestern U.S. cities than in northeastern and midwestern U.S.
cities, according to data from EPA's multicity studies conducted in 1984 and 1985. Rural
NMOC/NOX ratios tend to be higher than urban ratios. The NMOC/NOX ratios trended
downward to well below 10 in the South Coast Air Basin and in cities in the eastern United
States during the 1980s. Based on these low ratios, hydrocarbon control should be more
effective than NOX control within a number of cities. Morning (6:00-to-9:00 a.m.)
NMOC/NOX ratios are used in the EKMA type of trajectory model. The correlation of
NMOC/NOX ratios with maximum 1-h O3 concentrations, however, was weak in a recent
analysis.
Source Apportionment and Reconciliation
Source apportionment (regarded as synonymous with receptor modeling) refers to
determining the quantitative contributions of various sources of VOCs to ambient air pollutant
concentrations. Source reconciliation refers to the comparison of measured ambient VOC
1-5
-------
concentrations with emissions inventory estimates of VOC source emission rates for the
purpose of validating the inventories.
Recent findings have shown that vehicle exhaust was the dominant contributor to
ambient VOCs in seven of eight U.S. cities studied. Whole gasoline contributions were
estimated to be equal to vehicle exhaust in one study and to 20% of vehicle exhaust in a
second study.
Estimates of biogenic VOCs at a downtown site in Atlanta, GA, in 1990 indicated
a lower limit of 2% (24-h average) for the biogenic percentage of total ambient VOCs at that
location (isoprene was used as the biogenic indicator species). The percentage varies during
the 24-h period because of the diurnal (e.g., temperature, light intensity) dependence of
isoprene concentrations.
Source reconciliation data have shown disparities between emission inventory
estimates and receptor-estimated contributions. For biogenics, emission estimates are greater
than receptor-estimated contributions. The reverse has been true for natural gas contributions
estimated for Los Angeles, CA; Columbus, OH; and Atlanta; and for refinery emissions in
Chicago, IL.
Ozone Air Quality Models
Models and Their Components
Photochemical air quality models are used to predict how O3 concentrations change
in response to prescribed changes in source emissions of NOX and VOCs. These models
operate on sets of input data that characterize the emissions, topography, and meteorology of
a region and produce outputs that describe air quality in that region.
Two kinds of photochemical models are recommended in guidelines issued by
EPA: (1) the use of EKMA is accepted under certain circumstances, and (2) the grid-based
Urban Airshed Model (UAM) is recommended for modeling O3 over urban areas. The 1990
Clean Air Act Amendments mandate the use of three-dimensional (grid-based) air quality
models such as UAM in developing state implementation plans for areas designated as
"extreme", "severe", "serious", or "multistate moderate". General descriptions of EKMA and
grid-based models were given in the 1986 EPA criteria document for O3.
The EKMA-based method for determining O3 control strategies has limitations, the
most serious of which is that predicted emissions reductions are critically dependent on the
initial NMHC/NOX ratio used in the calculations. This ratio cannot be determined with any
certainty and is expected to be quite variable in time and space in an urban area.
Spatial and temporal characteristics of VOC and NOX emissions are major inputs to
a grid-based photochemical air quality model. Greater accuracy in emissions inventories is
needed for biogenics and for both mobile and stationary source components. Grid-based air
quality models also require as input the three-dimensional wind field for the photochemical
episode being simulated.
A chemical kinetic mechanism, representing the important chemical reactions that
occur in the atmosphere, is used in an air quality model to estimate the net rate of formation
of each pollutant simulated as a function of time.
Dry deposition is an important removal process for O3 on both urban and regional
scales and is included in all urban- and regional-scale models. Wet deposition is generally
not included in urban-scale photochemical models, because O3 episodes do not occur during
periods of significant clouds or rain.
1-6
-------
Concentration fields of all species computed by the model must be specified at the
beginning of the simulation ("initial conditions"). These initial conditions are determined
mainly with ambient measurements, either from routinely collected data or from special
studies; but interpolation can be used to distribute the surface ambient measurements.
Use of Ozone Air Quality Models
Photochemical air quality models are used for control strategy evaluation by first
demonstrating that a past episode or episodes can be simulated adequately. The hydrocarbon
or NOX emissions or both are reduced in the model inputs, and the effects of these reductions
on O3 in the region are assessed. The adequacy of control strategies based on grid-based
models depends, in part, on the nature of input data for simulations and model validation, on
input emissions inventory data, and on the mismatch between the spacial output of the model
and the current form of the NAAQS for O3. Uncertainties in models obviously can affect
their outputs. Uncertainties exist in all components of grid-based O3 air quality models:
emissions, meteorological modules, chemical mechanisms, deposition rates, and determination
of initial conditions.
Grid-based models that have been widely used to evaluate control strategies for
O3 or acid deposition, or both, are the UAM, the California Institute of Technology/Carnegie
Institute of Technology model, the Regional Oxidant Model, the Acid Deposition and Oxidant
Model, and the Regional Acid Deposition Model. The UAM (Version IV) is the grid model
approved nationwide for control strategy development at this time.
Despite the many uncertainties in photochemical air quality modeling, including
emission inventories, these models are essential for regulatory analysis and solving the
O3 problem. Grid-based O3 air quality modeling is superior to the available alternatives for
O3 control planning, but the chances of its incorrect use must be minimized.
Analytical Methods for Oxidants and Their Precursors
Oxidants
Current methods used to measure O3 are chemiluminescence (CL); UV absorption
spectrometry; and newly developed spectroscopic and chemical approaches, including
chemical approaches applied to passive sampling devices (PSDs) for O3.
The CL method has been designated as the reference method by EPA. Detection
limits of 0.005 ppm and a response time of <30 s are typical of currently available
commercial instruments. A positive interference from atmospheric water vapor was reported
in the 1970s and recently has been confirmed. Proper calibration can minimize this source of
error.
Commercial UV photometers for measuring O3 have detection limits of about
0.005 ppm and a response time of <1 min. Because the measurement is absolute, UV
photometry is also used to calibrate O3 methods. A potential disadvantage of UV photometry
is that atmospheric constituents that absorb 254-nm radiation, the wavelength at which O3 is
measured, will cause a positive interference in O3 measurements. Interferences have been
reported in two recent studies, but assessment of the potential importance of such
interferences (e.g., toluene, styrene, cresols, nitrocresols) is hindered by lack of absorption
spectra data in the 250-nm range and by lack of aerometric data for the potentially interfering
species. There also can be some interference from water, possibly from the condensation of
moisture in sampling lines.
1-7
-------
Calibration of O3 measurement methods (other than PSDs) is done by UV
spectrometry or by gas-phase titration (GPT) of O3 with NO. Ultraviolet photometry is the
reference calibration method approved by EPA. Ozone is unstable and must be generated in
situ at time of use to produce calibration mixtures.
Peroxyacetyl nitrate and the higher peroxyacyl nitrates normally are measured by
gas chromatography (GC) using an electron capture detector. Detection limits have been
extended to 1 to 5 ppt. The preparation of reliable calibration standards is difficult because
PAN is unstable, but several methods are available.
Volatile Organic Compounds
The method recommended by EPA for total NMOC measurement involves the
cryogenic preconcentration of NMOCs and the measurement of the revolatilized NMOCs
using flame ionization detection (FID). The primary technique for speciated NMOC/NMHC
measurements is cryogenic preconcentration followed by GC-FID. Systems for sampling and
analysis of VOCs have been developed that require no liquid cryogen for operation.
Stainless steel canisters have become the containers of choice for collection of
whole-air samples for NMHC/NMOC data. Calibration procedures for NMOC
instrumentation require the generation, by static or dynamic systems, of dilute mixtures at
concentrations expected to occur in ambient air.
Preferred methods for measuring carbonyl species (aldehydes and ketones) in
ambient air are spectroscopic methods; on-line colorimetric methods; and, the most common
method currently in use for measuring gas-phase carbonyl compounds in ambient air, the
high-performance liquid chromatography method, which employs 2,4-dinitrophenylhydrazine
derivatization in a silica gel cartridge. Use of an O3 scrubber has been recommended to
prevent interference by O3 in this method in ambient air.
Oxides of Nitrogen
Nitric oxide and NO2 comprise the NOX compounds involved as precursors to
O3 and other photochemical oxidants.
The most common method of NO measurement is the gas-phase CL reaction with
O3, which is essentially specific for NO. Commercial NO monitors have detection limits of a
few parts per billion by volume (ppbv) in ambient air but may not have sensitivity sufficient
for surface measurements in rural or remote areas or for airborne measurements. Direct
spectroscopic methods for NO exist that have very high sensitivity and selectivity for NO, but
their complexity, size, and cost restrict these methods to research applications. No PSDs exist
for measurement of NO.
Chemiluminescence analyzers are the tools of choice for NO2 measurement, even
though they do not measure NO2 directly. Minimum detection levels for NO2 have been
reported to be 5 to 13 ppb, but more recent evaluations have indicated detection limits of
0.5 to 1 ppbv. Reduction of NO2 to NO is required for measurement. These analyzers
actually measure NOy (NOX + PAN + FINO3 + other reactive nitrogen species); however, for
most urban atmospheres, NOX is the predominant species measured diurnally.
Several spectroscopic approaches to NO2 detection have been developed but share
the drawbacks of spectroscopic NO methods. Passive samplers for NO2 exist but are still in
the developmental stage for ambient air monitoring.
1-8
-------
Calibration of methods for NO measurement is done using standard cylinders of
NO in nitrogen. Calibration of methods for NO2 measurement include use of cylinders of
NO2 in nitrogen or air, use of permeation tubes, and GPT.
1.4 Environmental Concentrations, Patterns, and
Exposure Estimates
Ozone is measured at concentrations above the minimum detectable level at all
monitoring locations in the world. In this section, hourly average concentration and exposure
information is summarized for urban, rural forested, and rural agricultural areas in the United
States.
Because O3 from urban area emissions is transported to rural downwind locations,
elevated O3 concentrations can occur at considerable distances from urban centers. Urban
O3 concentration values are often depressed because of titration by NO. Because of the
absence of chemical scavenging, O3 tends to persist longer in nonurban areas than in urban
areas, and nonurban exposures may be higher than those in urban locations.
Trends
Ozone hourly average concentrations have been recorded for many years by the
state and local air pollution agencies who report their data to EPA. The 10-year (1983 to
1992) composite average trend for the second highest daily maximum hourly average
concentration during the O3 season shows that the 1992 composite average for the trend sites
was 21% lower than the 1983 average. The 1992 value was the lowest composite average of
the 10-year period and was significantly less than each of the previous nine years, 1983 to
1991. The relatively high O3 concentrations in 1983 and 1988 likely were attributable, in
part, to hot, dry, stagnant conditions in some areas of the country, which were especially
conducive to O3 formation.
From 1991 to 1992, the composite mean of the second highest daily maximum
1-h O3 concentrations decreased 7%, and the composite average of the number of estimated
exceedances of the O3 standard decreased by 23%. Nationwide VOC emissions decreased 3%
from 1991 to 1992. The composite average of the second daily maximum concentrations
decreased in 8 of the 10 EPA regions from 1991 to 1992, and remained unchanged in Region
VII. Except for Region VII, the 1992 regional composite means were lower than the
corresponding 1990 levels. Although meteorological conditions in the east during 1993 were
more conducive to O3 than those in 1992, the composite mean level for 1993 was the second
lowest composite average of the decade, 1984 to 1993.
Surface Concentrations
Published data provide evidence showing the occurrence at some sites of multihour
periods within a day of O3 at levels of potential health effects. Although most of these
analyses were made using monitoring data collected from sites in or near nonattainment areas,
in one analysis of five sites (two in New York state, two in rural California, and one in rural
Oklahoma), none of which was in or near a nonattainment area, O3 concentrations showed
only moderate peaks but showed multihour levels above 0.1 ppm.
1-9
-------
A small amount of the O3 concentration measured at a monitoring site is produced
by sources distant to the photochemical reactions occurring on an urban or regional scale.
Typical sources include stratospheric intrusions into the troposphere, photochemical
production by the CH4/CO/NOX cycle in the troposphere, and transport of very distant
anthropogenic or biogenic VOCs and NOX. The specific concentrations of this "background"
O3 vary with averaging times ranging from the daily 1-h maximum to daily, monthly,
seasonal, or annual values. The background concentrations also vary with geographical
region and with elevation of the monitoring site.
On the basis of O3 data from isolated monitoring sites, EPA has indicated that a
reasonable estimate of O3 background concentration near sea level in the United States is
from 0.020 to 0.035 ppm for an annual average, 0.025 to 0.045 ppm for an 8-h daily summer
seasonal average, and from 0.03 to 0.05 ppm for the average summertime 1-h daily
maximum. This estimate includes a 0.005 to 0.015 ppm O3 contribution from stratospheric
intrusions into the troposphere.
Diurnal Variations
Diurnal patterns of O3 may be expected to vary with location, depending on the
balance among the many factors affecting O3 formation, transport, and destruction. Although
they vary with locality, diurnal patterns of O3 typically show a rise in concentration from low
levels, or levels near minimum detectable amounts, to an early afternoon peak. The diurnal
pattern of concentrations can be ascribed to three simultaneous processes: (1) downward
transport of O3 from layers aloft, (2) destruction of O3 through contact with surfaces and
through reaction with NO at ground level, and (3) in situ photochemical production of O3.
Seasonal Patterns
Seasonal variations in O3 concentrations in urban areas usually show the pattern of
high O3 in the late spring or in the summer and low levels in the winter; however, weather
conditions in a given year may be more favorable for the formation of O3 and other oxidants
than during the prior or following year.
Average O3 concentrations tend to be higher in the second versus the third quarter
of the year for many isolated rural sites. This observation has been attributed to either
stratospheric intrusions or an increasing frequency of slow-moving, high-pressure systems that
promote the formation of O3. However, for several clean rural sites, the highest exposures
have occurred in the third quarter rather than in the second. For rural O3 sites in the
southeastern United States, the daily maximum 1-h average concentration was found to peak
during the summer months.
Spatial Variations
Concentrations of O3 vary with altitude and with latitude. There appears to be no
consistent conclusion concerning the relationship between O3 exposure and elevation.
Indoor Ozone
Until the early 1970s, very little was known about the O3 concentrations
experienced inside buildings; to date, the database on this subject is not large, and a wide
range of indoor/outdoor O3 concentration relationships can be found in the literature (reported
indoor/outdoor values for O3 are highly variable). Indoor/outdoor O3 concentration ratios
1-10
-------
generally fall in the range from 0.1 to 0.7 and indoor concentrations of O3 almost invariably
will be less than outdoors.
Estimating Exposure
Both fixed-site monitoring information and human exposure models are used to
estimate risks associated with O3 exposure. Because, for most cases, it is not possible to
estimate population exposure solely from fixed-station data, several human exposure models
have been developed. These models also contain submodels depicting the sources and
concentrations likely to be found in each microenvironment, including indoor, outdoor, and
in-transit settings. Two distinct types of O3 exposure models exist: (1) those that focus
narrowly on predicting indoor O3 levels and (2) those that focus on predicting O3 exposures
on a community-wide basis. These latter models and their distinguishing features are:
1. pNEM/O3 based on the National Air Quality Standards Exposure Model (NEM)
series of models
Uses mass-balance approach and seasonal considerations for I/O ratio
estimation.
Variables affecting indoor exposure obtained by Monte Carlo sampling
from empirical distributions of measured data.
2. Systems Applications International (SAI)/NEM
• More districts and microenvironments and more detailed mass-balance
model than pNEM/O3.
Human activity data outdated and inflexible.
3. Regional Human Exposure Model (REHEX)
More detailed geographic resolution than NEM.
Uses California-specific activity data and emphasizes in-transit and outdoor
microenvironments.
4. Event probability exposure model (EPEM)
• Estimates probability that a randomly selected person will experience a
particular exposure regime.
Lacks multiday continuity.
Few data are available for individuals using personal exposure monitors. Results
from a pilot study demonstrated that fixed-site ambient measurements may not adequately
represent individual exposures. Models based on time-weighted indoor and outdoor
concentrations explained only 40% of the variability in personal exposures.
Peroxyacyl Nitrates
Peroxyacetyl nitrate and peroxypropionyl nitrate (PPN) are the most abundant of
the non-O3 oxidants in ambient air in the United States, other than the inorganic nitrogenous
oxidants such as NO2, and possibly HNO3. Most of the available data on concentrations of
PAN and PPN in ambient air are from urban areas. The levels to be found in nonurban areas
will be highly dependent on the transport of PAN and PPN or their precursors from urban
areas, because the concentrations of the NOX precursors to these compounds are considerably
lower in nonurban areas than in urban areas.
1-11
-------
Co-occurrence
Studies of the joint occurrence of gaseous NO2/O3 and SO2/O3 at rural sites have
concluded that the periods of co-occurrence represent a small portion of the potential
plant-growing period. For human ambient exposure considerations, in most cases, the
simultaneous co-occurrence of NO2/O3 and SO2/O3 was infrequent. Some researchers have
reported the joint occurrence of O3, nitrogen, and sulfur in forested areas, combining
cumulative exposures of O3 with data on dry deposition of sulfur and nitrogen. One study
reported that several forest landscapes with the highest dry deposition loadings of sulfur and
nitrogen tended to experience the highest average O3 concentrations and largest cumulative
exposure. Although the authors concluded that the joint concentrations of multiple pollutants
in forest landscapes were important, nothing was mentioned about the hourly co-occurrences
of O3 and SO2 or O3 and NO2. Acid sulfates, which are usually composed of H2SO4,
ammonium bisulfate, and ammonium sulfate, have been measured at a number of locations in
North America. The potential for O3 and acidic sulfate aerosols to co-occur at some locations
in some form (i.e., simultaneously, sequentially, or complex-sequentially) is real and requires
further characterization. For human ambient exposures, the simultaneous co-occurrence of
NO2 and O3 was infrequent.
In one study, the relationship between O3 and hydrogen ions in precipitation was
explored using data from sites that monitored both O3 and wet deposition simultaneously and
within one minute latitude and longitude of each other. It was reported that individual sites
experienced years in which both hydrogen ion deposition and total O3 exposure were at least
moderately high. With data compiled from all sites, it was found that relatively acidic
precipitation occurred together with relatively high O3 levels approximately 20% of the time,
and highly acidic precipitation occurred together with a high O3 level approximately 6% of
the time. Sites most subject to relatively high levels of both hydrogen ions and O3 were
located in the eastern part of the United States, often in mountainous areas.
The co-occurrence of O3 and acidic cloudwater in high-elevation forests has been
characterized. The frequent O3-only and pH-only single-pollutant episodes, as well as the
simultaneous and sequential co-occurrences of O3 and acidic cloudwater, have been reported.
Both simultaneous and sequential co-occurrences were observed a few times each month
above cloud base.
1.5 Environmental Effects of Ozone and Related
Photochemical Oxidants
Ozone is the gaseous pollutant most injurious to agricultural crops, trees, and
native vegetation. Exposure of vegetation to O3 can inhibit photosynthesis, alter carbon
(carbohydrate) allocation, and interfere with mycorrhizal formation in tree roots. Disruption of
the important physiological processes of photosynthesis and carbon allocation can suppress
the growth of crops, trees, shrubs, and herbaceous vegetation by decreasing their capacity to
form the carbon (energy) compounds needed for growth and maintenance and their ability to
absorb the water and mineral nutrients that they require from the soil. In addition, loss of
vigor impairs the ability of trees and crops to reproduce and increases their susceptibility to
insects and pathogens. The following section summarizes key environmental effects
associated with O3 exposure.
1-12
-------
Effects on Agroecosystems
Methodologies Used in Vegetation Research
Most of the knowledge concerning the effects of O3 on vegetation comes from the
exposure-response studies of important agricultural crop plants and some selected forest and
urban tree species, mostly as seedlings. A variety of methodologies have been used, ranging
from field exposures without chambers to open-top chambers and to exposures conducted in
chambers under highly controlled conditions. In general, the more controlled conditions are
most appropriate for investigating specific responses and for providing the scientific basis for
interpreting and extrapolating results. The greatest body of knowledge is from OTC studies.
Mode of Action
Leaves are important regulators of plant stress and function. Stress resulting from
exposure to O3 produces a leaf-mediated response. Effects expressed within cells in the leaf
(i.e., inhibition of photosynthesis) affect a plant's carbon (energy) budget. Plant processes are
impaired only by the O3 that enters the plant through the stomata (opening in the leaves). An
effect will occur only if sufficient O3 reaches sensitive sites within the leaf cells. The uptake
and movement of O3 to sensitive cellular sites within a leaf are subject to various biochemical
and physiological controls. Leaf injury will not be detected if the rate of uptake is small
enough for the plant to detoxify or metabolize O3 and its derivatives, or the plant is able to
repair or compensate for the impact at a rate equal to or greater than the rate of uptake.
Impairment of leaf cellular processes is the basis for all other plant effects. The diurnal
pattern of stomatal opening plays a critical role in O3 uptake, particularly at the canopy level.
Visible injury is usually the first observable indication of cellular response; injury
can occur, however, with no visible effects. Early senescence of leaves or needles is also a
result of cellular response. Impairment of cellular processes inhibits the rate of
photosynthesis, reduces carbon (sugars, carbohydrate) production, and alters carbon allocation,
causing a shift in growth pattern that favors shoots over roots. The reduced allocation of
carbon to leaf repair and new leaf formation limits the availability of carbon for reproduction;
stem and root growth; and, particularly, the formation of the mycorrhizae on roots necessary
for nutrient and water uptake. Reduction of plant vigor by O3 can result in mortality,
particularly when plant susceptibility to insects and pathogens is increased.
Factors That Modify Plant Response
Plant response to O3 exposure is influenced by a variety of biological, chemical,
and physical factors. When determining the impact of O3 exposure on plants, both the
influence of environmental factors on plant response and the effects of O3 on that response
must be considered. Biological factors within plants that affect their response to stresses
include, genetic composition, stage of development, and the diurnal pattern of stomatal
opening. Genotype significantly influences plant sensitivity to O3. Individuals, varieties, and
cultivars of a species are known to differ greatly in their responses to a given O3 exposure.
Genotype also influences the ability of plants to compete with one another for space,
nutrients, light, and water.
The magnitude of response of a particular species, variety, or cultivar depends on a
number of environmental factors. The plant's present and past environmental milieu, which
includes the temporal exposure pattern and stage of development, dictates the plant response.
The corollary is also true: exposure to O3 can modify plant response to other environmental
1-13
-------
variables. Available light, temperature, atmospheric turbulence and moisture, in both the
atmosphere and soil; soil nutrition; and exposure to and interaction with other pollutants such
as agricultural chemical sprays also influence the magnitude of plant response.
Drought can reduce visible injury and the adverse effects of O3 on growth and
yield of crops. However, in the case of crops, drought, per se, much more adversely affects
yield than the effects of O3. Ozone, on the other hand, tends to reduce the water-use
efficiency of well-watered crops. In some plants, O3 exposure reduces cold/winter hardiness.
Although exposure to O3 tends to reduce attacks by obligate pathogens, susceptibility of
plants to facultative pests and pathogens increases.
Effects-Based Air Quality Exposure Indices
Environmental scientists for many years have attempted to characterize and
mathematically represent plant exposures to O3. A variety of averaging times have been used.
Although most studies have characterized exposure by using mean concentrations, such as
seasonal, monthly, weekly, daily, or peak hourly means, other studies have used cumulative
measures (e.g., the number of hours above selected concentrations). None of these statistics
completely characterizes the relationships among O3 concentration, exposure duration, interval
between exposures, and plant response.
The use of a mean concentration with long averaging times implies that all
concentrations of O3 are equally effective in causing plant responses and minimizes the
contributions of the peak concentrations to the response. Ozone effects are cumulative;
therefore, exposure duration should be included in any index if it is to be biologically
relevant. Present evidence suggests that cumulative effects of episodic exposures to either
peak or mid-range concentrations, or both, can play an important role in producing growth
responses. The key to plant response is timing because peak and mid-range concentrations do
not occur at the same time. Potentially, the greatest effect of O3 on plants will occur when
stomatal conductance is greatest. When peaks occur at the time of greatest stomatal
conductance, the effect of mid-range concentrations will not be observable. Atmospheric
conductivity also strongly influences plant response because O3 must be in contact with the
leaf surface if it is to be taken up by a plant. Effects on vegetation appear when the amount
of pollutant entering exceeds the ability of the plant to repair or compensate for the impact.
Increasing uptake of O3 will inhibit photosynthesis and result in increased reductions in
biomass production.
An index of ambient exposures that relates well to plant response should
incorporate, directly or indirectly, environmental influences (e.g., temperature, humidity, soil-
moisture status) and exposure dynamics. Peak indices (e.g., second highest daily maximum)
imply that a single high-concentration exposure (1- or 8-h concentration) during the course of
a 70- to 120-day growing season is related to eventual yield or growth reductions. On the
other hand, mean indices (e.g., 7-h seasonal mean) imply that duration of the exposure is not
important, and that all concentrations have equal effect on plants. Neither of these indices
relates ambient O3 concentrations to biological effects on plants because these indices do not
consider the duration of exposure. An index that cumulates all hourly concentration during
the season and gives greater weight to higher concentrations appears to be a more appropriate
index for relating ambient exposures to growth or yield effects.
No experimental studies have been designed specifically to evaluate the adequacy
of the various peak-weighted indices that have been proposed. In retrospective analyses in
which O3 is the primary source of variation in response, year-to-year variations in plant
1-14
-------
response are minimized by peak-weighted, cumulative exposure indices. However, a number
of different forms of peak-weighted, cumulative indices have been examined for their ability
to properly order yield responses from the large number of studies of the National Crop Loss
Assessment Network (NCLAN) program. These exposure indices (i.e., SUMOO, SUM06,
SIGMOID, W126) all performed equally well, and it is not possible to distinguish among
them on the basis of statistical fits of the data. The biological basis for these indices has not
been determined.
Exposure Response of Plant Species
The emphasis of experimental studies usually has been on the more economically
important crop plants and tree species, as seedlings. Crop species usually are monocultures
that are fertilized and, in many cases, watered. Therefore, because crop plants are usually
grown under optimal conditions, their sensitivity to O3 exposures can vary from that of native
trees, shrubs, and herbaceous vegetation.
The concept of limiting values was used in both the 1978 and 1986 criteria
documents to summarize visible foliar injury. Limiting values are defined as concentrations
and durations of exposure below which visible injury does not occur. The limit for visible
injury indicating reduced plant performance was an O3 exposure of 0.05 ppm for several
hours per day for more than 16 days. When the exposure period was decreased to 10 days,
the O3 concentration required to cause injury was increased to 0.10 ppm. A short, 6-day
exposure further increased the concentration to 0.30 ppm. These exposure and concentration
periods apply for those crops where appearance or aesthetic value (e.g, spinach, cabbage,
lettuce) is considered important. Limiting values for foliar injury to trees and shrubs range
from 0.06 to 0.10 ppm for 4 h.
The following assertions can be made based on information from the 1986 criteria
document, its 1992 supplement, and literature published since 1986. Ambient
O3 concentrations in several regions of the country are high enough to impair growth and
yield of sensitive plant species. This clearly is indicated by comparison of data obtained from
crop yield in charcoal-filtered and unfiltered (ambient) exposures. These elevated levels are
further supported by data from studies using chemical protectants. These response data make
possible the extrapolation to plants not studied experimentally. Both approaches mentioned
above indicate that effects occur with only a few exposures above 0.08 ppm. Data from
regression studies conducted to develop an exposure-response function for estimating yield
loss indicated that at least 50% of the species and cultivars tested could be predicted to
exhibit a 10% yield loss at 7-h seasonal mean O3 concentrations of 0.05 ppm or less.
Effects on Natural Ecosystems
The responses of the San Bernardino mixed forest of Southern California to 50 or
more years of chronic ozone exposures based on many studies, present a classic example of
ecosystem response to severe stress. Data from an inventory conducted from 1968 through
1972 indicated that for 5 mo of each year, trees were exposed to O3 concentrations greater
than 0.08 ppm for more than 1,300 h. Concentrations rarely decreased below 0.05 ppm at
night near the crest of the mountain slope, approximately 5,500 ft. In addition, during the
years 1973 to 1978, average 24-h O3 concentrations ranged from a background of 0.03 to 0.04
ppm in the eastern part of the San Bernardino Mountains to a maximum of 0.10 to 0.12 ppm
in the western part during May through September.
1-15
-------
Plants accumulate, store, and use the energy in carbon compounds (sugars)
produced during photosynthesis to build their structures and to maintain the physiological
processes necessary for life. The patterns of carbon allocation to roots, stems, and leaves
directly influence growth. The strategy for carbon allocation changes during the life of a
plant, as well as with environmental conditions. Mature trees have a higher ratio of
respiration to photosynthetic tissue. Impairment of photosynthesis shifts carbon allocation
from growth and maintenance to repair; increased respiration can result in resource
imbalances. The significant changes observed in the San Bernardino forest ecosystem were a
possible outcome of the combined influences of O3 on carbon, water, and nutrient allocation.
The biochemical changes within the leaves of ponderosa and Jeffrey pine in the
San Bernardino forest, expressed as visible foliar injury, premature needle senescence,
reduced photosynthesis, and reduced carbohydrate production and allocation, resulted in
reduced tree vigor, growth, and reproduction. Reduced vigor increases susceptibility of trees
to insect pests and fungal pathogens. Premature needle senescence alters microorganismal
succession on confer needles and changes the detritus-forming process and associated nutrient
cycling.
Altered carbon allocation is important in the formation of mycorrhizae (fungus
roots), which are an extremely important but unheralded component of all ecosystems; the
majority of all plants depend on them because they are integral in the uptake of mineral
nutrients and water from the soil. Carbon-containing exudates from the roots are necessary
for the formation of mycorrhizae. Reduced carbon allocation to plant roots affects
mycorrhizal formation and impacts plant growth. Exposure to ozone, therefore, affects plant
growth both above and below ground.
Small changes in photosynthesis or carbon allocation can alter profoundly the
structure of a forest. Ecosystem responses to stress begin with the response of the most
sensitive individuals of a population. Stresses, whose primary effects occur at the molecular
level (within the leaves), must be propagated progressively through more integrated levels of
organ physiology (e.g., leaf, branch, root) to whole plant physiology, then to populations
within the stand (community), and finally to the landscape level to produce ecosystem effects.
Only a small fraction of stresses at the molecular level become disturbances at the tree, stand,
or landscape level. The time required for a stress to be propagated from one level to the next
(it can take years) determines how soon the effects of the stress can be observed or measured.
The primary effect of O3 on ponderosa and Jeffrey pine, two of the more
susceptible members of the San Bernardino forest community, was that the trees were no
longer able to compete effectively for essential nutrients, water, light, and space. Decline in
the sensitive trees, a consequence of altered competitive conditions, permitted the enhanced
growth of more tolerant species. Removal of the ecosystem dominants at the population level
changed its structure and altered the processes of energy flow and nutrient cycling, returning
the ecosystem to a less complex stage.
The San Bernardino Mountains continue to experience exposure to O3; however,
there has been a gradual decline in concentrations and length of exposure. Ozone
concentrations of 0.06 ppm or higher of varying durations capable of causing injury to trees
in forest ecosystems have been observed during the past 5 years in the Sierra Nevada
Mountains and the Appalachian Mountains from Georgia to Maine. Visible injury to forest
trees and other vegetation in these areas has been observed.
1-16
-------
Injury to sensitive trees from exposure to ozone concentrations 0.06 ppm or greater
in the Sierra Nevada Mountains and the Appalachian Mountains has never had the impact on
these ecosystems that it did on the San Bernardino forest. Forest stands differ greatly in age,
species composition, stability, and capacity to recover from disturbance. In addition, the
position in the stand or community of the most sensitive species is extremely important.
Ponderosa and Jeffrey pine were the dominant species in the San Bernardino forest. Removal
of populations of these trees altered both ecosystem structure and function. Both the Sierra
Nevada Mountains and the Appalachian Mountains are biologically more diverse. Removal
of sensitive individuals of eastern white pine and black cherry has not visibly altered the
forest ecosystems along the Appalachian Mountains, possibly because of the absence of
population changes in these species. Decline and dieback of trees on Mt. Mitchell, NC, and
Camel's Hump, VT, cannot be related solely to O3 injury.
Effects on Agriculture, Forestry, and Ecosystems: Economics
A number of economic assessments of the effects of O3 on agriculture have been
performed over the last decade. All use NCLAN response data to predict crop yield changes.
Although these studies employ somewhat different economic assessment methodologies, each
shows national-level economic losses to major crops in excess of $1 billion (1990 dollars)
from exposure to ambient concentrations of O3. These studies also evaluate the sensitivity of
the economic estimates to uncertainties in data, including the NCLAN response data. The
economic assessment models used could be adapted to future O3-crop yield response findings,
if available.
The plant science literature shows that O3 adversely influences physiological
performance of both urban and native tree species; the limited economic literature also
demonstrates that changes in growth have economic consequences. However, the natural
science and economic literature on the topic are not yet mature enough to conclude
unambiguously that ambient O3 is imposing economic costs. The economic effects of O3 on
ecosystems have not yet been addressed in the published literature. There is, however, an
emerging interest in applying economic concepts and methods to the management of
ecosystems.
Effects on Materials
Over four decades of research show that O3 damages certain materials such as
elastomers, textile fibers, and dyes. The amount of damage to actual in-use materials and the
economic consequences of that damage are poorly characterized.
Natural rubber and synthetic polymers of butadiene, isoprene, and styrene, used in
products like automobile tires and protective outdoor electrical coverings, account for most of
the elastomer production in the United States. The action of O3 on these compounds is well
known, and concentration-response relationships have been established and corroborated by
several studies. These relationships, however, must be correlated with adequate exposure
information based on product use. For these and other economically important materials,
protective measures have been formulated to reduce the rate of oxidative damage. When
antioxidants and other protective measures are incorporated in elastomer production, the
O3-induced damage is reduced considerably, although the extent of reduction differs widely
according to the material and the type and amount of protective measures used.
1-17
-------
Both the type of dye and the material in which it is incorporated are important
factors in the resistance of a fabric to O3. Some dyed fabrics, such as royal blue and red
rayon-acetate and plum cotton are resistant to O3. On the other hand, anthraquinone dyes on
nylon fibers are sensitive to fading by O3. Field studies and laboratory work show a positive
association between O3 levels and dye fading of nylon materials. At present, the available
research is insufficient to quantify the amount of damaged materials attributable to O3 alone.
The degradation of fibers from exposure to O3 is poorly characterized. In general,
most synthetic fibers, such as modacrylic and polyester, are relatively resistant, whereas
cotton, nylon, and acrylic fibers have greater but varying sensitivities to O3. Ozone reduces
the breaking strength of these fibers, and the degree of strength reduction depends on the
amount of moisture present. The limited research in this area indicates that O3 in ambient air
may have a minimal effect on textile fibers, but additional research is needed to verify this
conclusion.
A number of artists' pigments and dyes are sensitive to O3 and other oxidants; in
particular, many organic pigments are subject to fading or other color changes when exposed
to O3. Although most, but not all, modern fine arts paints are more O3 resistant, many older
works of art are at risk of permanent damage due to O3-induced fading.
A great deal of work remains to be done to develop quantitative estimates of the
economic damage to materials from photochemical oxidants. Most of the available studies
are outdated in terms of O3 concentrations, technologies, and supply-demand relationships.
Additionally, little is known about the physical damage functions, so cost estimates have been
simplified to the point of not properly recognizing many of the scientific complexities of the
impact of O3.
1.6 lexicological Effects of Ozone and Related
Photochemical Oxidants
Respiratory Tract Effects of Ozone
Biochemical Effects
Knowledge of molecular targets provides a basis for understanding mechanisms of
effects and strengthening animal-to-human extrapolations. Ozone reacts with polyunsaturated
fatty acids and sulfhydryl, amino, and some electron-rich compounds. These elements are
shared across species. Several types of reactions are involved, and free radicals may be
created. Based on this knowledge, it has been hypothesized that the O3 molecule is unlikely
to penetrate the liquid linings of the respiratory tract (RT) to reach the tissue, raising the
possibility that reaction products exert effects.
In acute and short-term exposure studies, a variety of lung lipid changes occur,
including an increase in arachidonic acid, the further metabolism of which produces a variety
of biologically active mediators that can affect host defenses, lung function, the immune
system, and other functions.
The level of lung antioxidant metabolism increases after O3 exposure, probably as
a result of the increase in the number of Type 2 cells, which are rich in antioxidant enzymes.
Collagen (the structural protein involved in fibrosis) increases in O3-exposed lungs
in a manner that has been correlated to structural changes (e.g., increased thickness of the
tissue between the air and blood after prolonged exposure). Some studies found that the
increased collagen persists after exposure ceases.
1-18
-------
Generally, O3 enhances lung xenobiotic metabolism after both short- and long-term
exposure, possibly as a result of morphological changes (increased numbers of nonciliated
bronchiolar epithelial cells). The impact of this change is dependent on the xenobiotics
involved; for example, the metabolism of benzo[a]pyrene to active metabolites was enhanced
by03.
Lung Inflammation and Permeability Changes
Elevated concentrations of O3 disrupt the barrier function of the lung, resulting in
the entry of compounds from the airspaces into the blood and the entry of serum components
(e.g., protein) and white blood cells (especially polymorphonuclear leukocytes [PMNs]) into
the airspaces and lung tissue. This latter impact reflects the initial stage of inflammation.
These cells can release biologically active mediators that are capable of a number of actions,
including damage to other cells in the lung. In lung tissue, this inflammation also can
increase the thickness of the air-blood barrier.
Increases in permeability and inflammation have been observed at levels as low as
0.1 ppm O3 (2 h/day, 6 days; rabbits). After acute exposures, the influence of the time of
exposure (from two to several hours) increases as the concentration of O3 increases.
Long-term exposure effects are discussed under lung morphology.
The impacts of these changes are not fully understood. At higher
O3 concentrations (e.g., 0.7 ppm, 28 days), the diffusion of oxygen into the blood decreases,
possibly because the air-blood barrier is thicker; cellular death may result from the enzymes
released by the inflammatory cells; and host defense functions may be altered by mediators.
Effects on Host Defense Mechanisms
Exposure to elevated concentrations of ozone results in alterations of all defense
mechanisms of the RT, including mucociliary and alveolobronchiolar clearance, functional
and biochemical activity of the alveolar macrophage (AM), and immunologic competence.
These effects can cause susceptibility to bacterial respiratory infections.
Mucociliary clearance, which removes particles and cellular debris from the
conducting airways, is slowed by acute, but not repeated exposures to O3. Ciliated epithelial
cells that move the mucous blanket are altered or destroyed by acute and chronic exposures.
Neonatal sheep exposed to O3 do not have normal development of the mucociliary system.
Such effects could prolong the retention of unwanted substances (e.g., inhaled particles) in the
lungs, allowing them to exert their toxicity for a longer period of time.
Alveolar clearance mechanisms, which center on the functioning of AMs, are
altered by O3. Short-term exposure to levels as low as 0.1 ppm O3 (2 h/day, 1 to 4 days;
rabbits) accelerates clearance, but longer exposures do not. Even so, after a 6-week exposure
of rats to an urban pattern of O3, the retention of asbestos fibers in a region protected by
alveolar clearance is prolonged.
Alveolar macrophages engulf and kill microbes, as well as clear the deeper regions
of the lungs of nonviable particles; AMs also participate in immunological responses, but
little is known about the effects of O3 on this function. Acute exposures of rabbits to levels
as low as 0.1 ppm O3 decrease the ability of AMs to ingest particles. This effect is displayed
in decreases in the ability of the lung to kill bacteria after acute exposure of mice to levels as
low as 0.4 ppm O3.
1-19
-------
Both the pulmonary and systemic immune system are affected by O3, but in a
poorly understood way. It appears that the part of the immune system dependent on T-cell
function is more affected than is the part dependent on B-cell function.
Dysfunction of host defense systems results in enhanced susceptibility to bacterial
lung infections. For example, acute exposure to O3 concentrations as low as 0.08 ppm for
3 h can overcome the ability of mice to resist infection with streptococcal bacteria, resulting
in mortality. However, more prolonged exposures (weeks, months) do not cause greater
effects on infectivity.
Effects on antiviral defenses are more complex and less well understood. Only
high concentrations (1.0 ppm O3, 3 h/day, 5 days; mice) increase viral-induced mortality.
Apparently, O3 does not impact antiviral clearance mechanisms. Although O3 does not affect
acute lung injury from influenza virus infection, it does enhance later phases of the course of
an infection (i.e., postinfluenzal alveolitis).
Morphological Effects
Elevated concentrations of O3 cause similar types of alterations in lung structure in
all laboratory animal species studied, from rats to monkeys. In the lungs, the most affected
cells are the ciliated epithelial cells of the airways and Type 1 epithelial cells of the
gas-exchange region. In the nasal cavity, ciliated cells are also affected.
The centriacinar region (CAR; the junction of the conducting airways and gas-
exchange regions) is the primary target, possibly because this area receives the greatest dose
of O3. The ciliated cells can be killed and replaced by nonciliated cells (i.e., cells not capable
of clearance functions that also have increased ability to metabolize some foreign
compounds). Mucous-secreting cells are affected, but to a lesser degree. Type 1 cells, across
which gas exchange occurs, can be killed; they are replaced by Type 2 cells, which are
thicker and produce more lipids. An inflammatory response also occurs in the tissue. The
tissue is thickened further in later stages when collagen (a structural protein increased in
fibrosis) and other elements accumulate. Although fibrotic changes have been observed in the
CAR, they have not been distributed throughout the whole lung.
The distal airway is remodeled; more specifically, bronchiolar epithelium replaces
the cells present in alveolar ducts. Concurrent inflammation may play a role. This effect has
been observed at 0.25 ppm O3 (8 h/day, 18 mo) in monkeys; at a higher concentration, this
remodeling persists after exposure stops.
The progression of effects during and after a chronic exposure is complex. Over
the first few days of exposure, inflammation peaks and then drops considerably, plateauing
for the remainder of exposure, after which it largely disappears. Epithelial hyperplasia
increases rapidly over the first few days and rises slowly or plateaus thereafter; when
exposure ends, it begins to return toward normal. In contrast, fibrotic changes in the tissue
between the air and blood increase very slowly over months of exposure, and, after exposure
ceases, the changes sometimes persist or increase.
The pattern of exposure can make a major difference in effects. Monkeys exposed
to 0.25 ppm O3 (8 h/day) every other month of an 18-mo period had equivalent changes in
lung structure, more fibrotic changes, and more of certain types of pulmonary function
changes than did monkeys exposed every day over the 18 mo. From this work and rat
studies, it appears that natural seasonal patterns may be of more concern than more
continuous exposures. Thus, long-term animal studies with uninterrupted exposures may
underestimate some of the effects of O3.
1-20
-------
The morphologic lesions described in early publications on laboratory animals
exposed to O3 do not meet the current criteria for emphysema of the type seen in human
lungs.
Effects on Pulmonary Function
Pulmonary function changes in animals resemble those observed in humans after
acute exposure.
During acute exposure, the most commonly observed alterations are increased
frequency of breathing and decreased tidal volume (i.e., rapid, shallow breathing). This has
been reported at exposures as low as 0.2 ppm O3 for 3 h (rats). Typically, higher
concentrations (around 1 ppm) are required to affect breathing mechanics (compliance and
resistance). Extended characterizations of pulmonary function show types of changes
generally seen in humans. For example, there are decreased lung volumes at levels >0.5 ppm
O3 (a few hours; rats).
When rats are exposed to O3 for 2 h/day for 5 days, the pattern of attenuation of
pulmonary function responses is similar to that observed in humans. Other biochemical
indicators of lung injury did not return to control values by Day 5, and morphological
changes increased in severity over the period of exposure. Thus, attenuation did not result in
protection against all the effects of O3.
Long-term exposures have provided mixed results on pulmonary function,
including no or minimal effects, restrictive effects, and obstructive effects. When changes
occurred and postexposure examinations were performed, pulmonary function recovered.
Genotoxicity and Carcinogenicity of Ozone
The chemical reactivities of O3 give it the potential to be a genotoxic agent.
In vitro studies are difficult to interpret because the culture systems used allowed
the potential formation of artifacts, and high or very high concentrations of O3 often were
used. Generally, in these studies, O3 causes DNA strand breaks, sometimes is weakly
mutagenic, and causes cellular transformation and chromosomal breakage. The latter finding
has been investigated in vivo, with mixed results in animals.
The few earlier long-term carcinogenic studies in laboratory animals, with or
without coexposure to known carcinogens, are either negative or ambiguous.
The National Toxicology Program (NTP) completed chronic rat and mouse cancer
bioassays using commonly accepted experimental approaches and designs. Both male and
female rats and mice were studied. Animals were exposed for 2 years (6 h/day, 5 days/week)
to 0.12, 0.5, and 1.0 ppm O3 or for a lifetime to the same levels (except 0.12 ppm).
Following their standard procedures for determination of weight-of-evidence for
carcinogenicity, the NTP reported "no evidence" in rats, "equivocal evidence" in male mice,
and "some evidence" in female mice. The increases in adenomas and carcinomas were
observed only in the lungs. There was no concentration response. One of the reasons for the
designation of "some evidence" in female mice was that when the 2-year and lifetime
exposure studies were combined, there was a statistically significant increase in total tumors
at 1.0 ppm. Lung tumors from control and O3-exposed mice also were examined for the
presence of mutated Ha-ras oncogenes. Although the types of mutations found were similar
in both groups, a higher incidence of mutations was found in lung tumors from the
O3-exposed mice. At the present time, however, there is inadequate information to provide
1-21
-------
mechanistic support for the finding in mice. Thus, the potential for animal carcinogenicity
is uncertain.
In a companion NTP study, male rats were treated with a tobacco carcinogen and
exposed for 2 years to 0.5 ppm O3. Ozone did not affect the response and therefore had no
tumor promoting activity.
Systemic Effects of Ozone
Ozone causes a variety of effects on tissues and organs distant from the lung.
Because O3 itself is not thought to penetrate the lung, these systemic effects are either
secondary to lung alterations or result from reaction products of O3. Effects have been
observed on clinical chemistry, white blood cells, red blood cells, the circulatory system, the
liver, endocrine organs, and the central nervous system. Most of these effects cannot be
interpreted adequately at this time and have not been investigated in humans, but it is of
interest to note that O3 exposures causing effects on the RT of animals cause a wide array of
effects on other organs also.
Several behavior changes occur in response to O3. For example, 0.12 ppm O3 (6 h,
rats) decreases wheel-running activity, and 0.5 ppm (1 min) causes mice to avoid exposure.
These effects are not fully understood, but they may be related to lung irritation or decreased
ability to exercise.
Although cardiovascular effects, such as slowed heart rate and decreased blood
pressure, occur in O3-exposed rats, some observed interactions with thermoregulation prevent
qualitative extrapolation of these effects to humans at this time.
Developmental toxicity studies in pregnant rats summarized in the 1986 O3 criteria
document showed that levels up to about 2.0 ppm O3 did not cause birth defects. Rat pups
from females exposed to 1.0 ppm O3 during certain periods of gestation weighed less or had
delays in development of behaviors (e.g., righting, eye opening). No "classical" reproductive
assays with O3 were found.
Other studies have indicated that O3 can affect some endocrine organs (i.e.,
pituitary-thyroid-adrenal axis, parathyroid gland). It appears that the liver has less ability to
detoxify drugs after O3 exposure, but assays of liver enzymes involved in xenobiotic
metabolism are inconsistent.
Interactions of Ozone with Other Co-occurring Pollutants
Animal studies of the effects of O3 in combination with other air pollutants show
that antagonism, additivity, and synergism can result, depending on the animal species,
exposure regimen, and health endpoint. Thus, these studies clearly demonstrate the major
complexities and potential importance of interactions but do not provide a scientific basis for
predicting the results of interactions under untested ambient exposure scenarios.
1.7 Human Health Effects of Ozone and Related
Photochemical Oxidants
This section summarizes key effects associated with exposure to O3, the major
component of photochemical oxidant air pollution that is clearly of most concern to the health
of the human population. Another, often co-occurring photochemical oxidant component of
1-22
-------
"smog" is PAN, but this compound has been demonstrated to be primarily responsible for
induction of smog-related eye irritation (stinging of eyes). Limited pulmonary function
studies have shown no effects of PAN at concentrations below 0.13 to 0.30 ppm, which are
much higher than the generally encountered ambient air levels in most cities.
Controlled Human Studies of Acute Ozone Effects
Effects on Lung Function
Controlled studies in healthy adult subjects have demonstrated O3-induced
decrements in pulmonary function, characterized by alterations in lung volumes and flow and
airway resistance and responsiveness. Respiratory symptoms, such as cough and pain on deep
inspiration, are associated with these changes in lung function.
Ozone-induced decreases in lung volume, specifically forced vital capacity (FVC)
and forced expiratory volume in 1 s (FEVj), largely can be attributed to decreases in
inspiratory capacity (the ability to take a deep breath), although at higher exposure
concentrations, there is clearly an additional component that is not volume dependent. Lung
volumes recover to a large extent within 2 to 6 h; normal baseline function typically is
reestablished within 24 h, but not fully with more severe exposures.
Ozone causes increased airway resistance and may cause reductions in expiratory
flow and the FEVj/FVC ratio.
Ozone causes an increase in airway responsiveness to nonallergenic stimuli (e.g.,
histamine, methacholine) in healthy and asthmatic subjects. There is no clear evidence of a
relationship between O3-induced lung volume changes and changes in airway responsiveness.
Inflammation and Host Defense Effects
Controlled studies in healthy adult subjects also indicate that O3 causes an
inflammatory response in the lungs characterized by elevated levels of PMNs, increased
epithelial permeability, and elevated levels of biologically active substances (e.g.,
prostaglandins, proinflammatory mediators, cytokines).
Inflammatory responses to O3 can be detected within 1 h after a single 1-h
exposure with exercise to concentrations >0.3 ppm; the increased levels of some inflammatory
cells and mediators persist for at least 18 h. The temporal response profile is not defined
adequately, although it is clear that the time course of response varies for different mediators
and cells.
Lung function and respiratory symptom responses to O3 do not seem to be
correlated with airway inflammation.
Ozone also causes inflammatory responses in the nose, marked by increased
numbers of PMNs and protein levels suggestive of increased permeability.
Alveolar macrophages removed from the lungs of human subjects after 6.6 h of
exposure to 0.08 and 0.10 ppm O3 have a decreased ability to ingest microorganisms,
indicating some impairment of host defense capability.
Ozone Exposure-Response Relationships
Functional, symptomatic, and inflammatory responses to O3 increase with
increasing exposure dose of O3. The major determinants of the exposure dose are
O3 concentration (C), exposure duration (T), and the amount of ventilation (VE).
1-23
-------
Exercise increases response to O3 by increasing VE (greater mass delivered), tidal
volume, inspiratory flow (greater percentage delivery), and the intrapulmonary
O3 concentration.
Repeated daily exposures to relatively high levels of O3 doses (C x T x VE)
causing substantial reductions in FEVj (>20% decrement) typically cause exacerbation of the
lung function and respiratory symptom responses on the second exposure day. However,
attenuation of these responses occurs with continued exposures for a few days. Most
inflammatory responses also attenuate; for example, the PMN influx is absent after five
consecutive exposures.
Multihour exposures (e.g., for up to 7 h) to O3 concentrations as low as 0.08 ppm
cause small but statistically significant decrements in lung function, increases in respiratory
symptoms, and increases in PMNs and protein levels. Ozone C is a more important factor
than exercise VE or T in predicting responses to multihour low-level O3 exposure. There is
clear evidence of a response plateau in terms of lung volume response to prolonged
O3 exposure. This evidence suggests that for a given combination of exercise and
O3 concentration (i.e., dose rate), there is a response plateau; continued exposure (i.e,
increased T) at that dose rate will not increase response. Therefore, quantitative extrapolation
of responses to longer exposure durations is not valid.
Mechanisms of Acute Pulmonary Responses
The mechanisms leading to the observed pulmonary responses induced by O3 are
beginning to be better understood. The available descriptive data suggest a number of
mechanisms leading to the alterations in lung function and respiratory symptoms, including
O3 delivery to the tissue (i.e., the inhaled concentration, breathing pattern, airway geometry;
O3 reactions with the airway lining fluid and epithelial cell membranes; local tissue responses,
including injury and inflammation; and stimulation of neural afferents (bronchial C-fibers) and
the resulting reflex responses and symptoms. The cyclooxygenase inhibitors block production
of prostaglandin E2 and interleukin-6 as well as reduce lung volume responses; however, these
drugs do not reduce inflammation and levels of cell damage markers such as lactate
dehydrogenase.
Effects on Exercise Performance
Maximal oxygen uptake, a measure of peak exercise performance capacity, is
reduced in healthy young adults if preceded by O3 exposures sufficient to cause marked
changes in lung function (i.e., decreases of at least 20%) and increased subjective symptoms
of respiratory discomfort. Limitations in exercise performance may be related to increased
symptoms, especially those related to breathing discomfort.
Factors Modifying Responsiveness to Ozone
Many variables have the potential for influencing responsiveness to O3; however,
most are addressed inadequately in the available clinical data to make definitive conclusions.
Active smokers are less responsive to O3 exposure, which may reverse following
smoking cessation, but these results should be interpreted with caution.
The possibility of age-related differences in response to O3 has been explored,
although young adults historically have provided the subject population for controlled human
studies. Children and adolescents have lung volume responses to O3 similar to those of
young adults, but lack respiratory symptoms. Pulmonary function responsiveness in adults
1-24
-------
appears to decrease with age, whereas symptom rates remain similar to young adults. Group
mean lung function responses of adults over 50 years of age are less than those of children,
adolescents, and young adults.
The available data have not demonstrated conclusively that men and women
respond differently to O3. Likewise, pulmonary function responses of women have been
compared during different phases of the menstrual cycle, but the results are conflicting.
If gender differences exist for lung function responsiveness to O3, they are not based on
hormonal changes, differences in lung volume, or the ratio of FVC to VE.
There is no compelling evidence, to date, suggesting that any ethnic or racial
groups have a different distribution of responsiveness to O3.
Seasonal and ambient factors may vary responsiveness to O3, but further research
is needed to determine how they affect individual subjects. Individual sensitivity to O3 may
vary throughout the year, related to seasonal variations in ambient O3 concentrations.
The specific inhalation route appears to be of minor importance in exercising
adults. Exposure to O3 by oral breathing (i.e., mouthpiece) yields results similar to exposure
by oronasal breathing (i.e., chamber exposures).
Population Groups at Risk from Ozone Exposure
Population groups that have demonstrated increased responsiveness to ambient
concentrations of O3 consist of exercising healthy and asthmatic individuals, including
children, adolescents, and adults.
Available evidence from controlled human studies on subjects with preexisting
disease suggests that mild asthmatics have similar lung volume responses, but greater airway
resistance responses to O3 than nonasthmatics; and that moderate asthmatics may have, in
addition, greater lung volume responses than nonasthmatics.
Of all the other population groups studied, those with preexisting limitations in
pulmonary function and exercise capacity (e.g., chronic obstructive pulmonary disease,
chronic bronchitis, ischemic heart disease) would be of primary concern in evaluating the
health effects of O3. Unfortunately, limitations of subject selection, standardized methods of
subject characterization, and range of exposure hamper the ability to make definitive
conclusions regarding the relative responsiveness of most chronic disease subjects.
Effects of Ozone Mixed with Other Pollutants
No significant enhancement of respiratory effects has been demonstrated
consistently for simultaneous exposures of O3 mixed with SO2, NO2, H2SO4, HNO3,
particulate aerosols, or combinations of these pollutants. It is fairly well established that
simultaneous exposure of healthy adults and asthmatics to mixtures of O3 and other pollutants
for short periods of time (<2 h) induces pulmonary function responses not significantly
different from those following O3 alone when studies are conducted at the same
O3 concentration. Exposure to PAN has been reported to induce greater pulmonary function
responses than exposure to O3 alone, but at PAN concentrations (>0.27 ppm) much higher
than ambient levels. Unfortunately, only a limited number of pollutant combinations and
exposure protocols have been investigated, and subject groups are small and are representative
of only small portions of the general population. Thus, much is unknown about the
relationships between O3 and the complex mix of pollutants found in the ambient air.
1-25
-------
Prior exposure to O3 in asthmatics may cause an increase in response to other
pollutant gases, especially SO2. Likewise, prior exposure to other pollutants can enhance
responses to O3 exposure.
Controlled Human Studies of Ambient Air Exposures
Mobile laboratory studies of lung function and respiratory symptoms in a local
subject population exposed to ambient photochemical oxidant pollution provide quantitative
information on exposure-response relationships for O3. A series of these studies from
Los Angeles has demonstrated pulmonary function decrements at mean ambient
O3 concentrations of 0.14 ppm in exercising healthy adolescents and increased respiratory
symptoms and pulmonary function decrements at 0.15 ppm in heavily exercising athletes and
at 0.17 ppm in lightly exercising healthy and asthmatic subjects. Comparison of the observed
effects in exercising athletes with controlled chamber studies at comparable O3 concentrations
showed no significant differences in lung function and symptoms, suggesting that coexisting
ambient pollutants have a minimal contribution to the measured responses under typical
summer ambient conditions in Southern California.
Field and Epidemiology Studies of Ambient Air Exposures
Individual-level field studies and aggregate-level time-series studies have addressed
the acute effects of O3 on lung function decrements and increased morbidity and mortality in
human populations exposed to real-world conditions of O3 exposure.
Camp and exercise studies of lung function provide quantitative information on
exposure-response relationships linking lung function declines with O3 exposure occurring in
ambient air. Combined statistical analysis of six recent camp studies in children yields an
average relationship between decrements in FEVj and previous-hour O3 concentration of
-0.50 mL/ppb. Two key studies of lung function measurements before and after well-defined
outdoor exercise events in adults have yielded exposure-response slopes of -0.40 and
-1.35 mL/ppb. The magnitude of pulmonary function declines with O3 exposure is consistent
with the results of controlled human studies.
Daily life studies support a consistent relationship between O3 exposure and acute
respiratory morbidity in the population. Respiratory symptoms (or exacerbation of asthma)
and decrements in peak expiratory flow rate are associated with increasing ambient O3,
particularly in asthmatic children; however, concurrent temperature, particles, acidity
(hydrogen ions), aeroallergens, and asthma severity or medication status also may contribute
as independent or modifying factors. Aggregate results show greater responses in asthmatic
individuals than in nonasthmatics, indicating that asthmatics constitute a sensitive group in
epidemiologic studies of oxidant air pollution.
Summertime daily hospital admissions for respiratory causes in various locations of
eastern North America have consistently shown a relationship with ambient levels of O3,
accounting for approximately one to three excess respiratory hospital admissions per hundred
parts per billion O3 per million persons. This association has been shown to remain even
after statistically controlling for the possible confounding effects of temperature and
copollutants (e.g., hydrogen ions, sulfate, and particles less than 10 jim), as well as when
considering only concentrations below 0.12 ppm O3.
Many of the time-series epidemiology studies looking for associations between
O3 exposure and daily human mortality have been difficult to interpret because of
1-26
-------
methodological or statistical weaknesses, including the a failure to account for other pollutant
and environmental effects. One of the two most useful new studies on O3-mortality found a
small but statistically significant association in Los Angeles when peak 1-h maximum
O3 concentrations reached concentrations greater than 0.2 ppm during the study period.
A second study in regions with lower (<0.15 ppm) maximum 1-h O3 concentrations
(St. Louis, MO, and Kingston-Harriman, TN) did not detect a significant O3 association with
mortality.
Only suggestive epidemiologic evidence exists for health effects of chronic
ambient O3 exposure in the population. All of the available studies of chronic respiratory
system effects in exposed children and adults are limited by a simplistic assignment of
exposure or by their inability to isolate potential effects related to O3 from those of other
pollutants, especially particles.
1.8 Extrapolation of Animal lexicological Data to
Humans
There have been significant advances in O3 dosimetry since 1986 that better enable
quantitative extrapolation with marked reductions in uncertainty. Experiments and models
describing the uptake efficiency and delivered dose of O3 in the RT of animals and humans
are beginning to present a clearer picture than has existed previously.
The total RT uptake efficiency of rats at rest is approximately 50%. Within the
RT of the rat, 50% of the O3 taken up by the RT is removed in the head, 7% in the
larynx/trachea, and 43% in the lungs.
In humans at rest, the total RT uptake efficiency is between 80 and 95%. Total
RT uptake efficiency falls as flow increases. As tidal volume increases, uptake efficiency
increases and flow dependence lessens. Pulmonary function response data and O3 uptake
efficiency data in humans generally indicate that the mode of breathing (oral versus nasal
versus oronasal) has little effect on upper RT or on total RT uptake efficiency, although one
study suggests that the nose has a higher uptake efficiency than the mouth.
When all of the animal and human in vivo O3 uptake efficiency data are compared,
there is a good degree of consistency across data sets. This agreement raises the level of
confidence with which these data sets can be used to support dosimetric model formulations.
Several mathematical dosimetry models have been developed since 1986.
Generally, the models predict that net O3 dose to lung lining fluid plus tissue gradually
decreases distally from the trachea toward the end of the tracheobronchial region and then
rapidly decreases in the pulmonary region.
When the dose of O3 to lung tissue is computed theoretically, it is found to be very
low in the trachea; to increase to a maximum in the terminal bronchioles of the first
generation pulmonary region; and then to decrease rapidly, moving further into the pulmonary
region. The increased tidal volume and flow, associated with exercise in humans, shifts
O3 dose further into the periphery of the lung and causes a disproportionate increase in distal
lung dose.
Predictions of delivered dose have been used to investigate both acute and chronic
O3 responses in the context of intra- and interspecies comparisons. In the case of intraspecies
comparisons, for example, the distribution of predicted O3 tissue dose to a ventilatory unit in
a rat as a function of distance from the bronchoalveolar duct junction is very consistent with
1-27
-------
the distribution of alveolar wall thickening. In the case of interspecies comparisons (using the
delivered O3 dose to the proximal alveolar regions), although the functional responses (e.g.,
rapid, shallow breathing) differ markedly between rats and humans, there is similarity of acute
dose-response patterns in inflammation (influx of cells and protein) among species, with
humans and guinea pigs more responsive than rats and rabbits, and similarity of chronic dose-
response patterns for increased alveolar interstitial thickness in the CAR of the lung, with
monkeys being more responsive than humans and rats less responsive. In other words, the
quantitative relationship between animal and human responses is dependent on the animal
species and the endpoint.
In summary, there is an emerging consistency among a variety of O3 dosimetry
data sets and between the experimental data and theoretical predictions of O3 dose. The
convergence of experimental data with theoretical predictions lends a degree of confidence to
the use of theoretical models to predict total and regional O3 dose. The use of O3 dosimetry
data and models is beginning to provide a useful extrapolation of effects between animals and
humans. The data and models have thus far helped demonstrate that humans may be more
responsive to O3 than rats, but less responsive than monkeys with respect to acute and chronic
inflammatory responses. However, the monkey, with its similarity to the human in distal
airway structure, provides chronic effects data that may best reflect the degree to which a
comparably exposed human would respond. These findings, therefore, suggest that long-term
exposure to O3 could impart a chronic effect in humans.
1.9 Integrative Summary of Ozone Health Effects
This section summarizes the primary conclusions derived from an integration of
the known effects of O3 provided by animal lexicological, human clinical, and
epidemiological studies.
1. What are the effects of short-term (<8-h) exposures to ozone?
Recent epidemiology studies addressing the effects of short-term ambient exposure
to O3 in the population have yielded significant associations with a wide range of health
outcomes, including lung function decrements, aggravation of preexisting respiratory disease,
increases in daily hospital admissions and emergency department visits for respiratory causes,
and increased mortality. Results from lung function epidemiology studies generally are
consistent with the experimental studies in laboratory animals and humans.
Short-term O3 exposure of laboratory animals and humans causes changes in
pulmonary function, including tachypnea (rapid, shallow breathing), decreased lung volumes
and flows, and increased airway responsiveness to nonspecific stimuli. Increased airway
resistance occurs in both humans and laboratory animals, but typically at higher exposure
levels than other functional endpoints. In addition, adult human subjects experience
O3-induced symptoms of airway irritation such as cough or pain on deep inspiration. The
changes in pulmonary function and respiratory symptoms occur as a function of exposure
concentration, duration, and level of exercise. Adult human subjects with mild asthma have
responses in lung volume and airway responsiveness to bronchoconstrictor drugs that are
qualitatively similar to those of nonasthmatics. Respiratory symptoms are also similar, but
wheezing is a prevalent symptom in O3-exposed asthmatics in addition to the other
demonstrated symptoms of airway irritation. Airway resistance, however, increases relatively
1-28
-------
more in asthmatics from an already higher baseline. Recovery from the effects of O3 on
pulmonary function and symptoms is usually complete within 24 h of the end of exposure,
although other responses may persist somewhat longer.
An association between daily mortality and O3 concentration for areas with
high O3 levels (e.g., Los Angeles) has been suggested, although the magnitude
of such an effect is unclear.
Increased O3 levels are associated with increased hospital admissions and
emergency department visits for respiratory causes. Analyses from data in the
northeastern United States suggest that O3 air pollution is associated with a
substantial portion (on the order of 10 to 20%) of all summertime respiratory
hospital visits and admissions.
Pulmonary function in children at summer camps in southern Ontario, Canada,
in the northeastern United States, and in Southern California is associated with
O3 concentration. Meta-analysis indicates that a 0.5-mL decrease in FEVj is
associated with a 1-ppb increase in O3 concentration. For preadolescent
children exposed to 120 ppb (0.12 ppm) ambient O3, this amounts to an
average decrement of 2.4 to 3.0% in FEVj. Similar responses are reported for
children and adolescents exposed to O3 in ambient air or O3 in purified air for
1 to 2 h while exercising.
Pulmonary function decrements generally are observed in healthy subjects (8 to
45 years of age) after 1 to 3 h of exposure as a function of the level of
exercise performed and the O3 concentration inhaled during the exposure.
Group mean data from numerous controlled human exposure and field studies
indicate that, in general, statistically significant pulmonary function decrements
beyond the range of normal measurement variability (e.g., 3 to 5% for FEVj)
occur
(1) at >0.50 ppm O3 when at rest,
(2) at >0.37 ppm O3 with light exercise (slow walking),
(3) at >0.30 ppm O3 with moderate exercise (brisk walking),
(4) at >0.18 ppm O3 with heavy exercise (easy jogging), and
(5) at >0.16 ppm O3with very heavy exercise (running).
Smaller group mean changes (e.g., <5%) in FEVj have been observed at lower
O3 concentrations than those listed above. For example, FEVj decrements have
been shown to occur with very heavy exercise in healthy adults at 0.15 to 0.16
ppm O3, and such effects may occur in healthy young adults at levels as low as
0.12 ppm. Also, pulmonary function decrements have been observed in
children and adolescents at concentrations of 0.12 and 0.14 ppm O3 with heavy
exercise. Some individuals within a study may experience FEVj decrements in
excess of 15% under these exposure conditions, even when the group mean
decrement is less than 5%.
For exposures of healthy subjects performing moderate exercise during longer
duration exposures (6 to 8 h), 5% group mean decrements in FEVj were
observed at
(1) 0.08 ppm O3 after 5.6 h,
(2) 0.10 ppm O3 after 4.6 h, and
(3) 0.12 ppm O3 after 3 h.
1-29
-------
For these same subjects, 10% group mean FEVj decrements were observed at
0.12 ppm O3 after 5.6 and 6.6 h. As in the shorter duration studies, some
individuals experience changes larger than those represented by the group mean
changes.
An increase in the incidence of cough has been reported at O3 concentrations
as low as 0.12 ppm in healthy adults during 1 to 3 h of exposure with very
heavy exercise. Other respiratory symptoms, such as pain on deep inspiration,
shortness of breath, and lower respiratory scores (a combination of several
symptoms), have been observed at 0.16 to 0.18 ppm O3 with heavy and very
heavy exercise. Respiratory symptoms also have been observed following
exposure to 0.08, 0.10, and 0.12 ppm O3 for 6.6 h with moderate levels of
exercise.
Increases in nonspecific airway responsiveness in healthy adults have been
observed after 1 to 3 h of exposure to 0.40 but not 0.20 ppm O3 at rest and
have been observed at concentrations as low as 0.18 but not to 0.12 ppm
O3 during exposure with very heavy exercise. Increases in nonspecific airway
responsiveness during 6.6-h exposures with moderate levels of exercise have
been observed at 0.08, 0.10, and 0.12 ppm O3.
Short-term O3 exposure of laboratory animals and humans disrupts the barrier
function of the lung epithelium, permitting materials in the airspaces to enter lung tissue,
allowing cells and serum proteins to enter the airspaces (inflammation), and setting off a
cascade of responses.
Increased levels of PMNs and protein in lung lavage fluid have been observed
following exposure of healthy adults to 0.20, 0.30, and 0.40 ppm with very
heavy exercise and have not been studied at lower concentrations for 1- to 3-h
exposures. Increases in lung lavage protein and PMNs also have been
observed at 0.08 and 0.10 ppm O3 during 6.6-h exposures with moderate
exercise; lower concentrations have not been tested.
Short-term O3 exposure of laboratory animals and humans impairs AM clearance
of viable and nonviable particles from the lungs and decreases the effectiveness of host
defenses against bacterial lung infections in animals and perhaps in humans. The ability of
AMs to engulf microorganisms is decreased in humans exposed to 0.08 and 0.10 ppm O3 for
6.6 h with moderate exercise.
2. What are the effects of repeated, short-term exposures to ozone?
During repeated short-term exposures, some of the O3-induced responses are
partially or completely attenuated. Over a 5-day exposure, pulmonary function changes are
typically greatest on the second day, but return to control levels by the fifth day of exposure.
Most of the inflammatory markers (e.g., PMN influx) also attenuate by the fifth day of
exposure, but markers of cell damage (e.g., lactate dehydrogenase enzyme activity) do not
attenuate but continue to increase. Attenuation of lung function decrements is reversed
following 7 to 10 days without O3. Some inflammatory markers also are reversed during this
time period, but others still show attenuation even after 20 days without O3. The mechanisms
and impacts involved in attenuation are not known, although animal studies show that the
underlying cell damage continues throughout the attenuation process. In addition, attenuation
may alter the normal distribution of O3 within the lung, allowing more O3 to reach sensitive
1-30
-------
regions, possibly affecting normal lung defenses (e.g., PMN influx in response to inhaled
mi croorgani sm s).
3. What are the effects of long-term exposures to ozone?
Available data indicate that exposure to O3 for months and years causes structural
changes in several regions of the RT, but effects may be of the greatest importance in the
CAR (where the alveoli and conducting airways meet); this region typically is affected in
most chronic airway diseases of the human lung. This information on O3 effects in the distal
lung is extrapolated from animal toxicological studies because, to date, comparable data are
not available from humans. The apparent lack of reversal of effects during periods of clean
air exposure raises concern that seasonal exposures may have a cumulative impact over many
years. The role of adaptive processes in this response is unknown but may be critically
dependent on the temporal frequency or profile of exposure. Furthermore, the interspecies
diversity in apparent sensitivity to the chronic effects of O3 is notable, with the rat
representing the lower limit of response, and the monkey the upper limit. Epidemiological
studies attempting to associate chronic health effects in humans with long-term O3 exposure
provide only suggestive evidence that such a linkage exists.
Long-term exposure of one strain of female mice to high O3 levels (1 ppm) caused
a small, but statistically significant increase in lung tumors. There was no concentration-
response relationship, and rats were not affected. Genotoxicity data are either negative or
weak. Given the nature of the database, the effects in one strain of mice cannot yet be
extrapolated qualitatively to humans. Ozone (0.5 ppm) did not show tumor-promoting
activity in a chronic rat study.
4. What are the effects of binary pollutant mixtures containing ozone?
Combined data from laboratory animal and controlled human exposure studies of
O3 support the hypothesis that coexposure to pollutants, each at low-effect levels, may result
in effects of significance. The data from human studies of O3 in combination with NO2, SO2,
H2SO4, HNO3, or CO show no more than an additive response on lung spirometry or
respiratory symptoms. The larger number of laboratory animal studies with O3 in mixture
with NO2 and H2SO4 show that effects can be additive, synergistic, or even antagonistic,
depending on the exposure regimen and the endpoint studied. This issue of exposure to
copollutants remains poorly understood, especially with regard to potential chronic effects.
5. What population groups are at risk as a result of exposure to ozone?
Identification of population groups that may show increased sensitivity to O3 is
based on their biological responses to O3, preexisting lung disease (e.g., asthma), activity
patterns, personal exposure history, and personal factors (e.g., age, nutritional status).
The predominant information on the health effects of O3 noted above comes from
clinical and field studies on healthy, nonsmoking, exercising subjects, 8 to 45 years of age.
These studies demonstrate that, among this group, there is a large variation in sensitivity and
responsiveness to O3, with at least a 10-fold difference between the most and least responsive
individuals. Individual sensitivity to O3 also may vary throughout the year, related to
seasonal variations in ambient O3 exposure. The specific factors that contribute to this large
intersubject variability, however, remain undefined. Although differences in response may be
due to the dosimetry of O3 in the RT, available data show little difference on O3 deposition in
the lungs for inhalation through the nose or mouth.
1-31
-------
Daily life studies reporting an exacerbation of asthma and decrease in peak
expiratory flow rates, particularly in asthmatic children, appear to support the controlled
studies; however, those studies may be confounded by temperature, particle or aeroallergen
exposure, and asthma severity of the subjects or their medication use. In addition, field
studies of summertime daily hospital admissions for respiratory causes show a consistent
relationship between asthma and ambient levels of O3 in various locations in the northeastern
United States, even after controlling for independent contributing factors. Controlled studies
on mild asthmatics suggest that they have similar lung volume responses but greater airway
resistance changes to O3 than nonasthmatics. Furthermore, limited data from studies of
moderate asthmatics suggest that this group may have greater lung volume responses than
nonasthmatics.
Other population groups with preexisting limitations in pulmonary function and
exercise capacity (e.g., chronic obstructive pulmonary disease, chronic bronchitis, ischemic
heart disease) would be of primary concern in evaluating the health effects of O3.
Unfortunately, not enough is known about the responses of these individuals to make
definitive conclusions regarding their relative responsiveness to O3. Indeed, functional effects
in these individuals with reduced lung function may have greater clinical significance than
comparable changes in healthy individuals.
Currently available data follow on personal factors or personal exposure history
known or suspected of influencing responses to O3.
Human studies have identified a decrease in pulmonary function responsiveness
to O3 with increasing age, although symptom rates remain similar.
Toxicological studies are not easily interpreted but suggest that young animals
are not more responsive than adults.
Available toxicological and human data have not demonstrated conclusively
that males and females respond differently to O3. If gender differences exist
for lung function responsiveness to O3, they are not based on differences in
baseline pulmonary function.
Data are not adequate to determine whether any ethnic or racial group has a
different distribution of responsiveness to O3. In particular, the responses of
nonwhite asthmatics have not been investigated.
Information derived from O3 exposure of smokers is limited. The general trend
is that smokers are less responsive than nonsmokers. This reduced
responsiveness may wane after smoking cessation.
Although nutritional status (e.g., vitamin E deficiency) makes laboratory rats
more susceptible to O3-induced effects, it is not clear if vitamin E
supplementation has an effect in human populations. Such supplementation has
no or minimal effect in animals. The role of such antioxidant vitamins in
O3 responsiveness, especially their deficiency, has not been well studied.
Based on information presented in this document, the population groups that have
demonstrated increased responsiveness to ambient concentrations of O3 consist of exercising,
healthy and asthmatic individuals, including children, adolescents, and adults.
1-32
-------
2
Introduction
The photochemical oxidants found in ambient air in the highest concentrations are
ozone (O3) and nitrogen dioxide (NO2). Other oxidants, such as hydrogen peroxide (H2O2) and
the peroxyacyl nitrates, also have been observed, but in lower and less certain concentrations. In
1971, the U.S. Environmental Protection Agency (EPA) promulgated National Ambient Air
Quality Standards (NAAQS) to protect the public health and welfare from adverse effects of
photochemical oxidants. The 1971 photochemical oxidant standards were promulgated on the
basis of (1) commercially available measurement methodology,1 (2) uncertainties over the
concentrations of O3 and non-O3 photochemical oxidants in the atmosphere resulting from the
nonspecificity of the measurement methodology, and (3) uncertainties regarding the health and
welfare effects of the non-O3 photochemical oxidants found in ambient air. After 1971, however,
O3-specific commercial analytical methods became available, as did additional information on
concentrations and effects of the non-O3 photochemical oxidants. As a result, the chemical
designation of the standards was changed in 1979 from photochemical oxidants to Q. This
document focuses primarily on the scientific air quality criteria for C^ and, to a lesser extent, on
those for H2O2 and the peroxyacyl nitrates, particularly peroxyacetyl nitrate. The scientific air
quality criteria for NO2 are discussed in a separate document (U.S. Environmental Protection
Agency, 1993).
The previous O3 air quality criteria document (AQCD),^4/> Quality Criteria for Ozone
and Other Photochemical Oxidants (U.S. Environmental Protection Agency, 1986) was released
by EPA in August 1986 and a supplement, Summary of Selected New Information on Effects
of Ozone on Health and Vegetation (U.S. Environmental Protection Agency, 1992), was released
in January 1992. These documents were the basis for a March 1993 decision by EPA that
revision of the existing 1-h NAAQS for O3 was not appropriate at that time. That decision did
not take into account some of the newer scientific data that became available after completion of
the 1986 criteria document. The purpose of this document is to summarize the air quality criteria
for O3 available in the published literature through early 1995. This review was performed in
accordance with provisions of the Clean Air Act (CAA) to provide the scientific basis for periodic
reevaluation of the O3 NAAQS.
'The term "photochemical oxidants" historically has been defined as those atmospheric pollutants capable of
oxidizing neutral iodide ions (U. S. Environmental Protection Agency, 1978). A number of oxidants other than O3 are
measured, qualitatively if not quantitatively, by potassium iodide methods.
2-1
-------
This chapter provides a general introduction to the legislative and regulatory
background for decisions on the O3 NAAQS, as well as a general summary of the organization,
content, and major scientific topics presented in this document.
2.1 Legislative Background
Two sections of the CAA govern the establishment,review, and revision of the
NAAQS. Section 108 (U.S. Code, 1991) directs the Administrator of EPA to identify certain
ubiquitous pollutants that may reasonably be anticipated to endanger public health or welfare and
to issue air quality criteria for them. These air quality criteria are to reflect the latest scientific
information useful in indicating the kind and extent of all identifiable effects on public health or
welfare that may be expected from the presence of the pollutant in ambient air.
Section 109(a) of the CAA (U.S. Code, 1991) directs the Administrator of EPA to
propose and promulgate primary and secondary NAAQS for pollutants identified under Section
108. Section 109(b)(l) defines a primary standard as one the attainment and maintenance of
which, in the judgment of the Administrator and based on the criteria and allowing for an
adequate margin of safety, are requisite to protect the public health. The secondary standard, as
defined in Section 109(b)(2), must specify a level of air quality the attainment and maintenance of
which, in the judgment of the Administrator and based on the criteria, are requisite to protect the
public welfare from any known or anticipated adverse effects associated with the presence of the
pollutant in ambient air.
Section 109(d) of the CAA (U.S. Code, 1991) requires periodic review and, if
appropriate, revision of existing criteria and standards. Thus, the Administrator may find that
EPA's review and revision of criteria make appropriate the proposal of new or revised standards.
Alternatively, the Administrator may find that revision of the standards is inappropriate and
conclude the review by leaving the existing standards unchanged.
2.2 Regulatory Background2
On April 30, 1971, EPA promulgated primary and secondary NAAQS for
photochemical oxidants under Section 109 of the CAA (Federal Register, 1971). These standards
were set at an hourly average of 0.08 ppm total photochemical oxidants not to be exceeded more
than 1 h/year. On April 20, 1977, EPA announced (Federal Register, 1977) the first review and
updating of the 1970 Air Quality Criteria for Photochemical Oxidants in accordance with
Section 109(d) of the CAA. In preparing a revised AQCD, EPA made two external review drafts
of the document available for public comment, and these drafts were peer reviewed by the
Subcommittee on Scientific Criteria for Photochemical Oxidants of EPA's Science Advisory
Board (SAB). A final revised AQCD for O3 and other photochemical oxidants was published on
June 22, 1978.
Based on the 1978 revised AQCD and taking into account the advice and
recommendations of the Subcommittee and the comments received from the public, EPA
announced (Federal Register, 1979) a final decision to revise the NAAQS for photochemical
2This text is excerpted and adapted from the Proposed Decision on the National Ambient Air Quality
Standards for Ozone (Federal Register, 1992a).
2-2
-------
oxidants on February 8, 1979. The final ruling revised the level of the primary standard from
0.08 to 0.12 ppm, set the secondary standard identical to the primary standard, changed the
chemical designation of the standards from photochemical oxidants to Q, and revised the
definition of the point at which the standard is attained to "when the expected number of days per
calendar year with maximum hourly average concentrations above 0.12 ppm is equal to or less
than one" (see Table 2-1).
Table 2-1. National Ambient Air Quality Standards for Ozonea
Date of Promulgation Primary and Secondary NAAQS Averaging Time
February 8, 1979 0.12 ppmb (235//g/m3) 1 hc
aSee Appendix A for abbreviations and acronyms.
bl ppm = 1,962 Mg/m3, 1 Mg/m3 = 5.097 x 10'4 ppm at 25 °C, 760 mm Hg.
The standard is attained when the expected number of days per calendar year with a maximum hourly average
concentration above 0.12 ppm (235 ,ug/m3) is equal to or less than one.
On March 17, 1982, in response to requirements of Section 109(d) of the CAA, EPA
announced (Federal Register, 1982) that it was undertaking plans to revise the existing 1978
AQCD for O3 and other photochemical oxidants, and, on August 22, 1983, it announced (Federal
Register, 1983) that review of the primary and secondary NAAQS for Q had been initiated.
Public peer-review workshops on draft chapters of a revised AQCD were held December 15
through 17, 1982, and November 16 through 18, 1983. The EPA considered comments made at
both workshops in preparing the first external review draft that was made available (Federal
Register, 1984) on July 24, 1984, for public review.
On February 13, 1985 (Federal Register, 1985), and on April 2, 1986 (Federal
Register, 1986), EPA announced two public meetings of the Clean Air Scientific Advisory
Committee (CASAC) of EPA's SAB to be held March 4 through 6, 1985, and April 21 and 22,
1986, respectively. At these meetings, CAS AC reviewed external review drafts of the revised
AQCD for O3 and other photochemical oxidants. After completion of this review, CAS AC sent
the EPA Administrator a closure letter, dated October 22, 1986, indicating that the document
"represents a scientifically balanced and defensible summary of the extensive scientific literature."
The EPA released the final draft document in August 1986.
The first draft of the Staff Paper "Review of the National Ambient Air Quality
Standards for Ozone: Assessment of Scientific and Technical Information" was reviewed by
CASAC at a public meeting on April 21 and 22, 1986. At that meeting, CAS AC recommended
that new information on prolonged exposure effects of Oj be considered in a second draft of the
Staff Paper prior to closure. The CASAC reviewed this second draft and also a presentation of
new and emerging information on the health and welfare effects of 03 at a public review meeting
held on December 14 and 15, 1987. The CASAC concluded that sufficient new information
existed to recommend incorporation of relevant new data into a supplement to the 1986 AQCD
(O3 supplement) and in a third draft of the Staff Paper.
2-3
-------
A draft O3 supplement, Summary of Selected New Information on Effects of Ozone on
Health and Vegetation: Draft Supplement to Air Quality Criteria for Ozone and Other
Photochemical Oxidants, and the revised Staff Paper were made available to CAS AC and to the
public for review in November 1988. The O3 supplement reviewed and evaluated selected
literature concerning exposure- and concentration-response relationships observed for health
effects in humans and experimental animals and for vegetation effects. This literature appeared as
peer-reviewed journal publications or as proceedings papers from 1986 through late 1988.
On December 14 and 15, 1988, CASAC held a public meeting to review these
documents. The CASAC sent the EPA Administrator a closure letter dated May 1, 1989,
indicating that the draft O3 supplement, along with the 1986 AQCD, and the draft Staff Paper
"provide an adequate scientific basis for the EPA to retain or revise the primary and secondary
standards of ozone." The CASAC concluded that it would be some time before enough new
information on the health effects of multihour and chronic exposure to Q would be published in
scientific journals to receive full peer review and, thus, be suitable for inclusion in a criteria
document. The CASAC further concluded that such information could better be considered in the
next review of the O3 NAAQS. A final version of the O3 supplement has been published (U.S.
Environmental Protection Agency, 1992).
On October 22, 1991, the American Lung Association and other plaintiffs filed suit to
compel EPA to complete its review of the criteria and standards for Q. On May 4, 1992, the
U.S. District Court for the Eastern District of New York issued an order requiring the
Administrator of EPA to sign a proposed decision on whether to revise the standards for Q by
August 1, 1992, and to sign EPA's final decision by March 1, 1993.
On August 1, 1992, the Administrator signed a proposed decision not to revise the
existing NAAQS for O3 (Federal Register, 1992a), then, on March 1, 1993, signed EPA's final
decision, concluding that revision of the NAAQS was inappropriate at that time (Federal Register,
1993a). For reasons indicated in the proposed and final decisions, the March 1993 decision did
not take into consideration a number of recent studies on the health and welfare effects of Q that
had been published since the last literature review in early 1989. The EPA estimated that
approximately 3 years would be necessary to (1) incorporate the new studies into a revised
criteria document, (2) complete mandated CASAC review, (3) evaluate the significance of the key
information for regulatory decision-making purposes, and (4) publish a proposed decision on the
O3 NAAQS in the Federal Register.
The EPA intends to complete the current review of the criteria and standards for Q as
rapidly as possible. Accordingly, the National Center for Environmental Assessment (formerly the
Environmental Criteria and Assessment Office [ECAO]) of EPA's Office of Research and
Development, located in Research Triangle Park, NC, has given very high priority to review and
revision of the air quality criteria for O3. The ECAO began by announcing the commencement of
the review and identification of new information (Federal Register, 1992b). After assessing and
evaluating pertinent new studies, ECAO prepared a preliminary draft of a revised AQCD that was
reviewed in a series of expert peer reviewed workshops (Federal Register, 1993b,c). Comments
received at the workshops were used to revise the preliminary draft for external review (Federal
Register, 1994a). Public peer review meetings were held by CASAC to provide advice on the
scientific and technical adequacy of the external review draft (Federal Register, 1994b) and the
subsequent revised draft (Federal Register, 1995). The final document was prepared on the basis
of comments received from the public and CASAC reviews and provides a scientific basis for
review of the existing O3 standards. The EPA's Office of Air Quality Planning and Standards
2-4
-------
(OAQPS) is completing its preparation of a draft staff paper assessing the most significant
information contained in this AQCD and presenting staff recommendations on whether revisions
to the NAAQS for O3 are appropriate. After reviews of the draft staff paper by the public and
CASAC, the Administrator will decide whether to propose revisions to the Q NAAQS.
2.3 Summary of Major Scientific Topics Presented
A number of separate topics and issues are addressed in this Q criteria document.
Some of the key questions addressed are highlighted below by document section.
2.3.1 Air Chemistry
• What concerns still exist regarding precision and accuracy of measurements of Q
and its precursors?
• What is the order of magnitude of current estimates of natural emissions of
O3 precursors and emissions from anthropogenic sources and their relevance to
tropospheric O3 photochemistry?
• What new scientific information exists on the roles of meteorologic and
climatologic factors in O3 formation and transport?
• Are the reaction pathways of all major precursors to Q understood? Have all
major reaction products been identified? How are the reactions and products
represented in air quality models?
• What is the status of development, application, evaluation, and verification of air
quality models?
2.3.2 Air Quality
• What are the trends and geographic differences in Q concentrations across the
United States?
• What are diurnal and seasonal patterns of 1-h average Q concentrations for urban
and nonurban sites and for attainment versus nonattainment areas?
• What is known about patterns of co-occurrence of Q with other pollutants in the
atmosphere?
• What O3 exposure assessment data are available for agricultural crops and forests?
• To what level and to what extent are humans typically exposed to C^ in the course
of normal, everyday activities?
2.3.3 Environmental Effects
• What are the effects of ambient O3 concentrations on vegetation (i.e., agricultural
and horticultural crops; urban landscape trees, shrubs, and flowers; forest tree
species)?
• What characteristics of air quality (e.g., summary statistics) are relevant to these
effects on vegetation?
• What are the long-term effects of O3 exposures on natural ecosystems?
• Is there important new information on the effects of Q on nonbiological materials?
2-5
-------
2.3.4 Health Effects
• What O3 concentration and exposure duration relationships exist for effects on lung
structure, function, and host defense mechanisms, and what are the important
modifiers of these effects?
• What are the mechanisms of O3-induced lung injury?
• Can dosimetry models predict human population responses to Q on the basis of
laboratory animal data?
• Does long-term exposure to O3 lead to the development of chronic lung disease or
to an increased frequency or exacerbation of other chronic respiratory outcomes?
• What segments of the population are most susceptible to effects from exposure to
03?
2.4 Organization and Content of the Document
This document critically evaluates and assesses scientific information on the health and
welfare effects associated with exposure to the concentrations of 63 and related photochemical
oxidants present in ambient air. Although the document is not intended to be an exhaustive
literature review, it is intended to selectively cover the pertinent literature through 1995. The
references cited in the document should be reflective of the state of knowledge on those issues
most relevant to review of the NAAQS for O3, now set at 0.12 ppm for 1 h. Although emphasis
is placed on the presentation of health and welfare effects data, other scientific data will be
presented and evaluated in order to provide a better understanding of the nature, sources,
distribution, measurement, and concentrations of Q and related photochemical oxidants in
ambient air, as well as the characterization of population exposure to these pollutants.
To aid in the development of this document, summary tables of the relevant published
literature have been provided to supplement a selective discussion of the literature. Most of the
scientific information selected for review and comment in the text comes from the more recent
literature published since completion of the previous Q criteria document (U.S. Environmental
Protection Agency, 1986). Some of these newer studies were reviewed briefly in the supplement
to that document (U.S. Environmental Protection Agency, 1992), but more intense evaluation of
these studies has been included. Other studies, however, are included if they contain unique data,
such as the documentation of a previously unreported effect or of a mechanism of an effect, or if
they were multiple-concentration studies designed to provide exposure-response relationships.
Emphasis is placed on studies conducted at or near 63 concentrations found in ambient air. For
animal toxicology studies, typically only those studies conducted at less than 1 ppm Q are
considered. Studies that are presented in the previous criteria document and whose data were
judged to be significant because of their usefulness in deriving the current NAAQS are discussed
briefly in the text. Other, older studies also are discussed in the text if they were judged to be
(1) open to reinterpretation because of newer data or (2) potentially useful in deriving revised
standards for O3. The reader should, however, consult the more extensive discussion of these
"key" studies in the previous document. Generally, only published information that has
undergone scientific peer review is included in the criteria document.
Certain issues of direct relevance to standard setting are not explicitly addressed in this
document, but instead are analyzed in documentation prepared by OAQPS as part of its
regulatory review process. Such issues include (1) determining what constitutes an "adverse
2-6
-------
effect" and delineation of particular adverse effects that the primary and secondary NAAQS are
intended to protect against, (2) exposure assessment, (3) assessment of consequent risks based on
health and exposure analyses, and (4) factors to be considered in determining an adequate margin
of safely. Key points and conclusions from such analyses are summarized in the Staff Paper
prepared by OAQPS and reviewed by CASAC and the public. Although scientific data contribute
significantly to decisions regarding the above issues, their resolution cannot be achieved solely on
the basis of experimentally acquired information. Final decisions on items 1 and 4 are made by the
EPA Administrator, as mandated by the CAA.
A fourth issue directly pertinent to standard setting is identification of populations at
risk, which is basically a determination by EPA of the subpopulations to be protected by the
promulgation of a given standard. This issue is addressed only partially in the criteria document.
For example, information is presented on factors, such as preexisting disease, that biologically
may predispose individuals and subpopulations to more severe effects from exposures to Q. The
identification of a population at risk, however, requires information above and beyond data on
biological predisposition, such as information on levels of exposure, activity patterns, and
personal habits. Such information is included in the Staff Paper developed by OAQPS.
Finally, the O3 air quality document considers only the scientific and technical issues
that are important for standard setting, not those issues relative to implementation of the
O3 NAAQS. For example, certain issues related to the control strategies for attainment of the
standard and to possible atmospheric consequences of control strategy design are not discussed in
this document. This limitation also includes discussion of impacts consequent to possible changes
in the O3 NAAQS. These issues would be better addressed in regulatory impact analyses or cost-
benefit analyses that may be prepared as part of the C^ NAAQS decision package.
This document is structured as follows: Chapter 1 (executive summary and
conclusions) provides a concise presentation of key information and conclusions from all
subsequent chapters. This is followed by this brief introduction (Chapter 2) containing
information on the legislative and regulatory background for review of the Q NAAQS, as well as
an overview of the organization of this document. Chapter 3 provides information on the
chemistry, sources, emissions, measurement, and transport of Q, and related photochemical
oxidants and their precursors, and Chapter 4 covers environmental concentrations, patterns, and
exposure estimates of O3 and oxidant air quality. This is followed by Chapter 5, which deals with
environmental effects of O3 and related photochemical oxidants. Chapters 6, 7, and 8 discuss
animal toxicological studies, human health effects, and extrapolation of animal toxicological data
to humans, respectively. Finally, Chapter 9, provides an integrative and interpretive evaluation of
health effects associated with exposure to O3.
2-7
-------
References
Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30) 36: 8186-8201.
Federal Register. (1977) Review of the photochemical oxidant and hydrocarbon air quality standards. F. R. (April 20)
42: 20493-20494.
Federal Register. (1979) National primary and secondary ambient air quality standards: revisions to the national ambient
air quality standards for photochemical oxidants. F. R. (February 8) 44: 8202-8221.
Federal Register. (1982) Air quality criteria document for ozone and other photochemical oxidants. F. R. (March 17) 47:
11561.
Federal Register. (1983) Review of the national ambient air quality standards for ozone. F. R. (August 22) 48: 38009.
Federal Register. (1984) Draft air quality criteria document for ozone and other photochemical oxidants. F. R. (July 24)
49: 29845.
Federal Register. (1985) Science Advisory Board; Clean Air Scientific Advisory Committee; open meeting.
F. R. (February 13)50:6049.
Federal Register. (1986) Science Advisory Board; Clean Air Scientific Advisory Committee; open meeting. F. R. (April
2)51: 11339.
Federal Register. (1992a) National ambient air quality standards for ozone; proposed decision. F. R. (August 10) 57:
35542-35557.
Federal Register. (1992b) Air quality criteria for ozone and related photochemical oxidants: call for information.
F. R. (August 27) 57: 38832.
Federal Register. (1993a) National ambient air quality standards for ozone—final decision. F. R. (March 9)
58: 13008-13019.
Federal Register. (1993b) Peer-review workshops on the health effects of ozone and related photochemical oxidants:
notice of public meeting. F. R. (July 1) 58: 35454.
Federal Register. (1993c) Three review workshops on draft chapters of a revised air quality criteria for ozone and
related photochemical oxidants: notice of public meetings. F. R. (September 14) 58: 48063.
Federal Register. (1994a) External review draft of revised air quality criteria for ozone and related photochemical
oxidants: notice of availability of external review draft. F. R. (January 31) 59: 4278.
Federal Register. (1994b) Science Advisory Board Clean Air Scientific Advisory Committee review of a draft revised
air quality criteria for ozone and related photochemical oxidants. F. R. (June 3) 59: 28857-28858.
Federal Register. (1995) Science Advisory Board; notification of public advisory committee meetings. F. R. (March 3)
60: 11971-11974.
U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air quality
standards. U. S. C. 42: §§7408-7409.
U.S. Environmental Protection Agency. (1978) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; report no. EPA-600/8-78-004. Available from: NTIS, Springfield, VA; PB80-
124753.
2-8
-------
U.S. Environmental Protection Agency. (1986) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS, Springfield, VA;
PB87-142949.
U.S. Environmental Protection Agency. (1992) Summary of selected new information on effects of ozone on health and
vegetation: supplement to 1986 air quality criteria for ozone and other photochemical oxidants. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; report no. EPA/600/8-88/105F. Available from: NTIS, Springfield, VA;
PB92-235670.
U.S. Environmental Protection Agency. (1993) Air quality criteria for oxides of nitrogen. Research Triangle Park, NC:
Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
nos. EPA/600/8-91/049aF-cF. 3v. Available from: NTIS, Springfield, VA; PB95-124533, PB95-124525,
andPB95-124517.
2-9
-------
Tropospheric Ozone
and Its Precursors
3.1 Introduction
Ozone (O3) and other oxidants found in ambient air, such as peroxyacyl nitrates
(PANs) and hydrogen peroxide (H2O2), are formed as the result of atmospheric physical and
chemical processes involving two classes of precursor pollutants, volatile organic compounds
(VOCs) and nitrogen oxides (NOX). The formation of O3 and other oxidants from these
precursors is a complex, nonlinear function of many factors, including temperature, the
intensity and spectral distribution of sunlight, atmospheric mixing and related meteorological
conditions, the concentrations of the precursors in ambient air and the ratio between VOC and
NOX, and the reactivity of the organic precursors.
An understanding of the atmospheric chemistry and meteorological parameters and
processes responsible for the formation and occurrence of elevated concentrations of O3 in
ambient air is basic to the formulation of strategies and techniques for its abatement. Such an
understanding is required for representing those parameters and processes adequately in
predictive models used to determine the emission reductions needed for complying with the
National Ambient Air Quality Standards (NAAQS) for O3. In addition, the identification and
quantification of O3 precursors in ambient air are essential, along with emission inventories or
emission models, for the development, verification, and refinement of photochemical air
quality models and for comparisons of ambient concentrations with emission inventories
(source reconciliation), as a check on the accuracy of measurements and of inventories.
Product identification and quantification of yields, in both chambers and ambient
air, are helpful in the verification of photochemical air quality models and in testing
theoretical chemical mechanisms. Likewise, product identification and quantification are
useful in determining the need for research on the potential effects of the simultaneous or
sequential co-occurrence with O3 and related oxidants of multiple air pollutants.
The ability to measure O3 and its precursors, its reaction products, and the products
of the atmospheric reactions of its respective precursors is essential for understanding the
atmospheric chemistry of O3 formation, verifying chemical mechanisms and models,
quantifying emission rates, and adequately characterizing exposure-response factors for both
biological and nonbiological receptors.
For these reasons, this chapter presents information on a broad range of topics.
The chapter describes the chemical processes by which O3 and other photochemical oxidants
are formed in ambient air (Section 3.2). The chapter also characterizes the nature of the
precursors in terms of their sources and emissions into the atmosphere and their
3-1
-------
concentrations in ambient air (Section 3.4), as well as the methods by which their
concentrations in ambient air are measured (Section 3.5).
In addition to information on the chemistry of oxidants and their precursors, this
chapter includes a discussion of meteorological processes (Section 3.3) that contribute to the
formation of O3 and other oxidants and govern their transport and dispersion. Finally, an
overview is given (Section 3.6) of models of the relationships between precursor emissions
and O3 formation in the atmosphere.
Readers are referred to other sources (e.g., Finlayson-Pitts and Pitts, 1986;
Seinfeld, 1986; the U.S. Environmental Protection Agency, 1986a; National Research Council,
1991) for additional information on the chemical and physical aspects of photochemical air
pollution.
3.2 Tropospheric Ozone Chemistry
3.2.1 Background Information
Ozone is formed photochemically in the stratosphere and transported downward,
resulting in the presence of O3 in the natural or "clean" troposphere. The presence of O3 in
the clean troposphere, in the absence of perturbations caused by human activities, is highly
important because O3 is a precursor to the hydroxyl (OH) radical, the key intermediate species
in the tropospheric degradation of VOCs emitted into the atmosphere. Although O3 at
relatively low concentrations is an integral part of the clean troposphere, its presence at higher
concentrations is detrimental.
The chemical processes occurring in the atmosphere that lead to the formation of
O3 and other photochemical air pollutants are complex. Tropospheric O3 is formed as a result
of (1) the emissions of NOX and VOCs into the atmosphere from anthropogenic and natural
sources, (2) the transport of these emissions and their reaction products, and (3) chemical
reactions occurring in the atmosphere concurrent with transport and dispersion of the
emissions. These processes lead to the formation of O3 and other photochemical oxidants,
such as peroxyacetyl nitrate (PAN), nitric acid (FINO3), and sulfuric acid (H2SO4), and to
other compounds, such as particulate matter and formaldehyde (HCHO) and other carbonyl
compounds. Additionally, deposition of gases and particles along the trajectory of an air
parcel occurs, reducing the concentrations of precursors and products in the atmosphere, but
possibly leading to adverse impacts on the earth's environment.
The basic process leading to the photochemical formation of O3 in the troposphere
involves the photolysis of nitrogen dioxide (NO2) to yield nitric oxide (NO) and a ground-
state oxygen atom, O(3P),
NO2 + hv -» NO + O(3P), (3-1)
which then reacts with molecular oxygen to form O3:
O(3P) + O2 + M -» O3 + M, where M = air. (3-2)
The NO and O3 react to reform NO2:
3-2
-------
NO + (X -» NO, + O7. (3-3)
The presence of reactive VOCs leads to the conversion of NO to NO2 without the
intermediary of O3 (Reaction 3-3), and the photolysis of NO2 then leads to the formation of
elevated levels of O3:
voc
NO * NO2 . (3-4)
The photochemical cycles leading to O3 production are best understood through a
knowledge of the chemistry of the atmospheric oxidation of methane (CH4), which can be
viewed as being the chemistry of the clean or unpolluted troposphere (although this is a
simplification because vegetation releases large quantities of complex VOCs into the
atmosphere). Although the chemistry of the VOCs emitted from anthropogenic and biogenic
sources in polluted urban and rural areas is more complex, a knowledge of the CH4 oxidation
reactions aids in understanding the chemical processes occurring in the polluted atmosphere
because the underlying chemical principles are the same.
This section first describes the structure of the atmosphere, followed by discussions
of the formation of the OH radical and of tropospheric NOX chemistry. The photochemical
formation of tropospheric O3 from the oxidation of CH4 then is discussed in some detail
because the CH4 oxidation cycle serves as a model for the photochemical formation of O3. In
Section 3.2.4, the chemistry of the major classes of nonmethane VOCs and the formation of
O3 from these VOCs are discussed. Finally, in Section 3.2.5, a brief account of the
photochemical formation of aerosols is given because the same processes that lead to the
formation of elevated levels of O3 result in the formation of both paniculate matter (leading
to visibility degradation) and atmospheric acidity.
3.2.2 Structure of the Atmosphere
Earth's atmosphere is composed of a number of layers (Mcllveen, 1992). For the
purposes of this chapter, those of concern are the troposphere and the stratosphere, and the
boundary between them, the tropopause.
The troposphere extends from the earth's surface to the tropopause (-10 to 18 km
altitude, depending on latitude and season). The altitude of the tropopause is greatest in the
tropics and lowest in the wintertime polar regions, with an average altitude of -14 km. The
temperature in the troposphere decreases with increasing altitude from an average of 290 K at
the earth's surface to -210 to 220 K at the tropopause, and the pressure decreases from -760
torr at the earth's surface to -100 torr at the tropopause.
The stratosphere extends from the tropopause to an altitude of -50 km. In the
stratosphere, the temperature increases with increasing altitude from -210 to 220 K at the
tropopause to -270 K at the top of the stratosphere. The pressure in the stratosphere
decreases with increasing altitude from -100 torr at the tropopause to -1 torr at the top of the
stratosphere.
3-3
-------
3.2.2.1 Vertical and Horizontal Mixing in the Atmosphere
In the troposphere, temperature generally decreases with increasing altitude.
As will be discussed in Section 3.3, the lowest 1 to 2 km of the troposphere is influenced by
the planetary boundary layer (PEL) and, in certain locales, by inversion layers. These
boundary and inversion layers inhibit the vertical movement of pollutants into the free
troposphere. Above inversion and boundary layers, vertical mixing in the "free" troposphere
has a time scale of =10 to 30 days (Langner et al., 1990; World Meteorological Organization,
1990a).
Because temperature increases with increasing altitude in the stratosphere, vertical
mixing in the stratosphere is slow, with a time scale of months to a few years.
Horizontal mixing in the troposphere occurs both within and between the
hemispheres. The time scale for mixing between the Northern and Southern Hemispheres is
=1 year (Cicerone, 1989; Singh and Kanakidou, 1993). Transport within a hemisphere is
more rapid (Graedel et al., 1986a), and local, regional, and global transport distances of <100
km, 100 to 1,000 km, and >1,000 km, respectively, are observed. For a wind speed of
15 km h"1 (~4 m s"1), transport times over these local, regional, and global distances are a few
hours, a few hours to a few days, and >10 days, respectively.
3.2.2.2 Formation of Stratospheric Ozone
At altitudes between approximately 20 and 35 km, the stratosphere has a layerjrf
air containing O3 at mixing ratios up to approximately 10 ppm. The sun emits radiation>170
nm, and this radiation impacts the upper levels of the atmosphere. The bulk composition of
the atmosphere (78.1% nitrogen [N2], 21.0% molecular oxygen [O2], 0.9% argon [Ar], 0.03%
carbon dioxide [CO2], with variable trace gas concentrations) is invariant up to at least 50 km
(Mcllveen, 1992). The shorter wavelength radiation (175 to 240 nm) is absorbed by O2 in the
stratosphere, leading to dissociation into two ground-state oxygen atoms, O(3P),
O2 + hv -> 2 O(3P), (3-5)
followed by the reaction of O(3P) atoms with O2 in the presence of a third body, M, to form
03:
O(3P) + O2 + M -» O3 + M, where M = air. (3-2)
Ozone also photolyzes, at wavelengths <360 nm (DeMore et al., 1992),
O3 + hv -> O2 + O, (3-6)
where the oxygen atom produced can be in the ground state, O(3P), or electronically excited
state, O(1D). The O(JD) atoms produced are deactivated to the ground-state O(3P) atom by
N2, O2, CO2, and Ar:
3-4
-------
OfD) + M -> O(3P) + M, where M = N7, O7, Ar, and CO7. (3-7)
The reaction of O(3P) atoms with O3 is the termination step of this reaction sequence,
O(3P) + O3 -» 2 O2. (3-8)
These reactions, called the Chapman reactions (Chapman, 1930), are responsible
for the layer of O3 found in the stratosphere. Because the stratospheric O3 layer absorbs the
sun's radiation below -290 nm, only radiation of wavelengths>290 nm can penetrate into the
troposphere and impact the earth's surface. Any depletion of the stratospheric O3 layer allows
shorter wavelength ultraviolet radiation (<320 nm) to be transmitted through the stratosphere
and into the troposphere.
In addition to the biological effects expected from increased ultraviolet B (UV-B)
radiation (290 to 320 nm), increased penetration of UV-B into the troposphere can lead to
changes in tropospheric O3. Model calculations indicate that O3 in the troposphere could
increase with increasing UV-B in urban and rural areas impacted by anthropogenic NOX
emissions (Gery et al., 1988; Liu and Trainer, 1988; Thompson et al., 1989; Thompson, 1992)
but could decrease with increasing UV-B in remote tropospheric areas characterized by low
NOX levels (Liu and Trainer, 1988; Thompson et al., 1989). Besides the implications of long-
term trends in stratospheric O3 concentrations leading to corresponding changes in the
intensity of UV-B radiation impacting the troposphere, short-term changes (including daily
changes) in stratospheric O3 levels lead to short-term changes in the rates of photolysis of
several important species. These include the photolysis of formaldehyde to produce radicals
and of O3 to form the OH radical. These changes in photolysis rates affect the formation
rates and ambient concentrations of key radical intermediates, specifically of the OH radical,
in the troposphere. Information concerning such short-term changes in stratospheric
O3 concentrations is needed as input to urban and regional airshed computer models of
photochemical air pollution formation.
In the clean atmosphere, stratospheric O3 also is influenced by the emission of
nitrous oxide (N2O) from soils and oceans (World Meteorological Organization, 1992).
Because N2O is chemically inert in the troposphere and does not photolyze (Prinn et al.,
1990), it therefore is transported into the stratosphere, where it undergoes photolysis and also
reacts with OfD) atoms (DeMore et al., 1992; Atkinson et al., 1992a). The reaction of N2O
with the O(JD) atom is the major source of stratospheric NO, which then participates in a
series of reactions known as the NOX catalytic cycle (Crutzen, 1970; Johnston, 1971).
NO + O, -* NO, + O, (3-3)
The Chapman reactions and the NOX catalytic cycle reactions control the O3 concentrations in
the lower clean stratosphere.
Additional reaction sequences leading to the removal of stratospheric O3 arise from
the C1OX and BrOx catalytic cycles, which result when chlorine (Cl)- and bromine
3-5
-------
NO2 + O(3P) -» NO + O2 (3.9)
Net: O(3P) + O3 -> 2 O2
(Br)-containing organic compounds are emitted into the atmosphere. These O3-depleting
compounds include the chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HCFCs),
carbon tetrachloride (CC14), methyl chloroform, halons, and methyl bromide (Anderson et al.,
1991; Rowland, 1990, 1991; World Meteorological Organization, 1992). Analogous to N2O,
the CFCs, CC14, and certain halons (CF3Br and CF2ClBr) are inert in the troposphere and are
transported into the stratosphere, where they photolyze to generate Cl or Br atoms (World
Meteorological Organization, 1992). Methyl bromide and the HCFCs react to a large extent
in the troposphere, so that only a fraction of these Cl- and Br-containing species that are
emitted into the troposphere are transported into the stratosphere (World Meteorological
Organization, 1990b, 1992).
3.2.3 Background Ozone in the Troposphere
As noted in Section 3.2.1, O3 is present in the troposphere even in the absence of
human activities. Although this "natural" O3 has received widely varying estimates in the
literature, there has not been an attempt to standardize the definition because natural
background O3 is a multidimensional and complex concept. Concentrations of background O3
can vary with temperature; wind speed and direction; vertical motion; geographical location,
including latitude and altitude; and season of the year. Because of the decrease of total
pressure with increasing altitude, the O3 concentration in the clean troposphere may be taken
to be reasonably independent of altitude at ~7 x lo11 molecules cm"3.
For purposes of this document, the primary focus is on the background
O3 concentration near the surface over the United States during the O3 season. However, the
length of the O3 season varies from state-to-state, depending predominantly on latitude (see
Chapter 4). These variations in O3 season affect the determination of which seasonal
averaging period to apply when estimating O3 concentrations. Based on available assessments
of O3 monitoring measurements, a daytime, 7-h (0900 to 1559 hours), seasonal (April to
October) average O3 concentration of 25 to 45 ppb can be assumed (Altshuller and Lefohn,
1996) as an estimated background concentration (see Chapter 4). Although one component of
this background is of natural origin, the other can be attributed to anthropogenic contributions
associated with long-range transport of O3. This assumption is consistent with the relatively
long lifetime of O3 in the troposphere, which can be as long as 30 to 60 days.
The background of O3 can be attributed to the following sources: downward
transport of stratospheric O3 through the free troposphere to near ground level, in situ
O3 production from methane emitted from swamps and wetlands reacting with natural NOX
emitted from soils and lighting strikes and from the downward transport of NO from the
stratosphere into the troposphere, and in situ production of O3 from the reactions of biogenic
VOCs with natural NOX (National Research Council, 1991). Another source to be considered
is the long-range transport of O3 from distant pollutant sources.
It is important to appreciate that NOX has a limited lifetime, often estimated to be
as short as 6 h in plumes (Altshuller, 1986) and possibly up to 1 to 2 days under less polluted
conditions. Because of this relatively short lifetime, the NOX emitted from cultivated areas in
3-6
-------
the United States as a result of agricultural practices may not survive long enough to interact
with substantial emissions of biogenic VOCs in distant forested areas. There also is no direct
way to distinguish natural NOX from anthropogenic NOX in a rural or remote location.
3.2.3.1 Tropospheric Hydroxyl Radicals
It is now recognized that the key reactive species in the troposphere is the OH
radical, which is responsible for initiating the degradation reactions of almost all VOCs.
In the presence of NO, these OH radical reactions with VOCs lead to the formation of
O3 and, hence, to O3 concentrations above those encountered in the clean troposphere. The
OH radical is produced from the ultraviolet (UV) photolysis of O3. Ozone photolyzes in the
UV radiation at wavelengths <320 nm to generate the electronically excited O(1D) atom
(DeMore et al., 1992; Atkinson et al., 1992a),
O3 + hv -* O2 + OfD). (3-6a)
The O(1D) atoms either are deactivated to the ground state O(3P) atom by Reaction 3-7 or
they react with water vapor to form the OH radical:
OfD) + H2O -^ 2 OH. (3-10)
The O(3P) atoms formed directly in the photolysis of O3 or formed from deactivation of O(1D)
atoms (Reaction 3-7) reform O3 through Reaction 3-2. At room temperature and 50% relative
humidity, 0.2 OH radicals are formed per O(JD) atom generated from the photolysis of O3.
Hydroxyl radical production from Reactions 3-6a and 3-10 is balanced by reaction of the OH
radical with carbon monoxide (CO) and CH4. Because the water vapor mixing ratio decreases
with increasing altitude in the troposphere (Logan et al., 1981; World Meteorological
Organization, 1992) and the O3 mixing ratio generally increases with increasing altitude, the
OH radical concentration is expected to be reasonably independent of altitude (Dentener and
Crutzen, 1993).
A knowledge of ambient tropospheric OH radical concentrations is needed for an
understanding of tropospheric chemistry and to reliably calcuate the lifetimes of chemical
compounds. Because OH and hydroperoxyl (HO2) radicals are interrelated through a series of
reactions (Section 3.2.3.3), concurrent measurements of OH and HO2 radical concentrations
improve the knowledge of tropospheric chemistry. Only in the past few years have
measurements been made of lower tropospheric OH radical concentrations (see, for example,
Felton et al., 1990; Hofzumahaus et al., 1991; Eisele and Tanner, 1991; Mount and Eisele,
1992; Comes et al., 1992; Hard et al., 1992). The limited data available show that, as
expected, the OH radical concentrations exhibit a diurnal profile, with daytime maximum
concentrations of several times 106 molecules cm"3. A global, annually, seasonally, and
diurnally averaged tropospheric OH radical concentration also can be derived from the
estimated emissions and measured atmospheric concentrations of methylchloroform (CH3CC13)
and the rate constant for the reaction of the OH radical with CH3CC13 (its major tropospheric
loss process). Using this method, Prinn et al. (1992) have derived a 24-h average OH radical
concentration of 8 x 105 molecules cm"3 (equivalent to a 12-h daytime average of 1.6 x 106
molecules cm"3 [~0.1 ppt]). Ambient air measurements of the decay of nonmethane
3-7
-------
hydrocarbons in urban plumes (Blake et al., 1993) give OH radical concentrations of similar
magnitude to those derived from direct tropospheric measurements and globally averaged
estimates.
3.2.3.2 Tropospheric Nitrogen Oxides Chemistry
The presence of NOX is necessary for the formation of O3 from the oxidation of
CH4 and other VOCs. Sources of tropospheric NOX include downward transport from the
stratosphere, in situ formation from lightning (National Research Council, 1991; World
Meteorological Organization, 1992) (see Section 3.4.1.2), and emission from soils (National
Research Council, 1991; World Meteorological Organization, 1992). Recent measurements
show that the NOX concentrations over maritime areas increase slightly with increasing
altitude, from -15 ppt in the marine boundary layer (Carroll et al., 1990) to -30 to 40 ppt at
3 to 7 km altitude (Ridley et al., 1989; Carroll et al., 1990). Significantly higher NOX
concentrations (-100 ppt) have been observed in the boundary layer over relatively unpolluted
continental areas (Carroll et al., 1990), with the NOX concentrations decreasing with
increasing altitude to «50 ppt at 3 to 7 km (Ridley et al., 1989; Carroll et al., 1990).
In the troposphere, NO, NO2, and O3 are interrelated by the following reactions:
NO + O3 -* NO2 + O2 (3-3)
NO2 + hv -* NO + O(3P) (3-1)
O(3P) + O2 + M -^ O3 + M. (3-2)
Because Reaction 3-2 is fast (the lifetime of an O(3P) atom at 298 K and 760 torr of air is
-10"5 s), the O3 concentration at photoequilibrium is given by
[03] = JJN02]/k3[NO], (3-11)
where Jj and k3 are the photolysis rate of NO2 (-0.5 min"1 for an overhead sun) and the rate
constant for the reaction of NO with O3, respectively.
There are other important reactions involving NOX. The reaction of NO2 with
O3 leads to the formation of the nitrate (NO3) radical,
NO2 + O3 -+ NO3 + O2, (3-12)
which in the lower troposphere is nearly in equilibrium with dinitrogen pentoxide (N2O5):
NO3 + NO2 M N2O5. (3-13, -3-13)
3-8
-------
However, because the NO3 radical photolyzes rapidly (with a lifetime of ~5 s for an overhead
sun [Atkinson et al., 1992a]),
* NO + O2 (10%) (3-14a)
NO, +hv
0(3P) (90%), (3-14b)
its concentration remains low during daylight hours, but can increase after sunset to nighttime
concentrations of <5 x 107 to 1 x 1010 molecules cm"3 (<2 to 430 ppt) over continental areas
influenced by anthropogenic emissions of NOX (Atkinson et al., 1986). Nitrate radical
concentrations over marine areas are low because NOX concentrations are low over lower
tropospheric marine areas (Noxon, 1983), and an NO3 radical mixing ratio of 0.25 ppt has
been measured at 3 km altitude in Hawaii (Noxon, 1983). Atkinson (1991) has suggested the
use of a 12-h nighttime average NO3 radical concentration of 5 x 108 molecules cm"3 in the
lower troposphere over continental areas, with an uncertainty of a factor of =10.
The tropospheric chemical removal processes for NOX involve the daytime reaction
of NO2 with the OH radical and the nighttime wet and dry deposition of N2O5 to produce
HNO3.
OH + NO2 > HNO3 (3-15)
H2O
N2O5 ^ HNO3 (3-16)
The gas-phase reaction of the OH radical with NO2 is the major and ultimate removal process
for NOX in the troposphere. This reaction removes radicals (OH and NO2) and competes with
the reaction of the OH radical with VOCs. Gaseous HNO3 formed from Reaction 3-15
undergoes wet and dry deposition, including combination with gaseous ammonia (NH3) to
form particulate phase ammonium nitrate (N2H4O3). The tropospheric lifetime of NOX due to
chemical reaction (mainly Reaction 3-15) is =1 to 2 days. The tropospheric NOX reactions are
shown schematically in Figure 3-1. It should be noted that OH radicals also can react with
NO to produce nitrous acid (HNO2):
M
OH + NO ^ HNO2. (3-17)
In urban areas, HNO2 also can be formed during nighttime hours (Harris et al., 1982; Pitts
et al., 1984a; Rodgers and Davis, 1989), apparently from the heterogeneous hydrolysis of NO2
or NOX, or both (Sakamaki et al., 1983; Pitts et al., 1984b; Svensson et al., 1987;
3-9
-------
Emission
HNO
Figure 3-1. The cyclic reactions of tropospheric nitrogen oxides.
N205
Wet/Dry
Deposition
HNO,
Jenkin et al., 1988; Lammel and Perner, 1988; Notholt et al., 1992a,b).
HNO2 during the early morning hours,
HNO? + hv -^ OH + NO,
The photolysis of
(3-18)
thus can become an important source of OH radicals, leading to the rapid initiation of
photochemical activity (Harris et al., 1982).
In the troposphere, the initially emitted NO is converted to NOX (NO + NO2), then
to reservoir and termination species (PAN and its homologues, organic nitrates, HNO3, and
particulate nitrate). These reservoir and termination species are referred to as NOZ. The term
"NOy" refers to the total amount of nitrogen, with NOy = (NOX + NOZ). Parrish et al. (1993)
have investigated the partitioning between the individual nitrogen-containing species at several
rural sites in the eastern United States, and Trainer et al. (1993) and Olszyna et al. (1994)
have shown that, in rural areas in the eastern United States, there is a good correlation
between the O3 levels and NOy. Trainer et al. (1993) further showed that O3 levels correlate
even better with NOZ than with NOy, as may be expected because NOZ quantifies the amount
of initially emitted NO that has been processed photochemically, forming O3 in the process.
3.2.3.3 The Methane Oxidation Cycle
Methane is emitted into the atmosphere from swamps and wetlands, as well as
from ruminants (Fung et al., 1991a; World Meteorological Organization, 1992). The major
tropospheric removal process for CH4 is by reaction with the OH radical, with the CH4
lifetime equal to
(k21 [OH])'1,
(3-19)
3-10
-------
where k21 is the rate constant for Reaction 3-21, and [OH] is the (variable) atmospheric
OH radical concentration. The calculated lifetime of CH4 in the troposphere is =10 to
12 years. As for other saturated organic compounds, the OH radical reaction with CH4
proceeds by hydrogen (H)-atom abstraction from the carbon (C)-H bonds to form the methyl
(CH3) radical:
OH + CH4 -» H2O + CH3. (3-2°)
In the troposphere, the methyl radical reacts solely with O2 to yield the methyl peroxy
(CH3O2) radical (Atkinson et al., 1992a):
CH3 + O2 H CH3O2.
In the troposphere, the methyl peroxy radical can react with NO, NO2, HO2
radicals, and other organic peroxy (RO^ radicals, with the reactions with NO and HO2
radicals being the most important (see, for example, World Meteorological Organization,
1990b). The reaction with NO leads to the formation of the methoxy (CH36) radical,
CH3O2 + NO -» CH36 + NO2. (3~22)
The reaction with the HO2 radical leads to the formation of methyl hydroperoxide
(CH3OOH),
CH3O2 + HO2 -» CH3OOH + O2, (3~23)
which can photolyze or react with the OH radical (Atkinson et al., 1992a):
CH3OOH + hv -> CH36 + OH, (3~24)
H2O + CHgOj (3-25a)
OH + CH.OOH
H2O + CHjOOH (3-25b)
I fast
HCHO + OH.
Methyl hydroperoxide also undergoes wet and dry deposition or incorporation into cloud
water. The lifetime of CH3OOH in the troposphere due to photolysis and reaction with the
3-11
-------
OH radical is calculated to be ~2 days. Methyl hydroperoxide is then a temporary sink of
radicals, with its wet or dry deposition being a tropospheric loss process for radicals.
The only important reaction for the methoxy radical in the troposphere is with
O2 to form HCHO and the HO2 radical,
CH36 + O2 -» HCHO + HO2. (3-26)
Formaldehyde is a "first-generation" product that reacts further, by photolysis:
+ H2 + CO (55%) (3-27a)
HCHO + hv
H + HCO (45%), (3-27b)
where the percentages are for overhead sun conditions (Rogers, 1990) and also by reaction
with the OH radical,
OH + HCHO -» H2O + HCO. (3-28)
In the troposphere, the H atom and HCO (formyl) radical produced in these processes react
solely with O2 to form the HO2 radical:
H + O2 + M -» HO2 + M (3-29)
HCO + O2 -» HO2 + CO. (3-30)
The lifetimes of HCHO due to photolysis and OH radical reaction are ~4 h and 1.5 days,
respectively, leading to an overall lifetime of ~3 h for overhead sun conditions.
The final step in the oxidation of CH4 in the earth's atmosphere involves the
oxidation of CO by reaction with the OH radical (the only tropospheric reaction of CO) to
form CO2:
OH + CO -> H + CO, (3-31)
2
H + O + M -» HO, + M. (3-29)
The lifetime of CO in the lower troposphere is ~2 mo.
The overall reaction sequence leading to CO2 formation, through the HCHO and
CO intermediate products, is shown in Figure 3-2.
3-12
-------
Wet/Dry Deposition -— CH3OOH
OH
CH3O
OH + CH4 + H2O + CH3
Figure 3-2. Atmospheric reactions in the complete oxidation of methane.
There is competition between NO and the HO2 radical for reaction with the CH3O2
radical, and the reaction route depends on the rate constants for these two reactions and the
tropospheric concentrations of HO2 radicals and NO. The rate constants for the reaction of
the CH3O2 radicals with NO (Reaction 3-22) and HO2 radicals (Reaction 3-23) are of
comparable magnitude (Atkinson et al., 1992a). Based on the expected HO2 radical
concentration in the troposphere, Logan et al. (1981) calculated that the reaction of the CH3O2
radical with NO dominates for NO mixing ratios of >30 ppt (equivalent to an NO
concentration of >7 x 108 molecules cm"3 in the lower troposphere). For NO mixing ratios
<30 ppt, the reaction of the CH3O2 radical with HO2 dominates.
Hydroperoxy radicals formed from, for example, Reactions 3-26, 3-29, and 3-30
can react with NO, O3, or themselves, depending mainly on the concentration of NO. The
reaction with NO leads to regeneration of the OH radical,
HO + NO -^ OH + NO
(3-32)
whereas the reactions with O3 and HO2 radicals lead to a net destruction of tropospheric O3:
HO2 + HO2 -^ H2O2 + O2
(3-33)
HO, + O,
OH + 2
(3-34)
3-13
-------
This net loss of tropospheric O3 occurs because the photolytic production of the OH radical
from O3, via the intermediary of the O(1D) atom, represents a loss process for tropospheric
O3. Hence, the absence of any O3 formation from the CH4 oxidation cycle is equivalent to a
net O3 loss. Using the rate constants reported for Reactions 3-32 and 3-34 (Atkinson et al.,
1992a) and the tropospheric O3 mixing ratios given above, it is calculated that the HO2 radical
reaction with NO dominates over reaction with O3 for NO mixing ratios >10 ppt. The rate
constant for Reaction 3-33 is such that an NO mixing ratio of this magnitude also means that
the HO2 radical reaction with NO dominates over the self-reaction of HO2 radicals.
There are therefore two regimes, depending on the fate of HO2 and CH3O2 radicals:
(1) a high-NO regime in which HO2 and CH3O2 radicals react with NO to convert NO to
NO2, regenerate the OH radical, and, through the photolysis of NO2, produce O3; and
(2) a low-NO regime in which HO2 and CH3O2 radicals combine (Reaction 3-23), and HO2
radicals undergo self-reaction and react with O3 (Reactions 3-33 and 3-34), leading to a net
destruction of O3 and inefficient OH radical regeneration (see also Ehhalt et al., 1991; Ayers
et al., 1992).
Under high-NO conditions, the oxidation of CH4 leading to the formation of
HCHO can be written as the net reaction,
OH + CH4 + 2 NO + 2 O2 = H2O + HCHO + 2 NO2 + OH, (3-35)
indicating the conversion of two molecules of NO to NO2 and regeneration of the OH radical.
Because NO2 photolyzes to form O3 in the presence of O2,
N02 + hv °2 > NO + Oj , (3-1, 3-2)
the oxidation of CH4 to HCHO under high-NO conditions can be written as
OH + CH4 + 4 O2 = H2O + HCHO + 2 O3 + OH, (3-36)
showing the formation of O3 from CH4 oxidation in the troposphere. The reaction cycles
oxidizing CH4 to HCHO, converting NO to NO2, and forming O3 are shown schematically in
Figure 3-3.
In a similar manner, under high-NO conditions, the photolysis of HCHO and its
reaction with the OH radical is given approximately by
0.2 OH + HCHO + 0.92 NO -> CO + 0.44 H2 + 0.2 H2O + 0.92 NO2 + 0.92 OH.
(3-37)
3-14
-------
HNO
Emission
Emission
CHp2
CHp
HCHO
HNO,
Figure 3-3. Cyclic reactions of methane oxidation to formaldehyde, conversion of nitric
oxide to nitrogen dioxide, and concomitant formation of ozone in the
atmosphere.
Formaldehyde photooxidation is thus a source of HO2 radicals (and of OH radicals in high-
NO conditions) (Ehhalt et al., 1991), especially in urban areas where the concentration of
HCHO is elevated because it is produced during the oxidation of anthropogenic nonmethane
VOCs (Finlayson-Pitts and Pitts, 1986).
Nitric oxide mixing ratios are sufficiently low in the lower troposphere over
marine areas that oxidation of CH4 will lead to a net destruction of O3 (low-NO conditions),
as discussed by Carroll et al. (1990) and Ayers et al. (1992). However, in the upper
troposphere and over continental areas impacted by NOX emissions from combustion sources,
NO mixing ratios are high enough (high NO-conditions) for CH4 oxidation to lead to net
O3 formation (Carroll et al., 1990; World Meteorological Organization, 1992).
3.2.3.4 Cloud Processes in the Methane-Dominated Troposphere
In addition to the dry and wet deposition of certain products of the NOx-CH4-air
photooxidation (e.g., wet and dry deposition of HNO3 and CH3OOH [Atkinson, 1988, and
references therein; Hellpointner and Gab, 1989]), cloud processes can have significant effects
on the gas-phase chemistry of the clean troposphere (Lelieveld and Crutzen, 1990, 1991;
Warneck, 1991, 1992). Lelieveld and Crutzen (1990, 1991) have postulated from modeling
studies that the uptake of HCHO, HO2 radicals, and N2O5 into clouds can decrease markedly
the production of O3. The incorporation of HCHO into cloudwater removes HCHO from the
gas phase and, hence, reduces the gas-phase formation of HO2. The uptake of HO2 radicals
into cloudwater has the same effect. Moreover, the aqueous-phase reactions of CH2(OH)2 (the
hydrated form of HCHO) lead to the formation of O^ which reacts with dissolved O3 to act as
a sink for O3 during cloudy periods. During nighttime, N2O5 formed in the gas phase from
Reactions 3-3, 3-12, and 3-13 can be readily incorporated into cloudwater with hydrolysis to
HNO3, precluding the reformation of NOX during the following day from Reactions -3-13 and
3-14. Dentener and Crutzen (1993) have also concluded from a computer modeling study that
3-15
-------
the heterogeneous reactions of NO3 radicals and N2O5 on aerosols can have significant effects
on global O3 mixing ratios and on OH concentrations by reducing NOX levels.
The net effect of these cloud processes is to reduce the gas-phase concentrations of
HCHO, NOX, HOX, and O3. Additional, but related, processes can occur in the polluted
troposphere (see, for example, Jacob et al., 1989; Dentener and Crutzen, 1993; Section 3.2.5).
3.2.4 Photochemistry of the Polluted Atmosphere
Human activities lead to the emissions of NOX (NO + NO2) and both CH4- and
nonmethane organic compounds (NMOC) into the atmosphere (Table 3-1). Methane
emissions are important on a global scale (World Meteorological Organization, 1992),
whereas nonmethane VOC emissions are most important in urban and regional areas.
In addition to the emissions of nonmethane VOCs from anthropogenic sources, large
quantities of biogenic nonmethane VOCs (mainly of isoprene [2-methyl-l, 3-butadiene] and
monoterpenes, [C10H16]) are emitted, both in polluted and nonpolluted areas, into the
atmosphere from vegetation (see, for example, Isidorov et al., 1985; Lamb et al., 1987; Arey
et al., 1991a,b).
Table 3-1. Estimated Emissions of Methane, Nonmethane
Organic Compounds, Nitrous Oxide, and Nitrogen Oxides into
the Earth's Atmosphere from Biogenic and Anthropogenic Sources3
Emissions (Tg/yearb)
Chemical
CH4C
NMOCd
N2O (as N)e
NOX (as N)f
Biogenic Sources
=150
=1,000
~7
=10
Anthropogenic Sources
=350
=100
=6
=40
aSee Appendix A for abbreviations and acronyms.
bTeragram = 1012 g, or ~W6 metric tons.
Tung et al. (1991a); World Meteorological Organization (1992). Emissions from ruminants, rice paddies, and
biomass burning are considered as anthropogenic emissions.
dLogan et al. (1981); World Meteorological Organization (1992), with biogenic emissions being assumed to be
50% isoprene and 50% monoterpenes.
ePrinn et al. (1990).
National Research Council (1991); World Meteorological Organization (1992); biogenic sources =50% from
soils and =50% from lightning.
Analogous to the photooxidation of CH4, the interaction of NOX with nonmethane
VOCs from anthropogenic and biogenic sources under the influence of sunlight leads to the
formation of photochemical air pollution (National Research Council, 1991). In urban areas,
emissions of NOX and VOCs from human activities (combustion sources, including
transportation; industrial sources; solvent usage; landfills; etc.) dominate over biogenic
sources (National Research Council, 1991; Chameides et al., 1992). However, the emissions
3-16
-------
of VOCs from vegetation have been implicated in the formation of photochemical air
pollution in urban (Chameides et al., 1988; 1992) as well as rural areas (Trainer et al., 1987;
Roselle et al., 1991; Chameides et al., 1992).
In essence, the chemistry of the polluted urban and regional atmosphere is an
extension of that of the clean, CH4-dominated troposphere, with a number of additional
complexities due to the number and types of VOCs emitted from anthropogenic and biogenic
sources. At least in certain urban areas, the NMOC content of ambient air is similar to the
composition of typical gasolines (Mayrsohn and Crabtree, 1976; Mayrsohn et al., 1977;
Harley et al., 1992; see Section 3.4.3). For example, gasolines typically consist of-55 to
65% alkanes, -5 to 10% alkenes, and -25 to 35% aromatic hydrocarbons (Lonneman et al.,
1986; Sigsby et al., 1987), whereas in Los Angeles, CA, the ambient urban air composition is
-50 to 55% alkanes, -5 to 15% alkenes, -25 to 30% aromatic hydrocarbons, and -5 to 15%
carbonyls (Grosjean and Fung, 1984; California Air Resources Board, 1992). Emissions of
NOV and VOCs are dealt with in detail in Section 3.4.1.
•"x
3.2.4.1 Tropospheric Loss Processes of Volatile Organic Compounds
The chemical loss processes of gas-phase VOCs include photolysis and chemical
reaction with the OH radical during daylight hours, reaction with the NO3 radical during
nighttime hours, and reaction with O3, which often is present throughout the 24-h period
(Atkinson, 1988).
As discussed earlier, photolysis of chemical compounds in the troposphere is
restricted to the wavelength region above -290 nm. Because of the strength of chemical
bonds, the tropospheric wavelength region in which photolysis can occur extends from
-290 to 800 nm, and this wavelength region often is referred to as the "actinic" region. For
photolysis to occur, a chemical compound must be able to absorb radiation in the actinic
region (and hence have a nonzero absorption cross-section, ox, in this wavelength region).
Having absorbed radiation, a chemical compound must then undergo chemical change (i.e.,
have a nonzero quantum yield, (|)x, for photodissociation or photoisomerization). The quantum
yield, (|)x, is defined as (number of molecules of the chemical undergoing change)/(number of
photons of light absorbed). The photolysis rate, kpho,olysis, for the process,
C + hv -^ products (3-38)
is given by
k. f, . = f 1 a,
-------
For the reaction of a VOC with a reactive species, X (for tropospheric purposes,
X = OH, NO3, and O3), the lifetime for the reaction process, C + X — » products, is given by
l = kJX]-1 (3-41)
x
and depends on the concentration of the reactive species X and the rate constant (kx) for
reaction of the VOC with X. In general, the ambient atmospheric concentrations of OH
radicals, NO3 radicals, and O3 are variable, depending on time of day, season, latitude,
altitude, etc. For the purpose of comparing lifetime calculations for various classes of VOCs,
average ambient tropospheric concentrations of these three species often are used. The
concentrations used here have been presented in the sections above and are OH radicals, a 12-
h average daytime concentration of 1.6 x io6 molecule cm"3 (equivalent to a 24-h average
concentration of 8 x io5 molecule cm"3) (Prinn et al., 1992); NO3 radicals, a 12-h nighttime
average concentration of 5 x io8 molecule cm"3 (Atkinson, 1991); and O3, a 24-h average of
7 x io11 molecule cm"3 (30 ppb) (Logan, 1985).
The major classes of VOCs are the alkanes, alkenes (including alkenes from
biogenic sources), aromatic hydrocarbons, carbonyl compounds, alcohols, and ethers (see
California Air Resources Board, 1992). The calculated lifetimes with respect to the individual
atmospheric loss processes of compounds representing a range of reactivities in each class are
given in Table 3-2. Note that the lifetimes given are dependent on the reaction rate constants
and the assumed ambient concentrations of OH radicals, NO3 radicals, and O3. Uncertainties
in the ambient concentrations of the reactive species translate directly into corresponding
uncertainties in the lifetimes.
The following brief discussions of the tropospheric chemistry of the important
classes of VOCs are based on the recent review and evaluation article of Atkinson (1994),
which should be consulted for further details of the tropospheric reactions of VOCs.
Alkanes
Because gasoline and diesel fuels contain alkanes of carbon number C4 to >C15, a
large number of alkanes are present in ambient air (see, for example, Grosjean and Fung,
1984; California Air Resources Board, 1992; Section 3.4). Table 3-2 shows that the only
important tropospheric loss process for the alkanes is by reaction with the OH radical, with
calculated lifetimes of the C3 to C10 alkanes ranging from ~1 to 15 days. As for methane, the
OH radical reaction proceeds by H-atom abstraction from the various C-H bonds. The
nighttime reactions of the NO3 radical with alkanes (calculated to be generally of minor
importance, but see Penkett et al. [1993]) also proceed by initial H-atom abstraction. For an
alkane (RH), the initially formed radical is an alkyl radical (R),
OH + RH ^ H2O + R, (3-42)
3-18
-------
Table 3-2. Calculated Tropospheric Lifetimes of Selected
Volatile Nonmethane Organic Compounds Due to Photolysis
and Reaction with Hydroxyl and Nitrate Radicals and with Ozone3
Organic
w-Butane
2-Methylbutane
w-Octane
Ethane
Propene
Isoprene
Limonene
Benzene
Toluene
w-Xylene
Formaldehyde
Acetaldehyde
Acetone
2-Butanone
Methanol
Ethanol
Methyl ^-butyl ether
Ethyl f-butyl ether
Methylglyoxal
OH
5.7 days
3.7 days
1.7 days
1.7 days
6.6 h
1.7 h
l.Oh
12 days
2.4 days
7.4 h
1.5 days
11 h
66 days
13 days
15 days
4.4 days
4.9 days
1.6 days
10 h
Lifetime Due
NO3
2.8 years
290 days
250 days
230 days
4.9 days
0.8 h
3 min
>4 years
1.9 years
200 days
80 days
17 days
—
—
>77 days
>50 days
—
—
—
to Reaction with
03
>4,500 years
>4,500 years
>4,500 years
10 days
1.6 days
1.3 days
2.0 h
>4.5 years
>4.5 years
>4.5 years
>4.5 years
>4.5 years
>4.5 years
>4.5 years
—
—
—
—
>4.5 years
hv
4h
6 days
60 days
2h
aSee Appendix A for abbreviations and acronyms.
Sources: Lifetimes resulting from reaction with OH, NO3, and O3 were calculated using rate constants given in
Atkinson and Carter (1984) and Atkinson (1989, 1991, 1994); data for photolysis lifetimes are from Horowitz
and Calvert (1982), Meyrahn et al. (1982, 1986), Plum et al. (1983), and Rogers (1990). The OH radical, NO3
radical, and O3 concentrations used (molecule cm"3 were: OH, 12-h average of 1.6 x 106; NO3, 12-h average of 5
x 108; O3, 24-h average of 7 x lo11.
which rapidly adds O2 to form an alkyl peroxy radical (RO2),
R + O2 RO2,
(3-43)
3-19
-------
with the simplest of the RO2 radicals being the methylperoxy radical, described in
Section 3.2.3.3 dealing with methane oxidation. Alkyl peroxy radicals can react with NO,
NO2, and HO2 radicals, and other organic peroxy radicals (R'O2):
RO2 + NO -» RO + NO2 (3-44a)
RO; + NO2 * ROONO2 (3 -45)
RO2 + HO2 -» ROOH + O2 (3 -46)
RO; + R'O;-^ RO + R'6 + O, (3-47a)
RO; + R'CT2-» a carbonyl + (1 - a) ROH + O2 (3-47b)
+ products of R'O'2 .
The reactions with organic peroxy radicals are expected to be of less importance in the
troposphere than the other reactions listed. Because low NO conditions occur even in air
masses in urban areas, the HO2 radical reactions with RO2 radicals and the subsequent
chemistry must be considered. However, because of space constraints and a general lack of
knowledge concerning the tropospheric chemistry of RO2 radicals under low-NO conditions,
only the reactions occurring under high-NO conditions are presented and discussed here. For
the >C3 alkyl peroxy radicals, in addition to the reaction pathway leading to NO-to-NO2
conversion (Reaction 3-44a), a second reaction pathway leading to formation of an alkyl
nitrate becomes important:
+ NO > RONO2. (3-44b)
For a given alkyl peroxy radical, the alkyl nitrate yield increases with increasing pressure and
with decreasing temperature (Carter and Atkinson, 1989a).
Analogous to the case for the methoxy radical, those alkoxy radicals (ROj) formed
from the higher alkanes that have an abstractable H atom can react with O2 to form the HO2
radical and a carbonyl; for example,
(CH3)2CHO + O2 -> CH3C(O)CH3 + HO2. (3-48)
In addition, unimolecular decomposition by C-C bond scission and unimolecular isomerization
via a six-member transition state (Atkinson and Carter, 1991; Atkinson, 1994) can be
important for the larger alkoxy radicals. For example, the following chemistry can occur for
the 1-pentoxy radical:
3-20
-------
CH3CH2CH2CH2CH2O
Decomposition Q2
Isomerization
CH3CHCH2CH2CH2OH,
HCHO HO,
with the alkyl radicals C4H9 and HOCH2CH2CH2 CHCH3 undergoing further reaction.
The majority of the reaction rate constants and reaction pathways in the alkane
degradation schemes are arrived at by analogy from the chemistry of CrC3 alkyl, alkyl
peroxy, and alkoxy radicals (Atkinson, 1990, 1994; Carter, 1990; Atkinson et al., 1992a).
A number of areas of uncertainty still exist for the tropospheric chemistry of the alkanes
(Atkinson, 1993). These include the relative importance of alkoxy radical reaction with O2,
decomposition and isomerization, and the reactions occurring subsequent to the isomerization
reaction; the formation of alkyl nitrates from the reactions of the peroxy radicals with NO;
and reactions of the alkyl peroxy radicals with HO2 and other peroxy radicals, reactions that
can be important in the nonurban troposphere.
Alkenes (Anthropogenic and Biogenic)
The alkenes emitted from anthropogenic sources are mainly ethene, propene, and
the butenes, with lesser amounts of the >C5 alkenes. The major biogenic alkenes emitted
from vegetation are isoprene (2-methyl-1,3-butadiene) and C10H16 monoterpenes (Isidorov
et al., 1985; Winer et al., 1992), and their tropospheric chemistry is currently the focus of
much attention (see, for example, Hatakeyama et al., 1989, 1991; Arey et al., 1990; Tuazon
and Atkinson, 1990a; Pandis et al., 1991; Paulson et al., 1992a,b; Paulson and Seinfeld,
1992a; Zhang et al., 1992; Hakola et al., 1993, 1994).
As evident from Table 3-2, the alkenes react with OH and NO3 radicals and O3.
All three processes are important atmospheric transformation processes, and all three reactions
proceed by initial addition to the >C=C< bonds. These reactions are briefly discussed below.
Hydroxyl Radical Reactions. As noted above, the OH radical reactions with the
alkenes proceed mainly by OH radical addition to the >C=C< bonds. For example, the OH
radical reaction with propene leads to the formation of the two OH-containing radicals,
M
OH + CH3CH = CH2 ^> CH3CH(OH)CH2 + CH3CHCH2OH. (3-49)
The subsequent reactions of these radicals are similar to those of the alkyl radicals formed by
H-atom abstraction from the alkanes. Taking the CH3CHCH2OH radical as an example,
under high-NO conditions, the following chemistry occurs:
3-21
-------
CH3CHCH2OH + O2-^ CH3CH(OO)CH2OH
NO
CH?CH(ONO2)CH2OH
CH3CH(O)CH2OH + NO2
Decomposition
CH2OH + CH3CHQ CH?C(O)CH?OH + HO2.
|02
HCHO + HO2
The underlined species represent products that, although stable, can undergo further reaction;
and, hence, they can lead to "second-generation" products. For the simple
-------
cleavage of the >C=C< bonds, it is possible that the yields reported for these carbonyls
included contributions by other, as yet unidentified, carbonyl-containing products.
Nitrate Radical Reactions. The NO3 radical reactions proceed by reaction schemes
generally similar to the OH radical reactions, except that, when NO3 radicals are present, NO
concentrations are low (see above) and RO2 + RO2 and RO2 + HO2 radical reactions are
expected to dominate over RO2 + NO reactions. For propene, the initial reaction is
NO3 + CH3CH = CH2 -> CH3CHCH2ONO2
and CFLCH(ONO7)CFL
(3-50)
followed by a series of reactions that are expected (Atkinson, 1991) to lead to the formation
of, among others, carbonyls and nitrato-carbonyls (for example, HCHO, CH3CHO,
CH3CH(ONO2)CHO, and CH3C(O)CH2ONO2 from propene). Few data are presently available
concerning the products and detailed mechanisms of NO3-alkene reactions (Atkinson, 1991,
1994, and references therein). In particular, the reaction products and mechanisms for the
NO3 radical reactions with isoprene and the monoterpenes are still not quantitatively
understood (Kotzias et al., 1989; Barnes et al., 1990; Hjorth et al., 1990; Skov et al., 1992).
Ozone Reactions. The O3 reactions also proceed by addition of O3 to the alkene,
to form an energy-rich ozonide that rapidly decomposes to form carbonyls and energy-rich
biradicals ([ ]*):
O3 + CH3CH=CH2
o
o
CH3CH CH2
CH3CHO
HCHO + [CH3CHOO] .
The energy-rich biradicals, [CH2OO]* and [CH3CHOO]*, undergo collisional stabilization or
decomposition:
M
[R^COO]
->• Decomposition.
(3-51a)
(3-51b)
3-23
-------
There are still significant uncertainties concerning the reactions of the energy-rich biradicals
(Horie and Moortgat, 1991; Atkinson, 1990, 1994), with recent studies showing the
production of OH radicals in high yields for several alkenes (Niki et al., 1987; Paulson et al.,
1992a; Atkinson et al., 1992b; Paulson and Seinfeld, 1992b; Atkinson and Aschmann, 1993).
For isoprene, the major products are methacrolein and methyl vinyl ketone
(Kamens et al., 1982; Niki et al., 1983; Paulson et al., 1992b). Paulson et al. (1992b) derived
an OH radical and an O(3P) atom formation yield of 0.68 ±0.15 and 0.45 ± 0.2, respectively,
from the O3 reaction with isoprene, indicating the dominance of secondary reactions.
However, Atkinson et al. (1992b) derived a significantly lower OH radical formation yield of
0.27 (uncertain to a factor of ~1.5). Clearly, further studies of this important reaction are
needed.
The only quantitative studies of the gas-phase O3 reactions with the monoterpenes
are those of Hatakeyama et al. (1989) for a- and (3-pinene and Hakola et al. (1993, 1994) for
a series of monoterpenes. Additionally, Atkinson et al. (1992b) derived OH radical formation
yields from these reactions under atmospheric conditions.
Several groups (Gab et al., 1985; Becker et al., 1990, 1993; Simonaitis et al., 1991;
Hewitt and Kok, 1991) have reported the formation of H2O2 and organic peroxides from
O3 reactions with alkenes. However, there are significant disagreements in the quantitative
results reported by Becker et al. (1990, 1993) and Simonaitis et al. (1991).
Aromatic Hydrocarbons
The chemistry of aromatic hydrocarbons is one of the major sources of uncertainty
in the atmospheric chemistry of VOCs (National Research Council, 1991; Atkinson, 1994).
The most abundant aromatic hydrocarbons in urban atmospheres are benzene, toluene, the
xylenes, and the trimethylbenzenes (Grosjean and Fung, 1984; California Air Resources
Board, 1992). As shown in Table 3-2, the only tropospherically important loss process for
benzene and the alkyl-substituted benzenes is by reaction with the OH radical. For the alkyl-
substituted benzenes, the OH radical reactions proceed by two pathways: (1) H-atom
abstraction from the C-H bonds of the alkyl substituent groups and (2) OH radical addition to
the aromatic ring, as shown below, for/?-xylene:
OH +
(3-52a)
(3-52b)
with the OH radical addition pathway being reversible above -325 K (Atkinson, 1989).
3-24
-------
The radical formed in Reaction 3-52a reacts analogously to an alkyl radical
(Atkinson, 1994), leading, in the presence of NO, to aromatic aldehydes and organic nitrates.
CH3C6H4CH2 + O2 =-£ CH3C6H4CH2OO
NO
CH3C6H4CH2QNO2
CH3C6H4CH2O + NO2
CH3C6H4CHQ + HO2
(p-tolualdehyde)
The OH-containing radical formed in Reaction 3-52b can undergo reaction with both NO2 and
O2. Knispel et al. (1990) reported rate constants for the reactions of NO2 and O2 with the
OH-containing radicals formed from benzene and toluene. The magnitude of the rate
constants they obtained implies that, in the troposphere, the major reactions of these radicals
will be with O2. Laboratory studies of the formation of selected products from the gas-phase
reactions of the OH radical with toluene, o-xylene, and 1,2,3-trimethylbenzene (Atkinson and
Aschmann, 1994), and, in particular, of the formation of the ring-cleavage product
2,3-butanedione from o-xylene, are consistent with the kinetic data of Knispel et al. (1990).
Thus, at least for the monocyclic aromatic hydrocarbons such as benzene, toluene, and the
xylenes, the OH-aromatic adducts formed in Reaction 3-52b react with O2 under atmospheric
conditions. However, care must be taken in using product yields obtained in the laboratory at
higher than ambient NO2 concentrations because data may be influenced by the NO2 reaction
and, hence, may not be applicable to the O2 reaction with the OH-aromatic adduct. Clearly,
the products formed and their yields from the O2 and NO2 reactions with the OH-aromatic
adducts need to be determined, and the detailed reaction mechanisms elucidated.
Despite these uncertainties, however, products from the OH radical addition
pathway have been identified, and their formation yields determined (Atkinson, 1994, and
references therein). The major products identified from the OH radical addition pathway are
phenolic compounds (e.g., phenol from benzene and o-, TO-, and />-cresol from toluene) and
oc-dicarbonyls (glyoxal, methylglyoxal, and 2,3-butanedione) resulting from the cleavage of
the aromatic ring (see, for example, Atkinson, 1990, 1994, and references therein).
Significant fractions (>50% for benzene, toluene, and the xylenes) of the reaction products
are, however, still not accounted for.
3-25
-------
Carbonyl Compounds
As noted above, the OH radical reactions with the alkanes, alkenes, and aromatic
hydrocarbons lead, often in large yield, to the formation of carbonyl compounds. Likewise,
carbonyls are formed during the reactions of NO3 radicals and O3 with alkenes. As a first
approximation, the carbonyl compounds of tropospheric interest are HCHO (see
Section 3.2.3.3), acetaldehyde, and the higher aliphatic aldehydes; benzaldehyde; acetone,
2-butanone, and the higher ketones; and simple dicarbonyls such as glyoxal, methylglyoxal,
and 2,3-butanedione.
The tropospheric photooxidation of isoprene leads to the formation of methyl vinyl
ketone (CH3C(O)CH=CH2) and methacrolein (CH3C(CHO)=CH2). The OH radical-initiated
reactions of these two carbonyl compounds in the presence of NOX have been studied by
Tuazon and Atkinson (1989, 1990b).
The tropospherically important loss processes of the carbonyls not containing
>C=C< bonds are photolysis and reaction with the OH radical. As shown in Tables 3-2,
photolysis is a major tropospheric loss process for the simplest aldehyde (HCHO) and the
simplest ketone (CH3C(O)CH3), as well as for the dicarbonyls. For the higher aldehydes and
ketones, the OH radical reactions are calculated to be the dominant gas-phase loss process
(Table 3-2). For CH3CHO, the reaction proceeds by H-atom abstraction from the -CHO
group to form the acetyl (CH3CO) radical,
OH + CH3CHO -» H2O + CH3CO, (3~53)
which rapidly adds O2 to form the acetyl peroxy radical:
CH3CO + O2 H CH3C(O)OO. (3-54)
This O2 addition pathway is in contrast to the reaction of O2 with the formyl (HCO) radical
formed from HCHO, which reacts by an H-atom abstraction pathway (Reaction 3-30). The
acetyl peroxy radical reacts with NO and NO2,
CH3C(O)OO + NO-^ CH3C(O)6 + NO2 (3-55)
fast
r
CH3 + CO2
M
CH3C(O)OO + NO2 * CH3C(O)OONO2, (3-56, -3-56)
with the NO2 reaction forming the thermally unstable PAN. The higher aldehydes also lead
to PANs (Roberts, 1990); for example, propionaldehyde reactions lead to the formation of
3-26
-------
peroxypropionyl nitrate (PPN). Although the rate constant at atmospheric pressure for the
thermal decomposition of PAN (Atkinson et al., 1992a) is such that the lifetime of PAN, with
respect to thermal decomposition, is -30 min at 298 K in the lower troposphere, the thermal
lifetime of PAN is calculated to be several hundred years in the upper troposphere. Reaction
with OH radicals or photolysis, or both, therefore will dominate as the loss processes of
PANs in the upper troposphere (Atkinson et al., 1992a).
The transport of PAN out of urban areas into colder air masses (e.g., to higher
altitude) leads to PAN becoming a temporary reservoir of NOX, allowing the long-range
transport of NOX to less polluted areas. Release of NO2 in these less polluted areas via
Reaction -3-56, with subsequent photolysis of NO2, then leads to O3 formation and the
pollution of remote areas.
Because the CH3 radical formed from the NO reaction with CH3C(O)OO leads to
HCHO formation, the OH radical reaction with CH3CHO subsequently leads to the formation
of HCHO. The same process occurs for propionaldehyde, which reacts to form CH3CHO and
then HCHO. Benzaldehyde appears to behave as a phenyl-substituted aldehyde, with respect
to its OH radical reaction, and the analog to PAN is then peroxybenzoyl nitrate, PBzN
(C6H5C(O)OONO2).
The formation of HCHO from CH3CHO and of CH3CHO and then HCHO from
propionaldehyde are examples of "cascading", in which the photochemical degradation of
emitted VOCs leads to the formation of further VOCs, typically containing fewer carbon
atoms than the precursor VOC. This process continues until the degradation products are
removed by wet and dry deposition or until CO or CO2 are the degradation products. The
reactions of each of these VOCs (i.e., the initially emitted VOC and its first-, second-, and
successive-generation products), in the presence of high concentrations of NO, can lead to the
formation of O3.
As discussed in Section 3.2.3.3 for HCHO, the photolysis of carbonyl compounds
can lead to the formation of new radicals that result in enhanced photochemical activity. The
OH radical reactions of the ketones are generally analogous to the reaction schemes for the
alkanes and aldehydes.
Alcohols and Ethers
A number of alcohols and ethers are used in gasolines and in alternative fuels.
The alcohols include methanol, ethanol, and tert-butyl alcohol, and the ethers include methyl
tert-butyl ether (MTBE) and ethyl tert-butyl ether (ETBE). Table 3-2 shows that in the
troposphere these VOCs react only with the OH radical. The relatively long calculated
lifetime of methanol in the troposphere (15 days), due to reaction with the OH radical (Table
3-2), suggests that methanol also will be removed from the troposphere by wet and dry
deposition and that these physical loss processes may dominate the OH radical reactions
preceded by H-atom abstraction from the C-H bonds (and to a minor extent from the O-H
bonds in the alcohols), for example, for methanol,
OH + CH3OH -^ H2O + CH3O (15%) (3-57a)
OH + CH3OH -* H2O + CH2OH. (85%) (3-57b)
3-27
-------
In the troposphere, both of the CH3O and CH2OH radicals react only with O2 to form HCHO.
OLO + O, -» HCHO + HO, (3-58)
CH2OH + O2 -> HCHO + HO2 (3-59)
The overall reaction is then
OH + CH3OH + O2 -» H2O + HCHO + HO2. (3-60)
The reaction sequence for ethanol is similar (Atkinson, 1994). Product studies of the OH
radical-initiated reactions of MTBE and ETBE in the presence of NOX have been carried out
by Taper et al. (1990), Smith et al. (199la, 1992), Tuazon et al. (1991), and Wallington and
Japar (1991). The major products from MTBE are fert-butyl formate, HCHO, and methyl
acetate [CH3C(O)OCH3] and, from ETBE, tert-butyl formate, fert-butyl acetate, HCHO,
CH3CHO, and ethyl acetate. The available product data and the reaction mechanisms have
been reviewed by Atkinson (1994), and that reference should be consulted for further details.
In addition to the use of alcohols and ethers in gasolines and alternative fuels,
unsaturated alcohols have been reported as emissions from vegetation (Arey et al., 1991a;
Goldan et al., 1993), and kinetic and product studies have begun to be reported for these
biogenic VOCs (Grosjean et al., 1993a).
Primary Products and Areas of Uncertainty for the Tropospheric Degradation Reactions
of Volatile Organic Compounds
The tropospheric degradation reactions of the alkanes, alkenes (including those of
biogenic origin), aromatic hydrocarbons, carbonyls (often formed as products of the
degradation reactions of alkanes, alkenes, and aromatic hydrocarbons), and other oxygenates
have been briefly discussed above. A more lengthy and detailed discussion of the
atmospheric chemistry of alkanes, alkenes, aromatic hydrocarbons, and oxygen- and nitrogen-
containing organic compounds emitted into the atmosphere from anthropogenic and biogenic
sources and of their atmospheric transformation products is given in the review and evaluation
of Atkinson (1994), which also provides an assessment of the uncertainties in the product
yields and the reaction rate constants. The first-generation products of the alkanes, alkenes,
and aromatic hydrocarbons follow (unfortunately, complete product distributions have not
been obtained for most of the VOCs studied).
Alkanes.
• Carbonyl compounds (i.e., aldehydes and ketones) are formed as major products for the
smaller (C3 alkanes studied to date. The yields increase with
the size of the alkane from ~4% for propane to -30% for w-octane.
• 8-Hydroxycarbonyls are expected to be formed after the alkoxy radical isomerization
reaction. To date, no direct evidence for the formation of these compounds exists. For the
larger alkanes, the formation yields of these compounds could be high.
3-28
-------
• Alkyl hydroperoxides are formed under low-NO conditions.
• Alkyl peroxynitrates (ROONO2) are formed but have short lifetimes (a few seconds at
298 K) due to thermal decomposition.
• Alcohols are formed from the combination reactions of the peroxy radicals under low-NO
experimental conditions. These compounds are expected to be formed in low
concentrations in the troposphere.
The major uncertainties in the atmospheric chemistry of the alkanes concern the
formation of RONO2 from the reactions of the peroxy radicals with NO (Reaction 3-44b) and
the reactions of the alkoxy radicals in the troposphere. These uncertainties affect the amount
of NO to NO2 conversion occurring and, hence, the amounts of O3 that are formed during the
NOx-air photooxidations of the alkanes.
Alkenes.
• Carbonyl compounds (aldehydes and ketones) are formed as major products of the OH
radical, NO3 radical, and O3 reactions.
• Organic acids are formed from the O3 reactions, but probably in low yields.
• Hydroxynitrates and nitratocarbonyls are formed from the OH radical reactions and NO3
radical reactions, respectively. The hydroxynitrates are formed in low yields from the OH
radical reactions, whereas the nitratocarbonyls may be major products of the NO3 radical
reactions.
• Hydroxycarbonyls and carbonyl-acids are also expected to be formed, although few, if any,
data exist to date.
• Decomposition products are produced from the initially energy-rich biradicals formed in
the O3 reactions; these include CO, CO2, esters, hydroperoxides, and, in the presence of
NOX, peroxyacyl nitrates (RC(O)OONO2 and PANs).
The major areas of uncertainty concern the products and mechanisms of the
O3 reactions (in particular, the radical yields from these reactions that affect the O3 formation
yields from the NOx-air photooxidations of the alkenes) and the reaction products and
mechanisms of the OH radical reactions with the alkenes containing more than four carbon
atoms.
Aromatic Hydrocarbons.
• Phenolic compounds, such as phenol and cresols, have been shown to be major products of
the atmospheric reactions of the aromatic hydrocarbons under laboratory conditions.
• Aromatic aldehydes, such as benzaldehyde, are formed in <10% yield.
• oc-Dicarbonyls, such as glyoxal, methylglyoxal, and biacetyl, are formed in fairly high
(10 to 40%) yields. These dicarbonyls photolyze rapidly to form radicals and, therefore,
are important products with respect to the photochemical activity of the aromatic
hydrocarbons.
• Unsaturated carbonyl or hydroxycarbonyl compounds are formed, although there is little
direct information concerning the formation of these products.
There is a lack of knowledge of the detailed reaction mechanisms and products for
the aromatic hydrocarbons under tropospheric conditions (i.e., for the NOX concentrations
encountered in urban and rural areas). It is possible that the products observed in laboratory
studies and their formation yields are not representative of the situation in the troposphere.
This then leads to an inability to formulate detailed reaction mechanisms for the atmospheric
3-29
-------
degradation reactions of the aromatic hydrocarbons, and the chemical mechanisms used in
urban airshed models then must rely heavily on environmental (or "smog") chamber data.
Oxygenated Compounds.
• The products observed from the atmospheric photooxidations of oxygenated organics are
carbonyls, organic acids (e.g., RC(O)OH), esters, alcohols, and, in the presence of NOX,
PANs.
The major area of uncertainty concerns the importance of photolysis of carbonyl
compounds in the troposphere, and the products formed. In particular, there is a lack of
information concerning the absorption cross-sections and photoodissociation quantum yields
for most of the aldehydes and ketones other than HCHO, CH3CHO, and CH3COCH3.
3.2.4.2 Chemical Formation of Ozone in Polluted Air
Major Steps in Ozone Formation
As discussed earlier, NOX and VOCs interact under the influence of sunlight to
form O3 and other photochemical air pollutants. The major steps in this process are the
conversion of NO to NO2 by peroxy radicals, with the photolysis of NO2 leading to
O3 production. In the absence of a VOC, Reactions 3-1 through 3-3,
NO + O3 -> NO2 + O2 (3-3)
NO2 + hv -4 NO + O3, (3-1, 3-2)
do not lead to any net formation of O3. The reaction of a VOC with the OH radical, or its
photolysis, leads to the formation of HO2 and organic peroxy (RO2) radicals, which react with
NO under high-NO conditions:
VOC(+ OH, hv) -4 RO2 (3-61)
RO2 + NO -» RO + NO2 (3-44a)
NO2 + hv ^ NO + O3 (3-1, 3-2)
Net: VOC(+ OH, hv) -4 RO2 + O3. (3-62)
with the alkoxy (RO) radical producing further HO2 or RO2 radicals or both, and, hence,
resulting in further production of O3. This process is shown schematically in Figure 3-4.
3-30
-------
VOC
RO
Figure 3-4. Major steps in production of ozone in ambient air (R = H, alkyl or substituted
alkyl, or acyl).
The general time-concentration profiles for selected species during irradiation of an
NOx-VOC-air mixture are shown in Figure 3-5 for a constant light intensity and in Figure 3-6
for diurnally varying light intensity. These general features of an NMOC-NOx-air irradiation
are described by the reactions described below.
The conversion of NO to NO2 occurs through the oxidation reactions:
organic(+ OH, hv, O3) -> RO2
+ NO -» RO + NO,
(3-63)
(3-44a)
RO -? carbonyl + HO,
(3-64)
HO, + NO -> OH + NO,.
(3-32)
The maximum concentration of NO2 is less than the initial NO
concentration because NO2 is removed through the reaction
NO,
M
OH + NO2 -» HNO3.
(3-15)
This reaction removes radicals (OH and NO2) and NOX from the system. In addition to the
removal of NOX through Reaction 3-15, NOX also can be removed through the formation of
temporary reservoir species such as organic nitrates (Reaction 3-44b) and PAN
(Reaction 3-56).
3-31
-------
0.60 ,
0.50-
0.00 i
O Ozone
D NO2
X PAN
A NO
O Propene
1.0
2.0 3.0
Time (hours)
4.0
5.0
6.0
Figure 3-5. Time-concentration profiles for selected species during irradiations of a
nitrogen oxide-propene-air mixture in an indoor chamber with constant light
intensity.
0.80.
0.60-
O Ozone A NO
D NO2+PAN O Propene
X PAN
8.0
10.0 12.0 14.0
Time (hours)
16.0
18.0
Figure 3-6. Time-concentration profiles for selected species during irradiations of a
nitrogen oxide-propene-air mixture in an outdoor chamber with diurnally
varying light intensity.
3-32
-------
RO2 + NO -> RONO2 (3-44b)
CH3C(O)OO + NO2 -» CH3C(O)OONO2 (3-56)
The O3 concentration increases with the NO2/NO concentration ratio, and
O3 formation ceases when NO2 (and hence NOX) has been removed by reaction.
Formation of PAN occurs by Reaction 3-56. Because of Reactions 3-55 and 3-56,
the PAN concentration also increases with the NO2/NO concentration ratio, and PAN
formation also ceases when NOX has been depleted.
The removal processes for NOX are by reaction of NO2 with the OH radical to
form HNO3 (Reaction 3-16), the formation of organic nitrates from the ROO + NO Reaction
3-44b pathway, and the formation of PAN through Reaction 3-56. The initially present NOX
is converted to organic nitrates, HNO3, and thermally unstable PANs. At ambient
temperature, the PANs will gradually thermally decompose to yield NO2 and the acylperoxy
radicals; hence, the ultimate fate of NOX will be to form HNO3 and organic nitrates.
Effects of Varying Initial Nitrogen Oxide and Nonmethane Volatile Organic Compound
Concentrations
As discussed in Section 3.2.4.2, NOX and VOC interact in sunlight to form O3 and
other photochemical air pollutants. The formation of O3 from the NOX and VOC precursors is
nonlinear with respect to the precursor emissions (or ambient concentrations). As discussed
in detail in Section 3.6, computer models incorporating emissions, meteorology, and chemistry
are necessary for a full understanding of the complexities of the NOX-VOC-O3 system. The
major reactions in irradiated VOC-NOx-air mixtures (see Section 3.2.4.2) include
OH + VOC °2 RO2, (3-61)
which is in competition with Reaction 3-15,
OH + NO2 -> HNO3, (3-15)
which removes both radicals and NOX.
RO2 + NO -» RO + NO2 (3-44a)
NO2 + hv °2 NO + O3 (3-1, 3-2)
NO + O3 -» NO, + O, (3-3)
Based on these reactions, as the VOC/NOX ratio decreases, Reaction 3-15 competes more
successfully with Reaction 3-61 in removing radicals from the system for a constant source of
OH radicals, and, thus, slows down the formation of O3 (and vice-versa). The ultimate
amount of O3 formed depends on the NO2 available for photochemical processing through
3-33
-------
Reactions 3-44a, 3-1, 3-2, and 3-3. Because of the diurnal light intensity variation and the
length of daylight available for reactions and photolysis to occur, O3 formation from VOC-
NOx-air mixtures, especially as a function of the VOC/NOX ratio, is complex and depends on
the time or distance scale being considered (for example, transport of urban plumes to rural
areas) (Wolff, 1993; Finlayson-Pitts and Pitts, 1993). In general, reducing the VOC/NOX ratio
by reducing VOC emissions slows down the formation rate of O3, leading to lower O3 levels
in urban areas and in areas downwind of urban complexes. Reduction of NOX emissions
leads to a more rapid formation of O3, although with less O3 formed, and, hence, inner-urban
areas may experience higher O3 levels with NOX control than with VOC control, whereas
suburban and rural areas may have lower O3 levels with NOX control (Trainer et al., 1993;
Olszyna et al., 1994).
Effects of Biogenic Nonmethane Volatile Organic Compound Emissions
Biogenic VOC emissions can be important in urban and rural areas (Trainer et al.,
1987; Chameides et al., 1988, 1992; Roselle et al., 1991) and can contribute to O3 formation
in much the same way as anthropogenic VOCs. Modeling simulations in which urban
biogenic VOC emissions are first included and then excluded from the calculations generally
indicate little effect of the biogenic emissions on the predicted O3 levels. This is not
unexpected from the shape of the O3 isopleths at high VOC/NOX ratios (Chameides et al.,
1988; Section 3.6). However, results of modeling studies in which anthropogenic VOC
emissions are removed from the simulations (but anthropogenic NOX emissions are left
unaltered) suggest that anthropogenic NOX together with biogenic VOCs may form sufficient
O3 to exceed the NAAQS, at least in certain areas (Chameides et al., 1988). Thus, for the
anthropogenic and biogenic VOC emissions considered by Chameides et al. (1988) for
Atlanta, GA, changes in either the anthropogenic VOC emissions or biogenic VOC emissions
have little effect on O3 levels. Therefore, as discussed for the Atlanta region (Chameides
et al., 1988), NOX control may be more favorable than VOC control in urban areas with
substantial biogenic NMOC emissions.
Although it is known that isoprene is reactive with respect to the formation of
O3 (Section 3.2.4.3) and that the monoterpenes react rapidly with OH radicals, NO3 radicals,
and O3, the O3-forming potentials of the various monoterpenes emitted into the atmosphere
are not known.
3.2.4.3 Hydrocarbon Reactivity with Respect to Ozone Formation
As discussed in Section 3.2.4, VOCs are removed and transformed in the
troposphere by photolysis and by chemical reaction with OH radicals, NO3 radicals, and O3.
In the presence of sunlight, the degradation reactions of the VOCs lead to the conversion of
NO to NO2 and the formation of O3 and various organic products. However, different VOCs
react at differing rates in the troposphere because of their differing tropospheric lifetimes
(Table 3-2). The lifetimes of most VOCs with respect to reaction with OH radicals and
O3 are in the range ~1 h to ~10 years. In large part, because of these differing tropospheric
lifetimes and rates of reaction, VOCs exhibit a range of reactivities with respect to the
formation of O3 (Altshuller and Bufalini, 1971, and references therein).
A number of "reactivity scales" have been developed over the years (see, for
example, Altshuller and Bufalini, 1971, and references therein; Darnall et al., 1976), including
the rate of VOC disappearance in NOx-VOC-air irradiations, the rate of NO to NO2
conversion in NOx-VOC-air irradiations, O3 formation in NOx-single VOC-air irradiations, eye
3-34
-------
irritation, and the rate constant for reaction of the VOC with the OH radical. It appears,
however, that a useful definition of "reactivity" is that of "incremental reactivity" (IR),
defined as the amount of O3 formed per unit of VOC added or subtracted from the VOC
mixture in a given air mass under high-NO conditions (Carter and Atkinson, 1987, 1989b):
IR = A[OJ/A[VOC], (3-65)
at the limit of A[VOC] —» 0. The concept of incremental reactivity and some further details
of this approach are illustrated by the general reaction mechanism for the photooxidation of
an alkane, RH:
OH + RH -» H2O + R (3-42)
R + O2 -» RO2 (3-43)
RO2 + NO -» RO + NO2 (3-44a)
RO -» carbonyl + HO2 (3-65)
HO2 + NO -» OH + NO2. (3-32)
The net reaction,
OH + RH + 2 NO -4 carbonyl + 2 NO2 + OH, (3-66)
can be viewed as involving the two separate reaction sequences:
(1) the formation of organic peroxy (RO2) radicals from the reactions,
OH + RH -* H2O + R
R + O2 -+ RO2
Net:
3-35
-------
and (2) the conversion of NO to NO2 and the formation of O3 and other products,
RO2 + NO -» RO + NO2
RO -»> carbonyl + HO2
HO2 + NO -* OH + NO2
Net: RO2 + 2 NO -»• carbonyl + 2 NO2 + OH
The photolysis of NO2 then leads to O3 formation (Reactions 3-1 and 3-2). The first reaction
sequence determines how fast RO2 radicals are generated from the VOC, which is called
"kinetic reactivity" (Carter and Atkinson, 1989b). For the case given above, where the only
reaction of the VOC is with the OH radical, the kinetic reactivity depends solely on the OH
radical reaction rate constant. The second reaction sequence, leading to NO to NO2
conversion, regeneration of OH radicals, and the formation of product species, determines the
efficiency of formation of O3 from the RO2 radicals formed from the first reaction sequence
and is termed "mechanistic reactivity" (Carter and Atkinson, 1989b). The second reaction
sequence can be represented as
RO2 + ocNO -> (3NO2 + yOH + 8 products. (3'67)
In general, the faster a VOC reacts in the atmosphere, the higher the incremental
reactivity. However, the chemistry subsequent to the initial reaction does affect the
O3-forming potential of the VOC. Thus, the existence of NOX sinks in the reaction
mechanism (low values of (3 or values of a-(3 > 0) leads to a decrease in the amount of
O3 formed. Examples of NOX sinks are the formation of organic nitrates and PANs (which
are also sinks for radicals). The generation or loss of radical species can lead to a net
formation or net loss of OH radicals (y > 1 or y < 1, respectively). This, in turn, leads to an
enhancement or suppression of radical concentrations in the air parcel and to an enhancement
or suppression of the overall reactivity of all VOCs in that air parcel by affecting the rate of
formation of RO2 radicals.
These effects vary in importance depending on the VOC/NOX ratio. Nitrogen
oxides sinks are most important at high VOC/NOX ratios (NOx-limited), affecting the
maximum O3 formed; although the formation or loss of OH radicals is most important at low
VOC/NOX ratios, affecting the initial rate at which O3 is formed (Carter and Atkinson, 1989b).
In addition to depending on the VOC/NOX ratio (Table 3-3), incremental reactivity depends on
the composition of the VOC mixture and on the physical conditions encountered by the air
mass (including the dilution rate, light intensity, and spectral distribution (Carter and
Atkinson, 1989b; Carter, 1991).
3-36
-------
Table 3-3. Calculated Incremental Reactivities of Selected Volatile Organic
Compounds as a Function of the Volatile Organic Compound/Nitrogen Oxide
Ratio for an Eight-Component Volatile Organic Compound Mixture3 and
Low-Dilution Conditions'3
VOC/NOX Ratio (ppmC/ppm)
NMOC
CO
Ethane
w-Butane
w-Octane
Ethene
Propene
tmns-2-Butene
Benzene
Toluene
w-Xylene
Formaldehyde
Acetaldehyde
Methanol
Ethanol
Urban Mix
4
0.011
0.024
0.10
0.068
0.85
1.28
1.42
0.038
0.26
0.98
2.42
1.34
0.12
0.18
0.41
8
0.022
0.041
0.16
0.12
0.90
1.03
0.97
0.033
0.16
0.63
1.20
0.83
0.17
0.22
0.32
16
0.016
0.018
0.069
0.027
0.33
0.39
0.31
-0.002
-0.036
0.091
0.32
0.29
0.066
0.065
0.088
40
0.005
0.007
0.019
-0.031
0.14
0.14
0.054
-0.002
-0.051
-0.025
0.051
0.098
0.029
0.006
0.011
"Eight-component VOC mixture used to simulate NMOC emissions in an urban area.
bSee Appendix A for abbreviations and acronyms.
Source: Carter and Atkinson (1989b).
The O3-forming potentials of large numbers of VOCs, including emissions from
automobiles using gasoline and various alternative fuels such as methanol and ethanol blends
and compressed natural gas, have been investigated by airshed computer models (Chang and
Rudy, 1990; Chang et al., 1991a; Derwent and Jenkin, 1991; Andersson-Skold et al., 1992;
McNair et al., 1992, 1994; Carter, 1994). Consistent with the modeling studies of Carter and
Atkinson (1987, 1989b), these computer-modeling studies show that VOCs differ significantly
in terms of their O3-forming potential, for single-day as well as multiday conditions (Derwent
and Jenkin, 1991; Andersson-Skold et al., 1992; Carter, 1994; McNair et al., 1994).
However, there are some differences between the O3-forming potentials derived by Derwent
and Jenkin (1991) for multiday transport conditions over Europe and by Carter (1994) for 1-
day urban airshed "maximum incremental reactivity" conditions, especially for HCHO, an
important direct emission and atmospheric transformation product of most VOCs.
Recent modeling studies have been carried out by Carter (1994) and McNair et al.
(1994) to determine the O3-forming potential of alternative fuels. Emissions from vehicles
using 85% methanol and 15% gasoline (M-85) were shown to have -40% of the O3-forming
3-37
-------
potential of emissions from gasoline-fueled vehicles. Emissions from vehicles using liquefied
petroleum gas (LPG) and compressed natural gas (CNG) were shown to have O3-forming
potentials that are -50% and -18%, respectively, of the O3-forming potential of emissions
from gasoline-fueled vehicles (McNair et al., 1994).
3.2.5 Photochemical Production of Aerosols
The chemical processes involved in the formation of O3 and other photochemical
pollutants from the interaction of NOX and VOCs lead to the formation of OH radicals and
oxidized VOC reaction products that often are of lower volatility than the precursor VOC.
The OH radicals that oxidize the VOCs and lead to the generation of RO2 radicals and
conversion of NO to NO2 (with subsequent photolysis of NO2 form O3) also react with NO2
and sulfur dioxide (SO2) to form HNO3 and H2SO4, respectively, which can become
incorporated into aerosols as particulate nitrate (NO3) and sulfate (SO42"). The low-volatility
VOC reaction products can condense onto existing particles in the atmosphere to form
secondary organic aerosol matter. Hence, O3 formation, acid formation, and secondary
aerosol formation in the atmosphere are so related that controls aimed at reducing O3 levels
can impact (positively or negatively) acid and secondary aerosol formation in the atmosphere.
3.2.5.1 Phase Distributions of Organic Compounds
Chemical compounds are emitted into the atmosphere in both gaseous and particle-
associated forms. The emissions from combustion sources (e.g., vehicle exhaust) are initially
at elevated temperature, and compounds that may be in the particle phase at ambient
atmospheric temperature may be in the gas phase when emitted. In addition, atmospheric
reactions of gas-phase chemicals can lead to the formation of products that then condense
onto particles, or self-nucleate (Pandis et al., 1991; Wang et al., 1992; Zhang et al., 1992).
Measurements of ambient atmospheric gas- and particle-phase concentrations of several
classes of organic compounds indicate that the phase distribution depends on the liquid-phase
vapor pressure, PL (Bidleman, 1988; Pankow and Bidleman, 1992). The available
experimental data and theoretical treatments show that, as a rough approximation, organic
compounds with PL > 10"6 torr at ambient temperature are mainly in the gas phase (Bidleman,
1988). As expected, the gas-particle phase distribution in the atmosphere depends on the
ambient temperature, with the chemical being more particle-associated at lower temperatures.
The gas-to-particle adsorption-desorption process can be represented as,
A + TSP * F, (3'68)
where A is the gas-phase compound, F is the particle-phase compound, and TSP is the total
suspended particulate matter. The relationship among these three species is expressed using a
particle-gas partition coefficient, K:
K = F/(TSP)A. (3-69)
3-38
-------
Because K is a constant at a given temperature, if TSP increases (for example, in going from
a clean remote atmosphere to an urban area), F/A also must increase and the chemical
becomes more particle-associated (Pankow and Bidleman, 1991, 1992).
Gaseous and particulate species in the atmosphere are subject to wet and dry
deposition. Dry deposition refers to the uptake of gases and particles at the earth's surface by
vegetation, soil, and water, including lakes, rivers, oceans, and snow-covered ground. Wet
deposition refers to the removal of gases and particles from the atmosphere through
incorporation into rain, fog, and cloud water, followed by precipitation to the earth's surface.
These processes are discussed further in Section 3.6.
For gases, dry deposition is important primarily for HNO3, SO2, and H2O2 as well
as for O3 and PAN, whereas wet deposition is important for water-soluble gases such as
HNO3, H2O2, phenols, and, under atmospheric conditions, SO2. Dry deposition of particles
depends on the particle size; those of a mean diameter of -0.1 to 2.5 jim have lifetimes with
respect to dry deposition of -10 days (Graedel and Weschler, 1981; Atkinson, 1988),
sufficient for long-range transport. However, particles are efficiently removed from the
atmosphere by wet deposition (Bidleman, 1988).
Particles can form in the atmosphere by condensation or by coagulation, generally
occurring by coagulation in urban and rural areas. The photooxidation reactions of VOCs
typically lead to the formation of more oxidized and less volatile product species. When the
vapor pressures exceed the saturated vapor pressure (i.e., vapor pressure < 10"6 Torr), the
products will become particle-associated (Pandis et al., 1991, 1992). Accumulation-size
particles are in the size range 0.08 to 2.5 jim diameter (Whitby et al., 1972).
In urban areas, the major sources of particulate matter (Larson et al., 1989;
Solomon et al., 1989; Wolff et al., 1991; Hildemann et al., 1991a,b; Rogge et al., 1991, 1993;
Chow et al., 1993) are
direct emissions of elemental carbon from, for example, diesel-powered
vehicles (Larson et al., 1989);
direct emissions of primary organic carbon from, for example, meat cooking
operations, paved road dust, and wood-burning fireplaces and other combustion
sources (Hildemann et al., 1991a,b; Rogge et al., 1991, 1993);
secondary organic material formed in the atmosphere from the atmospheric
photooxidations of gas-phase NMOC (Turpin and Huntzicker, 1991; Pandis
et al., 1992);
the conversion of NO and NO2 to HNO3, followed by neutralization by NH4 or
through combination with other cations to form aerosol nitrates:
NH3(gas) + HNO3(gas) * MLNO3(aerosol); (3-69a)
the conversion of SO2 (and other sulfur-containing species) to H2SO4, which
has sufficiently low volatility to move to the aerosol phase; and
emission into the atmosphere of "fine dust", for example, crustal material.
Because the fine-particle size range is the same magnitude as the wavelength of
visible light, particulate matter present in the atmosphere leads to light scattering and
absorption, and hence to visibility reduction (Larson et al., 1989; Eldering et al., 1993).
3-39
-------
3.2.5.2 Acid Deposition
As noted above, the chemical processes involved in the formation of O3 and other
photochemical pollutants from the interaction of NMOC and NOX also lead to the formation
of acids in the atmosphere. The two major acidic species in ambient air are HNO3 and
H2SO4, arising from the atmospheric oxidation of NOX and SO2, respectively. Reduced sulfur
compounds emitted from biogenic sources and certain anthropogenic sources also may lead to
SO2 or sulfonic acids (Tyndall and Ravishankara, 1991).
The major sulfur-containing compound emitted into the atmosphere from
anthropogenic sources is SO2. In the troposphere, the important loss processes of SO2 are dry
deposition (Atkinson, 1988, and references therein), reactions within cloud water, and gas-
phase reaction with the OH radical. The rate constant for the reaction of SO2 with the OH
radical is such that the lifetime of SO2 with respect to gas-phase reaction with the OH radical
is -15 days. The reaction proceeds by (Stockwell and Calvert, 1983; Atkinson et al., 1992a)
OH + SO2 ™ HOSO2 (3-70)
HOSO2 + O2 -» HO2 + SO3 (3-71)
S03 + H20 -» H2S04. (3-72)
The reaction of SO3 with water vapor is slow in the gas-phase (Atkinson et al., 1992a) and,
hence, this may be a heterogeneous reaction. Because of its low vapor pressure, H2SO4 exists
in the aerosol or particle phase in the atmosphere.
Dry deposition is an important atmospheric loss process for SO2, because SO2 has
a fairly long lifetime, due to gas-phase chemical processes, and also has a high deposition
velocity. A lifetime, in relation to dry deposition of 2 to 3 days, appears reasonable
(Schwartz, 1989).
Sulfur dioxide is not very soluble in pure water (Schwartz, 1989). However, the
presence of pollutants such as H2O2 or O3, in the aqueous phase, displaces the equilibrium and
allows gas-phase SO2 to be incorporated into cloud, rain, and fog water, where it is oxidized
rapidly (Schwartz, 1989; Pandis and Seinfeld, 1989, and references therein):
S02(gas) - S02(aqu) (3-73)
SO2(aqu) + H2O - HSO3" + H+ - SO32" + 2 H+ (3-74)
HSO3" + H2O2 -> SO42" + H+ + H2O (3-75)
SO32" + O3 -» SO42" + O2. (3-76)
In addition, aqueous sulfur can be oxidized in a process catalyzed by transition metals such as
iron(III) (Fe3+) and manganese(II) (Mn2+) (Graedel et al., 1986b; Weschler et al., 1986; Pandis
and Seinfeld, 1989).
3-40
-------
SO32" + 1/2 O2 -> SO42" (3-77)
The oxidation rate of aqueous sulfur by O3 decreases as the pH decreases (i.e., as the acidity
increases) and this oxidation route is therefore self-limiting and generally of minor importance
in the atmosphere. The oxidation of SO2 by H2O2 appears to be the dominant aqueous-phase
oxidation process of SO2 (Chandler et al., 1988; Gervat et al., 1988; Schwartz, 1989; Pandis
and Seinfeld, 1989; Fung et al., 1991b), although the transition metal-catalyzed oxidation of
SO2 may also be important (Jacob et al., 1989). It should be noted that aqueous-phase H2O2
arises, in part, from the absorption of HO2 radicals and H2O2 into the aqueous phase, with
HO2 radicals being converted into H2O2 (Zuo and Hoigne, 1993).
The oxidation of SO2 to sulfate in clouds and fogs is often much faster than the
homogeneous gas-phase oxidation of SO2 initiated by reaction with the OH radical. The gas-
phase oxidation rate is -0.5 to 1% h"1, whereas the aqueous-phase (cloud) oxidation rate may
be as high as 10 to 50% h'1 (Schwartz, 1989).
The oxidation of NOX to HNO3 and nitrates was discussed in Section 3.2.3.
During daylight hours, oxidation occurs by the gas-phase reaction of NO2 with the OH
radical:
OH + NO2 HNO3, (3-15)
with the lifetime of NO2 due to Reaction 3-15 calculated to be ~1.4 days. Nitric acid is
removed from the troposphere by wet and dry deposition, with wet deposition being efficient.
During nighttime hours, NO2 can be converted into NO3 radicals and N2O5:
NO2 + O3 -+ NO3 + O2 (3-12)
M
NO3 + NO2 * N2O5, (3-13, -3-13)
with N2O5 undergoing wet or dry deposition, or both. The reader is referred to Schwartz
(1989) for further discussion of the conversion of NOX to NO3 and HNO3 and of acid
deposition.
3.3 Meteorological Processes Influencing Ozone
Formation and Transport
Day-to-day variability in O3 concentrations is, to a first approximation, the result of
day-to-day variations in meteorological conditions. This section presents a succinct overview
of those atmospheric processes that affect the concentrations of O3 and other oxidants in
urban and rural areas. Included in this list of processes are the vertical structure and
3-41
-------
dynamics of the PEL; transport processes, including thermally-driven mesoscale circulations
such as lake and sea breeze circulations; complex terrain effects on transport and dispersion;
vertical exchange processes; deposition and scavenging; and meteorological controls on
biogenic emissions and dry deposition.
3.3.1 Meteorological Processes
3.3.1.1 Surface Energy Budgets
Knowledge of the surface energy budget is fundamental to an understanding of the
dynamics of the PEL. The PEL is defined as that layer of the atmosphere in contact with the
surface of the earth and directly influenced by the surface characteristics. In combination
with synoptic winds, it provides the forces for the vertical fluxes of heat, mass, and
momentum. The accounting of energy inputs and outputs provides a valuable check on
modeled PEL dynamics.
Figure 3-7 illustrates the surface radiation budget for short-wave (wavelength
roughly <0.4 jim) and long-wave radiation. The radiation budget for the surface can be
described in terms of its components as
Qsfc = Ki - KT + Li - Lt + QH + QE . (3-78)
where K-l is the incoming short-wave radiation, KT is the outgoing short-wave radiation, L-l
is the incoming long-wave radiation from the atmosphere, LT is the outgoing long-wave
radiation, and QH and QE are the heat flux and latent heat flux to the soil, respectively. On a
global annual average, Qsfc is assumed to be near zero (i.e., the planet is not heating or
cooling systematically, an assumption clearly being questioned with the growing debate on
climatic change). On a day-to-day basis, however, Qsfc will certainly vary from zero and will
cause changes in surface temperature. Cloud cover, for example, will reduce the amount of
short-wave radiation reaching the surface and will modify all the subsequent components of
the radiation budget. Moreover, the redistribution of energy through the PEL creates
thermodynamic conditions that influence vertical mixing. The treatment of energy budgets
has been attempted on the scale of individual urban areas. These studies are summarized by
Oke (1978).
For many of the modeling studies of the photochemical production of O3, the
vertical mixing has been parameterized by a single, well-mixed layer. However, because this
is a great simplification of a complex structure, and because the selection of rate and extent
of vertical mixing may influence local control options, future modeling and observational
studies need to address the energy balances so that more realistic simulations can be made of
the structure of the PEL.
3-42
-------
/
Scattered
Absorbed
Absorbed Re-emitted
Latent
and
Sensible
Heat
Figure 3-7. Surface radiation budget for short- and long-wave radiation. The surface
radiation budget is driven by the input of short-wave radiation (1). This direct
input is reduced by scatter (2) and absorption passing through the atmosphere.
The amount that remains can be absorbed or reflected at the surface. The
reflected light (3) also can be scattered back to the surface (4). The short-
wave energy absorbed at the surface will ultimately be emitted back to the
atmosphere as long-wave radiation (5). The atmosphere absorbs much of this
radiation and radiates it back to the surface (7) and out to space (6). This
energy cycle is completed as some of the absorbed energy is transmitted to the
atmosphere as sensible and latent heat (8).
3.3.1.2 Planetary Boundary Layer
The concentration of an air pollutant depends significantly on the degree of mixing
that occurs between the time a pollutant or its precursors are emitted and the arrival of the
pollutant at the receptor. Atmospheric mixing is the result of either mechanical turbulence,
often associated with wind shear, or thermal turbulence associated with vertical redistribution
of heat energy. The potential for thermal turbulence can be characterized by atmospheric
stability. The more stable the air layer, the more work is required to move air vertically.
As air is moved vertically through the atmosphere, as might happen in a
convective thermal, its temperature will decrease with height as the result of adiabatic
expansion. It is the comparison of how the temperature should change with height in the
absence of external heating or cooling against the actual temperature change with height that
is a measure of atmospheric stability. Those layers of the atmosphere where temperature
increases with height (inversion layers) are the most stable as air, cooling as it rises, becomes
denser than its new warmer environment. In an atmospheric layer with relatively low
turbulence, pollutants do not redistribute vertically as rapidly as they do in an unstable layer.
3-43
-------
Also, because a stable layer has a relatively low rate of mixing, pollutants in a lower layer
will not mix through it to higher altitudes.
The stability of the atmosphere is often measured through computation of potential
temperature as
6 =
YR/C
P
(3-79)
where 6 is the virtual potential temperature, P is the pressure of the air parcel, P0 is the
reference pressure to which the air parcel will be moved (usually 1,000 mb), R is the gas law
constant, and cp is the specific heat of air at constant pressure. The faster 6 increases with
height, the less the potential for mixing.
A stable layer can also act as a trap for air pollutants lying beneath it. Hence, an
elevated inversion is often referred to as a "trapping" inversion. On the other hand, if
pollutants are emitted into a stable layer aloft, such as might occur from an elevated stack, the
lack of turbulence will keep the effluents from reaching the ground while the inversion
persists.
Traditionally, atmospheric mixing has been treated through use of a mixing height,
which is defined as the base of an elevated inversion layer. In this model, the O3 precursors
are mixed uniformly through the layer below the mixing height. As this layer grows, it both
entrains remnant O3 from previous days and redistributes fresh emissions aloft. The vertical
mixing profile through the lower layers of the atmosphere is assumed to follow a typical and
predictable cycle on a generally clear day. In such a situation, a nocturnal surface inversion
would be expected to form during the night as LT exceeds L-l. This surface layer inversion
persists until surface heating becomes significant, probably 2 or 3 h after sunrise. Pollutants
initially trapped in the surface inversion may cause relatively high, local concentrations, but
these concentrations will decrease rapidly when the surface inversion is broken by surface
heating. The boundary formed between the rising, cooling air of the growing mixing layer
and that of the existing PEL is often sharp and can be observed as an elevated temperature
inversion.
Elevated temperature inversions, when the base is above the ground, are also
common occurrences (Hosier, 1961; Holzworth, 1964, 1972). This condition can form simply
as the result of rapid vertical mixing from below, but is exacerbated in regions of subsiding
air when the sinking air warms to a point at which it is warmer than the rising and cooling
underlying air. Because these circumstances are associated with specific synoptic conditions,
they are less frequent than the ubiquitous nighttime radiation inversion. An elevated
inversion is, nevertheless, a very significant air pollution feature, because it may persist
throughout the day and, thus, restrict vertical mixing.
When compared to a source near the surface and the effects of a radiation (surface)
inversion, the pollutant dispersion pattern is quite different for an elevated source plume
trapped in a layer near the base of an elevated inversion. This plume will not be in contact
with the ground surface in the early morning hours because there is no mixing through the
surface radiation inversion. Thus, the elevated plume will not affect surface pollutant
concentrations until the mixing processes become strong enough to reach the altitude of the
plume. At that time, the plume may be mixed downward quite rapidly in a process called
3-44
-------
fumigation. During fumigation, surface O3 concentrations will increase if the morning
O3 concentration is higher aloft than at the ground and if insufficient scavenging by NO
occurs at ground-level. In fact, the rapid rise in O3 concentrations in the morning hours is
often the result of vertical (downward) transport from an elevated reservoir of O3. After this
initial increase, surface concentrations can continue to increase as a result of photochemistry
or transport of O3-rich air to the receptor.
When surface heating decreases in the late afternoon and early evening, the surface
inversion will form again under most conditions. The fate of the elevated inversion is less
clear, however. Although O3 and its precursors have been mixed vertically, the reduction of
turbulence and mixing at the end of the daylight hours leaves O3 in a remnant layer that is
often without a well-defined thermodynamic demarcation. This layer is then transported
through the night, often to regions far removed from pollution sources, where its pollutants
can influence concentrations at remote locations the next morning, as mixing entrains the
elevated remnant layer. This overnight transport can be aided by the development of a
nocturnal jet that forms many nights at the top of the surface inversion layer.
Geography can have a significant impact on the dispersion of pollutants (such as
along the coast of an ocean or one of the Great Lakes). Near the coast or shore, the
temperatures of land and water masses can be different, as can the temperature of the air
above these land and water masses. When the water is warmer than the land, there is a
tendency toward reduction in the frequency of surface inversion conditions inland over a
relatively narrow coastal strip (Hosier, 1961). This in turn tends to increase pollutant
dispersion in such areas. The opposite condition also occurs if the water is cooler than the
land, as in summer or fall. Cool air near the water surface will tend to increase the stability
of the boundary layer in the coastal zone, and thus decrease the mixing processes that act on
pollutant emissions. These conditions occur frequently along the New England coast (Hosier,
1961). Similarly, pollutants from the Chicago area have been observed to be influenced by a
stable boundary layer over Lake Michigan (Lyons and Olsson, 1972). This has been observed
especially in summer and fall when the lake surface is most likely to be cooler than the air
that is carried over it from the adjacent land.
Sillman et al. (1993) investigated abnormally high concentrations of O3 observed in
rural locations on the shore of Lake Michigan and on the Atlantic coast in Maine, at a
distance of 300 km or more from major anthropogenic sources. A dynamical-photochemical
model was developed that represented formation of O3 in shoreline environments and was
used to simulate case studies for Lake Michigan and the northeastern United States. Results
suggest that a broad region with elevated O3, NOX, and VOC forms as the Chicago plume
travels over Lake Michigan, a pattern consistent with observed O3 at surface monitoring sites.
Near-total suppression of dry deposition of O3 and NOX over the lake is needed to produce
high O3. Results for the East Coast suggest that the observed peak O3 can be reproduced only
by a model that includes suppressed vertical mixing and deposition over water, 2-day
transport of a plume from New York, and superposition of the New York and Boston, MA,
plumes. Hence, the thermodynamics associated with the water bodies seem to play a
significant role in some regional-scale episodes of high O3 concentrations.
There is concern that the strict use of mixing height unduly simplifies the complex
atmospheric processes that redistribute pollutants within urban areas. There is growing
evidence that some O3 precursors may not be evenly redistributed over some urban areas.
These are cases where the sources are relatively close to the urban area and atmospheric
mixing is not strong enough to redistribute the material over a short travel time. In these
3-45
-------
cases, it is necessary to treat the turbulent structure of the atmosphere directly and
acknowledge the vertical variations in mixing. Methods that are being used to investigate
these processes include the use of a diffusivity parameter to express the potential for mixing
as a function of height. A simple expression of how the mean concentration, %, changes with
time, t, in an air parcel, assuming all concentrations are homogeneous in the horizontal, is
(3-80)
where w'%' is the vertical turbulent eddy flux of pollutant. The term on the right hand side
of the equation changes mean concentration through flux divergence (i.e., turbulence either
disperses the pollutant to or from the point being considered). The problem with this
representation is that the flux divergence term is virtually impossible to measure directly.
The turbulent eddy flux needed to understand the vertical distribution of O3 and its
precursors often is parameterized in photochemical models, if included at all, through use of
eddy diffusivity. The eddy diffusivity is set using an analogy to mixing length theory as
(3-81)
which allows estimation of flux divergence from measured or estimated vertical gradients in
concentration and estimation of the eddy diffusivity. The selection of diffusivity is often
somewhat arbitrary, but can be related to the eddy diffusivity for heat or momentum,
depending on circumstances. Large values result in rapid mixing. Thus, the appropriate
selection of eddy diffusivity is necessary to simulate whether elevated plumes will enter an
urban airshed.
The use of an eddy diffusivity approach to turbulent diffusion assumes local down-
gradient diffusion, a situation not always realistic in the atmosphere. It is important to note
that eddy diffusivity is not valid under convective conditions because counter-gradient flows
occur (Sun, 1986). Eddy diffusivity also does not work in the presence of multiple stable
layers. Moreover, the form of the eddy diffusivity used in the existing air quality models is
rather arbitrary. More research will be needed to remove this arbitrariness.
Another method used for convective situations is a technique called "large-eddy
simulation", employed to recreate the probability of redistribution within the mixing height.
This method explicitly simulates the larger eddies occurring under convective situations.
These techniques require meteorological information that is not normally available from the
National Weather Service but that is now becoming available as part of several O3 field
experiments.
3.3.1.3 Cloud Venting
Vertical redistribution of O3 out of the PEL is achieved by the venting of
pollutants in clouds. Clouds represent the top-most reaches of thermals of air rising through
the PEL and can act as chemical reactors for soluble pollutants, returning the "processed" air
to the PEL. They also can result in physical redistribution of O3 and its precursors from the
3-46
-------
PEL if convection is sufficiently vigorous (Greenhut, 1986; Dickerson et al., 1987). Clouds
also act to influence photolysis rates and chemical transformation rates.
Greenhut (1986) showed that the net O3 flux in the cloud layer was a linear
function of the difference in O3 concentration between the boundary and cloud layers. Ozone
fluxes between clouds were usually smaller than those found within clouds, but the slower
rate is at least partially offset by the larger region of cloud-free air relative to cloudy air.
Large clouds, such as cumulonimbus, offer considerably more potential for
redistribution of O3 and its precursors. Additionally, the cumulonimbus clouds also are
associated with precipitation, a scavenger of pollutants, and with lightning, a potential source
for NOX. Using CO as a tracer, Dickerson et al. (1987) and Pickering et al. (1990) have
illustrated the redistribution potential of cumulonimbus cloud systems. Lyons et al. (1986)
provided an illustration of the potential for groups of cumulonimbus clouds to vent the
polluted boundary layer.
The role of cloud venting is thought to be largely a cleansing process for the
boundary layer, although a portion of the material lifted into the free troposphere could be
entrained back to the surface in subsequent convection. Aircraft observations frequently have
documented the occurrence of relatively high O3 concentrations above lower concentration
surface layers (e.g., Westberg et al., 1976). This is a clear indication that O3 is preserved
essentially in layers above the surface and can be transported over relatively long distances,
even when continual replenishment through precursor reactions is not a factor, such as at
night.
3.3.1.4 Stratospheric-Tropospheric Ozone Exchange
The fact that O3 is formed in the stratosphere, mixed downward, and incorporated
into the troposphere, where it forms a more or less uniformly mixed background
concentration, has been known in various degrees of detail for many years (Junge, 1963).
The mechanisms by which stratospheric air is mixed into the troposphere have been examined
by a number of authors, as documented previously by the U.S. Environmental Protection
Agency (EPA) (U.S. Environmental Protection Agency, 1986a, and references therein).
Although the portion of background O3 near the surface attributed to stratospheric-
tropospheric O3 can be in the 5 to 15 ppb range for a seasonal average, this amount of O3 by
itself cannot account for peak urban O3 values or regional episodes of elevated O3 levels
(Johnson and Viezee, 1981; Ludwig et al., 1977; Singh et al., 1980; and Viezee et al., 1979).
Johnson and Viezee (1981) concluded that the O3-rich intrusions studied sloped downward
toward the south. In terms of dimensions, the average crosswind width (north to south), at an
altitude of 5.5 km (ca. 18,000 ft), for six spring intrusions, averaged 226 km, and, for four
fall tropopause fold systems, averaged 129 km. Ozone concentrations at 5.5 km averaged 108
ppb in the spring systems and 83 ppb in the fall systems. From this and other research
described in the previous criteria document for O3 and other photochemical oxidants (U.S.
Environmental Protection Agency, 1986a), Viezee and coworkers (Viezee and Singh, 1982;
Viezee et al., 1983) concluded (1) that direct ground-level impacts by stratospheric O3 may be
infrequent, occurring <1% of the time; (2) that such ground-level events are short-lived and
episodic; and (3) that they are most likely to be associated on a 1- to 4-h average with
O3 concentrations in the range 60 to 100 ppb.
The monthly stratospheric-tropospheric total O3 flux from tropospheric folding
events in the Northern Hemisphere has been estimated. These fluxes, in units of
1035 O3 molecules per month per tropospheric folding event, increase from 1.0 in January to
3-47
-------
2.1 in April and May, decline to 1.0 in August, and reach a minimum of 0.5 in October and
November (Viezee et al., 1983). The spring-to-fall variation resembles the seasonal variations
of O3 near ground-level often observed at more remote sites (e.g., Logan, 1985). Four of the
10 episodes in which ground-level O3 has been attributed to stratospheric O3 transport
occurred during March, but several such episodes during summer months have been reported
(see below).
Using the 7Be-to-O3 ratio as an indicator of O3 of stratospheric origin and SO42"
concentrations as a tracer for anthropogenic sources, Altshuller (1987) estimated stratospheric
contributions of O3 in the range 0 to 40 ppb (0 to 95% of observed O3) at ground level at
Whiteface Mountain, NY, for July 1975 and mid-June to mid-July 1977. Monthly average
stratospheric contributions were estimated at 5 to 10 ppb. Extant 7Be and O3 data for a
number of lower-elevation rural locations in the western, midwestern, and southeastern United
States also were examined, and stratospheric or upper tropospheric contributions of 6 to 8 ppb
were calculated. The author concluded that the calculated values for such contributions
should be viewed with caution and regarded as probable upper limits because of scatter in the
7Be and corresponding O3 data that hindered definition of the 7Be-to-O3 ratio. Altshuller also
concluded that removal and dilution processes result in the loss of most stratospheric O3
before it reaches ground level. In other work performed in England, O3 of stratospheric
origin was estimated to contribute 10 to 15 ppb to the daily maximum hourly mean
O3 concentrations in an April through October period (Derwent and Key, 1988; United
Kingdom Photochemical Oxidant Review Group, 1993).
3.3.2 Meteorological Parameters
This section focuses on analyses of data from previous and ongoing measurement
programs to address two key questions: (1) are there meteorological parameters which are
systematically associated with O3 levels? and (2) are relationships between O3 and
meteorological parameters sufficiently strong such that meteorological fluctuations can be
filtered from the data to allow examination of longer term trends?
The meteorological factors that theoretically could influence surface O3 levels
include ultraviolet radiation, temperature, wind speed, atmospheric mixing and transport, and
surface scavenging. The following examines the theoretical basis for each of these factors
and identifies to what degree empirical evidence supports the hypotheses.
3.3.2.1 Sunlight
Ultraviolet radiation plays a key role in initiating the photochemical processes
leading to O3 formation. Sunlight intensity (specifically the UV portion of sunlight) varies
with season and latitude, but the latter effect is strong only during winter months. The
importance of photolysis to the formation of O3 provides a direct link between O3 and time of
year. However, during the summer, the maximum UV intensity is fairly constant throughout
the contiguous United States, and only the duration of the solar day varies to a small degree
with latitude.
The effects of light intensity on individual photolytic reaction steps and on the
overall process of oxidant formation have been studied in the laboratory (Peterson, 1976;
Demerjian et al., 1980). Early studies, however, employed constant light intensities, in
contrast to the diurnally varying intensities that occur in the ambient atmosphere. The diurnal
variation of light intensity was subsequently studied as a factor in photochemical oxidant
3-48
-------
formation (e.g., Jeffries et al., 1975, 1976). Such studies showed that the effect of this factor
varies with initial reactant concentrations. Most important was the observation that similar
NMOC/NOX systems showed different oxidant-forming potential depending on whether studies
of these systems were conducted using constant or diurnal light. This led to incorporation of
the effects of diurnal or variable light into photochemical models (Tilden and Seinfeld, 1982).
There is little empirical evidence in the literature, however, linking day-to-day
variations in observed UV radiation levels with variations in O3 levels. Samson and Shi
(1988) illustrated that the number of O3 concentrations exceeding 120 ppb did not track well
with potential solar radiation, as shown in Figure 3-8. Although variations in day-to-day
concentrations could well be influenced by cloud cover or attenuated by haze, the seasonal
peak in O3 concentrations usually lags the peak in potential solar radiation that occurs at the
Summer Solstice on or about June 23.
3.3.2.2 Temperature
There is an association between tropospheric O3 concentration and tropospheric
temperature that has been demonstrated from measurements in outdoor smog chambers and
from measurements in ambient air. A linear relationship between maximum O3 and
temperature was obtained in the smog chambers with little scatter around the regression line
(Kelly and Gunst, 1990). Numerous ambient studies done over more than a decade have
reported that successive occurrences or episodes of high temperatures characterize seasonally
high O3 years (Clark and Karl, 1982; Kelly et al., 1986). The relationship has been observed
for the South Coast Air Basin of California (Kuntasal and Chang, 1987), in New England
(Wolff and Lioy, 1978; Atwater, 1984; Wackter and Bayly, 1988), and elsewhere.
Figures 3-9 and 3-10 show the daily maximum O3 concentrations versus maximum
daily temperature for summer months (May to October), 1988 to 1990, for Atlanta and New
York City, NY, and for Detroit, MI, and Phoenix, AZ, respectively. There appears to be an
upper-bound on O3 concentrations that increases with temperature. Likewise, Figure 3-11
shows that a similar qualitative relationship exists between O3 and temperature even at a
number of rural locations.
The notable trend in these plots is the apparent upper-bound to O3 concentrations
as a function of temperature. It is clear that, at a given temperature, there is a wide range of
possible O3 concentrations because other factors (e.g., cloudiness, precipitation, wind speed)
can reduce the O3 production. The upper bound presumably represents the maximum
O3 concentration achieved under the most favorable conditions. These plots, based on
ambient air measurements, show wide scatter in O3 concentration with temperature because of
the contributions to variations from several of these factors that do not influence the results
from smog chamber studies (Kelly and Gunst, 1990). Table 3-4 lists the results of a
statistical regression performed on the paired O3-temperature data used in Figures 3-9 and
3-10 with separate slopes listed for temperatures above and below 30 °C. Results show that,
for T > 30 °C, the O3-temperature relationship is statistically significant at all sites. The rate
of increase for T > 30 °C is 3 to 5 ppb/°C at eastern United States rural sites and ranges from
4 to 9 ppb/°C at the three eastern U.S. urban sites (New York, Detroit, and Atlanta). At two
western sites, Williston, ND, and Billings, MT, there is a much weaker
Figure 3-8. The number of reports of ozone concentrations >120 ppb at the 17 cities
studied in Samson and Shi (1988). (1 April = Week 14, 1 May = Week 18,
1 June = Week 22, 1 July = Week 27, 1 August = Week 31, 1 September =
3-49
-------
35
30
4i 25
cT
I-
Al
•6 15
o 10
o
Annual Variation
in Solar
Radiation
/\
A
700
600
500
300
200
100
"
14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46
Week of Year
Week 35, 1 October = Week 40, 1 November = Week 44). A representation of
the annual variation in solar radiation reaching the earth's surface at 40°N
latitude (units = cal cm'2) is shown.
Source: Samson and Shi (1988).
dependence on temperature, possibly reflecting the lower level of anthropogenic activity.
At a third western site, Medford, OR, the O3-temperature relationship is comparable to that at
rural eastern sites.
Relationships between peak O3 and temperature also have been recorded by
Wunderli and Gehrig (1991) for three locations in Switzerland. At two sites near Zurich,
peak O3 increased 3 to 5 ppb/°C for diurnal average temperatures between 10 and 25 °C, and
little change in peak O3 occurred for temperatures below 10 °C. At the third site, a
high-altitude location removed from anthropogenic influence, much less variation of O3 with
temperature was observed.
The hypotheses for this correlation of O3 with temperature include, but are not
necessarily limited to:
Reduction in photolysis rates under meterological conditions associated with
low temperatures;
Reduction in H2O concentrations at low temperatures;
Thermal decomposition of PAN and its homologues;
Increased anthropogenic emissions of reactive hydrocarbons or NOX, or both;
Increased natural emissions of reactive hydrocarbons; and
3-50
-------
240
Maximum Temperature (°C)
Figure 3-9. A scatter plot of maximum daily ozone concentration in Atlanta, GA, and New
York, NY, versus maximum daily temperature.
Maximum Temperature (°C)
Figure 3-10. A scatter plot of maximum daily ozone concentration in Detroit, MI, and
Phoenix, AZ, versus maximum daily temperature.
3-51
-------
• Ann Arbor, Ml
Williamsport, PA >• Mammoth Cave, KY O MoKensie City, ND
210
Maximum Daily Temperature (°C)
Figure 3-11. A scatter plot of maximum ozone concentration versus maximum daily
temperature for four nonurban sites. The relationship with temperature is
still apparent, although the slope is reduced from that of the urban areas.
Table 3-4. Rates of Increase of Peak Ozone with Diurnal Maximum
Temperature (ppb/°C) for Temperature <300 K (27 °C) and Temperature
>300 K, Based on Measurements for April 1 to September 30, 19881
Location
Urbanized Regions
Nonurban Sites
NY-NJ-CT
Detroit
Atlanta
Phoenix
Southern California
Williamsport, PA
Saline, MI
Mammoth Cave, KY
Kentucky, cleanest site 3
Williston, ND
Billings, MT
Medford, OR
T<
ACVAT
1.5
1.4
3.2
—
11.3
1.2
0.8
0.1
0.3
0.2
0.1
0.5
300 K
T-Statistic
-5.2
-6.4
-4.2
—
-8.9
-5.0
-3.5
-0.3
-0.7
-1.0
-0.5
-2.6
T>
AOj/AT
8.8
4.4
7.1
1.4
—
4.0
3.1
4.4
3.4
0.8
0.7
3.3
300 K
T-Statistic
-7.4
-6.3
-5.9
-4.1
—
-7.4
-4.9
-7.3
-6.6
-3.7
-2.2
-13.7
aSee Appendix A for abbreviations and acronyms.
3-52
-------
The relationship with temperature is well known, but not yet reproduced by air
quality models. Although it has been argued that this striking relationship with temperature is
an indirect result of the stagnant synoptic meteorological conditions that lead to higher
O3 levels, the correlation is not strong with other parameters of stagnation, notably wind
speed, which is discussed later.
Reduction in Photolysis Rates
It is possible that, on a seasonal scale, the correlation between temperature and
O3 may be an indirect correlation with UV radiation variability. This is insufficient, however,
to explain the day-to-day correlation between the two variables.
Changes in photolysis rates and in H2O concentrations are related in that both are
linked to the supply of OH radicals, which determines the rate of O3 production in clean
atmospheres and contributes to O3 production in polluted atmospheres. A reduction in either
photolysis rates or H2O would reduce the source of OH radicals. Calculations by Sillman and
Samson (1995) showed that the difference between summer and fall photolysis rates (at 40° N
latitude) has a significant impact on the rate of O3 production in urban photochemical
simulations, roughly equal to the impact of PAN thermal decomposition (discussed below).
However the impact of photolysis rates and of water vapor was much lower in simulations for
polluted rural environments. In the simulations by Sillman et al. (1993), O3 production in
urban environments was limited largely by the supply of OH radicals to react with
hydrocarbons; whereas in rural environments the limiting factor was the source of NOX.
Consequently, photolysis rates and H2O had less impact on O3 production in rural
environments.
Thermal Decomposition of Peroxyacetyl Nitrate
Temperature-dependent photochemical rate constants provide a link between
O3 and temperature (Sillman et al., 1990a; Cardelino and Chameides, 1990). The reason for
the decline in O3 in rural areas when the PAN decomposition rate decreases is that PAN
represents a sink for NOX in rural environments. When the rate of PAN decomposition is
decreased, NOX drops sharply, whereas OH and HO2 remain largely unaffected.
Consequently, the rate of the important HO2 + NO reaction (see Section 3.2) shows a
substantial decrease.
The photochemical response in an urban environment is fundamentally different,
although the final result, a decrease in O3 with temperature, is similar. The impact of PAN in
urban environments is attributable to its role as a sink for odd hydrogen rather than to its
effect on NOX (Cardelino and Chameides, 1990). Sillman et al. (1990a) have shown that the
well-known division of O3 photochemistry into NOX- and VOC-sensitive regimes is associated
with the relative magnitude of odd-hydrogen sinks.
Sillman and Samson (1995) found that the thermal decomposition of PAN was
enough to explain an increase of 1 to 2 ppb peak O3/°C increase in temperature in rural
locations in the eastern United States, based on photochemical simulations. This increase
represents a significant fraction of the observed increase in peak O3 with a rise in temperature
(3 to 5 ppb/°C, Table 3-4).
Increased Anthropogenic Emissions
Emission rates for anthropogenic hydrocarbons (VOCs) also can increase with
temperature (U.S Environmental Protection Agency, 1989; Stump et al., 1992). Increased
3-53
-------
anthropogenic VOC emissions might be expected to cause increased rates of O3 production in
urban areas where O3 is sensitive to VOC, but would be less likely to have impact on rural
areas where biogenic VOC emissions can predominate. However, O3 in rural areas is NOX
dependent. The NOx-sensitive rural areas also would show increased O3 production with a
rise in temperature as biogenic NOX emissions increase with temperature.
Increased Natural Emissions
Emissions of biogenic hydrocarbons increase sharply with a rise in temperature
(Lamb et al., 1987). In ambient temperatures from 25 to 35 °C, the rate of natural
hydrocarbon emissions from isoprene-emitting deciduous trees increased by about a factor
of 4. From coniferous trees, the increase was on the order of 1.5.
Recently, Jacob et al. (1993) found that the photochemistry of O3 production in a
polluted rural environment (Blue Ridge Mountains, VA) is significantly different in
September and October, when natural emissions from deciduous forests have ceased. The
difference in chemistry between summer and fall leaf production also may have an impact on
the O3-temperature correlation.
Correlation with Stagnation
Recently, Jacob et al. (1993b) found that model-simulated O3 formation in the rural
United States shows a tendency to increase with a rise in temperature, based solely on the
difference in atmospheric circulation between relatively warm and relatively cool days. The
model-simulated O3-temperature correlation was less than observed but large enough to
represent a significant component of the observed correlation. However, the temperature-
meteorology correlation identified by Jacob et al. (1993b) was based on simulated
meteorology from a General Circulation Model rather than on direct observations.
3.3.2.3 Wind Speed
Ozone is expected to be influenced by wind speed because lower wind speeds
should lead to reduced ventilation and the potential for greater buildup of O3 and its
precursors. Abnormally high temperatures are frequently associated with high barometric
pressure, stagnant circulation, and suppressed vertical mixing resulting from subsidence
(Mukammal et al., 1982), all of which may contribute to elevated O3 levels. However, in
reality this relationship varies from one part of the country to another. Figure 3-12 shows the
frequency of 24-h trajectory transport distances to Southern cities on days with resulting
concentrations of O3 > 120 ppb (Samson and Shi, 1988). The frequency for Southern cities is
biased toward lower wind speeds. The same bias was shown for all 17 cities in the study.
A similar plot for cities in the northeastern United States (Figure 3-13) shows an opposite
pattern, in which the bias is toward higher wind speeds than normal. It is unclear how much
meteorological information is needed in order to perform accurate urban-area O3 simulations
using advanced photochemical models. To understand the significance of variations between
upper-air wind measurements during the Southern Oxidant Study (SOS), 1992, Atlanta,
an intensive, intercomparison test of the precision of upper-air measurements was conducted.
Collocated measurements were made at an SOS measurement site, using a boundary-layer
lidar, a wind profiler, and a rawinsonde balloon. There was generally good agreement
between the profiler and rawinsonde, although some large outliers existed. Figure 3-14
illustrates that the root-mean-square difference (RMSD) varied with altitude. The RMSD
reached a minimum near 1,200 m above ground level (AGL) of about 2 m/s, rising to over
3-54
-------
D Dallas
D Baton
Rouge
D Atlanta
• Houston
-•- All 17 cities
50 150 250 350 450 550 650 750 850 950 >1,000
Distance traveled (km)
Figure 3-12. The frequency of 24-h trajectory transport distance en route to city when
ozone was > 120 ppb in four Southern U.S. cities, compared with the percent
frequency distribution for all 17 cities (scale on right) of a nationwide study,
1983 to 1985.
Source: Samson and Shi (1988).
10
I 7
£• e
B
fr
§
D Providence
• New Haven
D Portland, [
ME
-^ All 17 cities
D Boston
-, 15
10
LJ- 50 150 250 350 450 550 650 750 850 950 > 1,000
Distance traveled (km)
Figure 3-13. The frequency of 24-h trajectory transport distance en route to city when
ozone was >120 ppb in four New England cities, compared with the percent
frequency distribution for all 17 cities (scale on right) of a nationwide study,
1983 to 1985.
Source: Samson and Shi (1988).
3-55
-------
2,000
1,600
(D
1,200
(5
£
g>
'
800
400
Mean RMSD
o Minimum
RMSD
•••• Maximum
RMSD
12
15
RMSD (m/s)
Figure 3-14. The root-mean-square difference (RMSD) between CLASS observations and
profiler observations as a function of height above ground level.
3 m/s near the surface and above 1,200 m AGL. Figure 3-15 illustrates the RMSD for the
lidar comparison with CLASS observations. There is slightly greater RMSD at all heights
than for the profiler-rawinsonde comparison, with a relative minimum observed at about 1,200
m.
Although the measurements were significantly correlated, the results illustrate that
there was still considerable disagreement between methods. The profiler had better precision
than the lidar had, although the differences were negligible if the first four runs were
excluded from the data set. The profiler obtained values biased slightly higher than the
CLASS system (+0.2 m/s), whereas the lidar system was biased low (-0.3 m/s or -0.5 m/s).
The statistical comparisons of both the profiler and the boundary-layer lidar with the
rawinsonde system suggest that variations in wind speed at a particular level must be larger
than about 3 m/s to be considered significant.
3.3.2.4 Air Mass Characteristics
In meteorology, an "air mass" is a region of air, usually of multistate dimensions,
that exhibits similar temperature, humidity, or stability characteristics. Air masses are created
when air becomes stagnant over a "source region" and subsequently takes on the
characteristics of the source region. Similarly, when dealing with air pollution meteorology,
it is possible to identify a "chemical air mass" as a region of air that has become stagnant
over an emissions source area. Air that is stagnant over, say, the center of Canada will
exhibit relatively cold, dry conditions and will be relatively devoid of pollutants. Air that
3-56
-------
2,000
Mean RMSD
B Minimum
RMSD
• • Maximum
RMSD
RMSD (mis)
Figure 3-15. The root-mean-square difference (RMSD) between CLASS observations and
lidar observations as a function of height above ground level.
resides over the industrial regions of the midwestern United States will exhibit low visibility
and, often, high O3 levels on a regional scale. Meteorological processes play an important
role in determining the amount of "accumulation" of O3 and its precursors that occurs under
such stagnant conditions.
Episodes of high O3 concentrations in urban areas often are associated with high
concentrations of O3 in the surroundings. This accumulated O3 forms under the same
atmospheric conditions that lead to high O3 levels in urban areas and exacerbates the urban
problem by supplying relatively high O3 and precursor concentrations to the urban area from
upwind. The transport of O3 and its precursors beyond the urban scale (<50 km) to
neighboring rural and urban areas has been well documented (e.g., Wolff et al., 1977a,c;
Wolff and Lioy, 1978; Clark and Clarke, 1984; Sexton, 1982; Wolff et al., 1982; Altshuller,
1988). A summary of most of these reports was given in the 1986 O3 criteria document (U.S.
Environmental Protection Agency, 1986a) and will not be reiterated here. The phenomena of
high nonurban O3 levels was illustrated by Stasiuk and Coffey (1974) for transport within
New York State; by Ripperton et al. (1977) for sites in the Middle Atlantic States; and by
Samson and Ragland (1977) for the midwestern United States.
These areas of O3 accumulation are characterized by synoptic-scale subsidence of
air in the free troposphere, resulting in development of an elevated inversion layer; relatively
low wind speeds associated with the weak horizontal pressure gradient around a surface high
pressure system; lack of cloudiness; and high temperatures.
On occasion, O3 at levels greater than 120 ppb can occur in rural areas far
removed from urban or industrial sources. Ozone levels at the summit of Whiteface
Mountain exceeded this value during the summer of 1988 when O3 accumulated across a
3-57
-------
wide expanse of the eastern United States at levels >120 ppb. Nonetheless, even when the
regional accumulation is at a level below the current O3 NAAQS, the increment needed to
bring the level above the NAAQS in an urban area is not large.
The identification and understanding of the transport of photochemical O3 and
other oxidants and their precursors by weather systems represent a significant advance in
comprehending photochemical air pollution and the potential extent of its effects.
Considerable progress has been made in the development of long-range photochemical
modeling techniques so that the likely impact of synoptic systems can be anticipated. Such
tools are very much in the research stage, however, because the local impact of O3 and other
oxidants results from a complex interaction of distant and local precursor sources, urban
plumes, mixing processes, atmospheric chemical reactions, and general meteorology.
3.3.3 Normalization of Trends
The degree to which meteorological factors can be "normalized" out of the
O3 concentration and "trends" data depends in large part on the strength of the relationships
between O3 and meteorological components. As part of the SOS Atlanta intensive field
campaign, an attempt was made to model statistically the O3 levels in Atlanta to build a
predictive tool for forecasting days of specialized measurement. Figure 3-16 shows the fit of
the data used to create the model to the model simulations. Figure 3-17 shows the fit
obtained from independent data collected in 1992.
40 80 120
OB (Observed) (ppb)
160
200
Figure 3-16. Model of ozone (OJ levels using regression techniques. The use of wind
speed, temperature, and previous-day ozone provided a means to forecast O3
levels.
3-58
-------
200
40 80 120 160 200
O3 (observed) (ppb)
Figure 3-17. Simulated versus observed ozone levels using regression techniques on an
independent data set obtained in the summer of 1992 in Atlanta, GA.
This model was used successfully to predict next-day O3 levels in Atlanta. Ozone
levels in a number of American cities should be analyzed using regression tools such as this
in order to normalize meteorological variability. Through such analyses, it is possible that
trends, if any, represented as systematic deviations from the model, may become observable.
A summary of other techniques for removing meteorological variability is contained in the
recent monograph from the National Research Council (1991). Table 3-5 lists a sample of
studies aimed at evaluation of O3 trends.
3.4 Precursors of Ozone and Other Oxidants
3.4.1 Sources and Emissions of Precursors
3.4.1.1 Introduction
As described elsewhere in this chapter, O3 is formed in the atmosphere through a
series of chemical reactions that involve VOCs and NOX. Control of O3 depends on reducing
emissions of VOCs or NOX or both. Thus, it is important to understand the sources and
source strengths of these precursor species in order to devise the most appropriate oxidant
control strategies. In the following sections, anthropogenic and biogenic NOX and VOC
sources will be described, and the best estimates of their current emission levels and trends
will be provided. Confidence levels for the assigned source strengths will be discussed.
Both English and metric units have been utilized in emission inventories. Thousands
or millions of short tons are the common scales in the English system. The metric unit most
3-59
-------
Table 3-5. Recent Studies Examining Trends in Ozone Data After Removal
of Variability Associated with Meteorological Factors
Study
Variables
Approach
Jones et al. (1989)
Pollock et al. (1988)
Kuntasal and Chang (1987)
Wakim (1989)
Chock et al. (1982)
Kumar and Chock (1984)
Korsog and Wolff (1991)
Surface temperature
Surface temperature
850-mb temperature
Surface temperature
Surface temperature, wind speed,
relative humidity, sky cover, wind
direction, dew point temperature,
sea level pressure, precipitation.
Surface temperature, wind speed,
relative humidity, sky cover, wind
direction, dew point temperature,
sea level pressure, precipitation.
Surface temperature, wind speed,
relative humidity, sky cover, wind
direction, dew point temperature,
sea level pressure, precipitation.
Compared number of days with
ozone concentrations above
120 ppb to days with temperature
above 30 °C.
Compared number of days with
ozone concentrations above
105 ppb to days with temperature
above 30 °C.
Regression of ozone versus
temperature for Southern
California.
Regression of ozone versus
temperature for Houston, New
York, and Washington, DC.
Regression versus a variety of
meteorological parameters.
Regression versus a variety of
meteorological parameters.
Regression versus a variety of
meteorological parameters.
Source: National Research Council (1991).
often employed is millions of metric tons, which is equivalent to teragrams (Tg). To convert
English tons to teragrams, multiply English tons by 0.907 x 10"6. For consistency, teragrams
have been employed throughout the ensuing discussion.
3.4.1.2 Nitrogen Oxides
Manmade Emission Sources
Anthropogenic NOX sources are associated with combustion processes. The
primary pollutant is NO, which is formed from nitrogen and oxygen atoms that are produced
at high combustion temperatures when air is present. In addition, NOX is formed from
nitrogen contained in the combustion fuel. Major NOX source categories include
transportation, stationary source fuel combustion, industrial processes, solid waste disposal,
3-60
-------
and some miscellaneous combustion-related activities. Table 3-6 provides a more detailed
summary of each of these source categories. The transportation category includes gasoline-
and diesel-powered motor vehicles, aircraft, railcars, vessels, and off-highway vehicles.
Electric utilities, industrial and commercial/institutional boilers, industrial furnaces, and space
heaters comprise the stationary source fuel combustion category. Industrial processes include
petroleum refining and paper, glass, steel, cement, and chemical production. The incineration
and open burning of waste leads to emissions of NOX in the solid waste disposal category.
The miscellaneous sources category includes prescribed forest slash burning, agricultural
burning, coal refuse burning, and structure fires. It should be noted at this point that, even
though NO is the pollutant emitted, NOX emission inventories are quantified relative to NO2
(mol wt = 46). Nitrogen dioxide is a secondary pollutant produced via oxidation of NO in
the atmosphere.
Quantifying NOX emissions in all of these categories generally requires multiplying
an emission factor and an activity level. Nitrogen oxides emission factors are obtained from
Compilation of Air Pollution Factors, AP-42 (U.S. Environmental Protection Agency, 1985),
and from the current mobile source emission factor model (e.g., MOBILES) recommended by
EPA. Activity levels are derived from information sources that provide consumption levels.
This takes the form of fuel type and amount consumed for stationary sources and, for
transportation sources, the number of vehicle miles traveled (VMT). Point-source emissions
are tallied at the individual plant level. These plant-by-plant NOX emissions are first summed
at the state level, and then state totals are added to arrive at the national emissions total. Data
on VMT are published for three road categories: (1) highways, (2) rural roads, and (3) urban
streets.
Table 3-7 provides a summary of NOX emissions from the various categories
mentioned previously (U.S. Environmental Protection Agency, 1993b). The 1991 total is
21.39 Tg of NOX emissions in the United States. About half of the emissions (10.69 Tg) is
associated with the stationary source fuel combustion category. Transportation-related
activities are the second largest source, accounting for about 45% of the national total. The
remaining 7% of emissions are divided among the industrial processes, solid waste disposal,
and miscellaneous sources categories. The two largest single NOX emission sources are
electric power generation and highway vehicles. Local NOX source apportionment may differ
substantially from these national figures.
Because of the dominance of the electric utility and transportation sources, the
geographical distribution of NOX emissions is related to areas with a high density of
power-generating stations and urban regions with high traffic densities. Figure 3-18 shows
the location of the 50 largest electric power generating sources of NOX in the United States.
The majority of these power plants are concentrated in the upper Mississippi-Ohio River
corridor. Because of this congregation of large point sources, 69% of U.S. NOX emissions
occur within U.S. Environmental Protection Agency Regions III, IV, V, and VI
(Figure 3-19). It is interesting to compare the annual NOX emissions from a large electrical
generating plant with the yearly transportation-related emissions in a major metropolitan
region. The largest utility plants currently release between 0.06 and 0.09 Tg of NOX annually,
which compares to approximately 0.12 Tg of NOX emitted by transportation sources in the
Atlanta urban area (U.S. Environmental Protection Agency, 1993b).
3-61
-------
Table 3-6. Source Categories Used to Inventory Nitrogen Oxides Emissions
Transportation
Stationary Source
Fuel Combustion
Industrial Processes
Solid Waste Disposal
Miscellaneous
CO
Highway vehicles
Gasoline-powered
Passenger cars
Light trucks - 1
Light trucks - 2
Heavy-duty vehicles
Motorcycles
Diesel-powered
Passenger cars
Light trucks
Heavy-duty vehicles
Aircraft
Railcars
Vessels
Farm machinery
Construction machinery
Industrial machinery
Other off-highway vehicles
Coal
Electric utilities
Industrial
Commercial/Institutional
Residential
Pulp mills
Organic chemicals
Ammonia
Nitric acid
Petroleum refining
Glass
Cement
Lime
Iron and steel
Incineration
Open burning
Forestries
Other burning
Fuel oil
Electric utilities
Industrial
Commercial/Institutional
Residential
Natural gas
Electric utilities
Industrial
Commercial/Institutional
Residential
Wood
Industrial
Residential
Other Fuels
Industrial
Residential
Source: U.S. Environmental Protection Agency (1992a).
-------
Table 3-7. 1991 Emission Estimates for Manmade Sources
of Nitrogen Oxides in the United States
Source Category
Emissions (Tg)
Transportation
Highway vehicles
Off-highway vehicles
Stationary fuel combustion
Electric utilities
Industrial
Other
Industrial processes
Solid waste disposal
Miscellaneous
Forest burning
Other burning
Miscellaneous organic solvents
Total of all sources
7.20
2.51
6.74
3.27
0.68
9.71
10.69
0.80
0.07
0.12
21.39
Source: U.S. Environmental Protection Agency (1993b).
Legend
^0.065 to 0.100 Tg/year
# 0.045 to 0.065 Tg/year
-------
^vf
II \ V
Figure 3-19. Nitrogen oxides emissions (Tg/year) from manmade sources in the
10 U.S. Environmental Protection Agency regions of the United States, 1991.
Source: U.S. Environmental Protection Agency (1993b).
Seasonal variations are available for the 1993 NOX emissions from 14 source
categories in the United States (U.S. Environmental Protection Agency, 1994). Very little
seasonal variation occurred for categories contributing approximately three-quarters of the
total annual NOX emissions, including the categories of highway vehicles, electric utilities, and
industrial combustion sources. For off-highway sources (comprising 20% of the total annual
NOX emissions), 29% of the off-highway NOX emissions occur in the summer and 21% in the
winter. In contrast, the category of other combustion sources (comprising 5% of the total
annual NOX emissions) emit 47% of the NOX emissions in the winter and only 8% in the
summer. An earlier inventory for 1985 (U.S. Environmental Protection Agency, 1989), which
considered only total point and area anthropogenic NOX emissions, indicated that very little
variation occurred in NOX emissions among the winter, spring, summer, and fall seasons. The
contributions of these NOX categories also were shown to vary seasonally by region of the
United States.
Trends in Nitrogen Oxides Emissions
Estimates of NOX emissions have been made back to 1900, when approximately
2.3 Tg were emitted into the atmosphere in the United States (U.S. Environmental Protection
Agency, 1992a). Figure 3-20 summarizes the growth in NOX emissions at 10-year intervals
since the 1940s. Emissions grew rapidly until the 1970s and then leveled off at about
Figure 3-20. Changes in nitrogen oxides emissions from manmade sources in the United
States, 10-year intervals, 1940 through 1990.
3-64
-------
25-r
1940 1950 1960 1970 1980 1990
Years
Source: U.S. Environmental Protection Agency (1992a).
20 Tg/year. Currently, more than 90% of the national NOX emissions result from
transportation activities and stationary fuel combustion. Figure 3-21 illustrates the growth in
each of these categories over the last 50 years. Transportation-related NOX emissions grew
steadily until the 1980s and then exhibited a moderate decrease. However, the recent trends
in transportation-related NOX emissions shown in Table 3-8 indicate no trend between 1987
and 1991. Emissions of NOX from fuel combustion sources have increased continually from
1940 to the present time.
Recent trends in the major NOX emission categories are shown in Table 3-8.
Between 1987 and 1991, transportation-related NOX emissions have remained essentially
constant, whereas the stationary source NOX emissions have increased about 10%.
Transportation and stationary source fuel combustion will likely show downward
trends in their NOX emissions during the next 20 years. This will result from the provisions
of the Clean Air Act, which was passed in 1990. Emission limits for electric utility boilers
have been prescribed to reduce acidic deposition, automobile tailpipe emission standards will
be tightened, and current technology-based applications will be required for industrial boilers
(non-utility) in O3 nonattainment areas. In addition, the average grams of NOX per mile from
passenger cars is expected to decrease because of new on-board diagnostic systems and
expanded inspection and maintenance requirements.
As a result of new emission limits and revised performance standards, NOX
emissions from electric utilities are expected to decrease by 16% by the year 2000. Control
requirements in the industrial non-utility sector are expected to reduce NOX emissions by 10%
between 1990 and 2000. Projections based on VMT and emission factors from the MOBILE
model suggest nearly a 50% decrease in NOX emissions from highway vehicles manufactured
from 1990 to 2000 (U.S. Environmental Protection Agency, 1992a).
3-65
-------
SF
TR
1940
1950
1960 1970
Years
1980
1990
Figure 3-21. Changes in nitrogen oxides emissions from stationary source fuel combustion
(SF) and transportation (TR) from 1940 through 1990."
The values for 1990 do not agree with those in Table 3-8 because different models were employed for deriving
the short- and long-term trends.
Source: U.S. Environmental Protection Agency (1992a).
Table 3-8. Recent Trends in Nitrogen Oxides Emissions
for Major Manmade Source Categories (Tg)
Year
1991
1990a
1989
1988
1987
Transportation
9.7
9.9
9.7
9.9
9.7
Stationary Source
Fuel Combustion
11.0
10.7
10.7
10.6
10.1
The values for 1990 do not agree with those in Figure 3-21 because different models were employed for
deriving the short- and long-term trends.
Source: U.S. Environmental Protection Agency (1993b).
3-66
-------
Uncertainty of Anthropogenic Nitrogen Oxides Emission Estimates
Because a large proportion of the U.S NOX emissions are derived from distinct
point sources, it generally is believed that published estimates are very reliable. For example,
the National Acid Precipitation Assessment Program (NAPAP) NOX inventory for U.S.
emissions in 1985 (18.6 Tg) was assigned a 90% relative confidence interval in the range of
6 to 11% (Placet et al., 1991). This confidence level was based on judgments used to assign
uncertainty to component inputs of emission models and on statistical assumptions used to
aggregate uncertainty values.
Sources of error are associated with both the emission factors and the activity
levels utilized in the inventorying process. Emission factors provide quantitative estimates of
the average rate of emissions from many sources. Consequently, these factors are best
applied to a large number of sources over relatively long time periods. In other words, an
NOX emission estimate for a single point source on a particular day in 1990 may be highly
inaccurate; but the emission value for this same source for the entire year of 1990 may be
very good. It appears that the emission factors assigned to the transportation sectors may be
the most uncertain. This results from the emission factors having been derived from mobile
source models that require multiple inputs. This type of model requires information on
temperatures, vehicle speeds, gasoline volatility, and several other parameters.
Recent attempts to validate NOX emission factors or inventories have involved
comparing ambient NOX concentrations with values predicted using emissions-based models.
These generally have taken one of two forms: (1) comparisons between NOX concentrations
measured in a tunnel and those predicted from emission factors, activity levels, and dilution
factors in the tunnel; or (2) whole-city integration procedures in which ambient NOX
concentrations are compared to ambient NOX levels that have been predicted using a model
such as the Urban Airshed Model (UAM). The latter approach has been applied in the South
Coast Air Basin (Fujita et al., 1992). It was reported that measured and predicted NOX
concentrations agreed within 20% for a 2-day period in August 1987. Likewise, the results
from tunnel studies (Pierson et al., 1990; Robinson et al., 1996) have shown reasonably good
agreement between predicted and measured NOX concentrations. It is important to keep in
mind that ambient NOX levels predicted using a modeling method cannot be assigned true
value status. There could be as much or more uncertainty in the model outputs as there is in
the emission inputs that are being tested. The fact, however, that an emissions-based model
predicts ambient concentrations that are close to those measured tends to lend credence to the
NOX emission estimates. No systematic study of the effect of these uncertainties on model
predictions has been published, but a limited summary of sensitivity analyses appears in
Seinfeld (1988).
In addition, NOX inventory validation has involved comparing annual emission
estimates reported by different groups. Table 3-9 shows several annual U.S. NOX emission
estimates. In 1982, the estimates vary by less than 12% and this decreases to about 9% in the
1985 comparison.
Natural Emission Sources
Natural sources of NOX include lightning, soils, wildfires, stratospheric intrusion,
and the oceans. Of these, lightning and soils are the major contributors. The convention is to
include emissions from all soils in the biogenic or natural category even though cultivated
soil emissions are in a sense anthropogenic; cultivated soils also appear to produce higher
emissions than those from undisturbed forest and prairie soils, as discussed later. Although
3-67
-------
Table 3-9. Comparison of Estimates of Nitrogen Oxides Emissions
from Manmade Sources in the United States3
Emissions/year (Tg)
Inventory 1982 1985
NAPAP — 18.6
EPA 19.6 19.8
MSCET 18.8 18.2
EPRI 20.7 —
aSee Appendix A for abbreviations and acronyms.
Source: U.S. Environmental Protection Agency (1993a).
NOX emitted from large wildfires can be significant on a regional scale, this source is overall
considered to be of minor importance for the United States. Injection of NOX into the upper
troposphere via subsidence from the stratosphere is estimated at less than 0.1 Tg/year for all
of North America. Because of the relatively short lifetime of NOX (1 to 3 days) and small
NOX emissions from sea water, transport of NOX from oceans is thought to be a negligible
source in the United States.
Lightning. Lightning produces high enough temperatures to allow N2 and O2 to be
converted to NO. Two methods have been employed to estimate the NOX source strength
from lightning:
(1) Multiply the frequency of lightning flashes by the energy dissipated per flash
and the NO production per unit of energy dissipated; or
(2) Relate NOX production to NOj deposition in remote areas where lightning-
produced NOX is thought to be the dominant NOj precursor.
Method 1 yields an annual NOX production of approximately 1.2 Tg for North America
(Placet et al., 1991). The deposition-based estimate (Method 2) gives a somewhat larger
value of 1.7 Tg/year (Placet et al., 1991). The NAPAP inventory included lightning-produced
NOX on a gridded 10° x 10° latitude-longitude scale. Most of the continental United States
fits within 30° to 50° N latitude and 80° to 120° W longitude. The estimated annual
lightning-produced NOX for this region (continental United States) is about 1.0 Tg. Roughly
60% (0.6 Tg) of this NOX is generated over the southern tier of states (30° to 40° N latitude;
80° to 120° W longitude).
Soils. Both nitrifying and denitrifying organisms in the soil can produce NOX.
The relative importance of these two pathways is probably highly variable from biome to
biome. Nitric oxide is the principal NOX species emitted from soils, with emission rates
depending mainly on fertilization levels and soil temperature. Several reports have noted a
large increase in NOX emissions from agricultural soils treated with NOj-containing fertilizers
(Johansson and Granat, 1984; Kaplan et al., 1988; Johansson, 1984). Measurements of soil
NOX emissions have established that the relationship with temperature is exponential,
3-68
-------
consisting of approximately a twofold increase for each 10 °C rise in temperature (Williams
et al., 1992; Valente and Thornton, 1993).
Inventorying soil NOX emissions is difficult because of the large temporal and
spatial variability in emissions. The existing inventories have been developed using emission
algorithms that are functions of soil temperature and land-use type. Two broad, land-use
categories—natural and agricultural—have been assigned. The natural soils are broken down
into biome types, and the agricultural soils subdivided according to fertilizer applications.
The highest biogenic NO emissions are in corn-growing regions of the midwest (Nebraska,
Iowa, and Illinois) during summer months. Of the total U.S. biogenic emissions of NO from
soils, 85% occur during the spring and summer months.
Table 3-10 provides a summary of the annual soil NOX emissions from the 10 U.S.
Environmental Protection Agency regions. Approximately 60% of this NOX is emitted in
Regions V, VII, and VIII (see Figure 3-19), which contain the central U.S. corn belt. The
total estimate for U.S. soil emissions is 1.2 Tg.
Table 3-10. Annual Nitrogen Oxides Emissions from Soils
by U.S. Environmental Protection Agency Regiona
U.S. Environmental Protection Agency Region NOX Emissions (Tg)
I, II, and III
IV
V
VI
VII
VIII
IX
X
Total
0.05
0.11
0.26
0.18
0.27
0.21
0.04
0.01
1.20b
aSee Appendix A for abbreviations and acronyms.
bValues do not sum to total due to independent rounding.
Source: Placet et al. (1991).
Uncertainty in Estimates of Natural Nitrogen Oxides Emissions
As previously indicated, inventorying NOX produced from lightning requires
multiplying the number of flashes by average energy factors. No attempt has been made to
assign confidence limits to these variables. A measure of the uncertainty associated with
lightning-produced NOX is provided, however, by comparing emission estimates generated
independently. Two estimates of the amount of lightning-generated, summertime NOX in the
southeastern United States (2.4 and 8.5 x 10"2 Tg) varied by approximately a factor
of 4 (Placet et al., 1991).
3-69
-------
Sources of uncertainty when inventorying NOX emissions from soils include:
land-use assignments, soil temperature, and emission algorithm development. Confidence
levels assigned to categories 1 and 2 are about ±50%. The emission algorithm is developed
from field measurements of NOX emission rates versus temperature for various land-use
categories. Measurement accuracy is approximately ±30%. However, because of the natural
variability of NOX emissions within a specific soil category, uncertainty in the exponential
relationship that relates emission rate to temperature is estimated to be in the range of a factor
of 2 to 4.
Comparison of Emissions from Manmade and Natural Sources
On an annual basis, natural sources (lightning and soils) contribute approximately
2.2 Tg of NOX to the troposphere over the United States. This compares to the 1990
anthropogenic emission estimate of 19.4 Tg. Annual NOX emissions from soils (1.2 Tg) are
about 6% of the manmade emissions in the United States. This percentage increases to about
14% when the comparison includes only the summer months of July, August, and September.
Even larger biogenic contributions can occur in certain regions of the United States. For
example, it is estimated that biogenic NOX emissions from soils account for about 19% of
summertime NOX emissions in Tennessee (Valente and Thornton, 1993) and actually exceed
emissions from manmade sources during the summer months in the states of Nebraska and
South Dakota (Williams et al., 1992).
3.4.1.3 Volatile Organic Compounds
Manmade Emission Sources
Volatile organic compounds are emitted into the atmosphere by evaporative and
combustion processes. Many hundreds of different organic species are released from a large
number of source types. The species commonly associated with atmospheric O3 production
contain from 2 to about 12 carbon atoms. They can be true hydrocarbons, which possess
only carbon and hydrogen atoms (e.g., alkanes, alkenes, aromatics), or substituted
hydrocarbons that contain a functional group such as alcohol, ether, carbonyl, ester, or
halogens. The emissions of methane have been ignored because of their largely natural origin
and the fact that the importance of methane is limited primarily to global scale processes.
In addition, the atmospheric oxidation rate of methane is very slow compared to the higher
molecular weight organics.
In 1991, the total U.S. emissions of VOCs was estimated to be 21.0 Tg (U.S.
Environmental Protection Agency, 1993b). The two largest source categories were industrial
processes (10.0 Tg) and transportation (7.9 Tg). Lesser contributions were attributed to waste
disposal and recycling (2.0 Tg), stationary source fuel combustion (0.7 Tg), and miscellaneous
area sources (0.5 Tg). Table 3-11 provides a more detailed breakdown of VOC source
contributions. Within the industrial category, solvent utilization, petroleum product storage
and transfer, and chemical manufacturing are the major contributors. Volatile organic
compounds released from highway vehicles account for almost 75% of the transportation-
related emissions.
Speciated hydrocarbon emissions from manmade sources were reported in the 1985
NAPAP Emissions Inventory. Emissions of each main hydrocarbon family exceeded 1 Tg.
Alkanes comprised about 33%, aromatics 19%, and alkenes 11% of anthropogenic VOC
emissions in the 1985 inventory (Placet et al., 1991). None of the major oxygenated
3-70
-------
Table 3-11. Estimated 1991 Emissions of Volatile Organic Compounds
from Manmade Sources in the United States
Source Category Emissions (Tg)
Transportation 7.87
Highway vehicles 6.00
Off-highway vehicles 1.87
Stationary fuel combustion 0.68
Electric utilities 0.03
Industrial 0.26
Other 0.39
Industrial processes 9.97
Chemical manufacture 1.61
Petroleum and related industries 0.68
Solvent utilization 5.50
Petroleum product storage and transport 1.69
Other 0.49
Waste disposal and recycling 2.01
Miscellaneous 0.51
Total all sources 21.04
Source: U.S. Environmental Protection Agency (1993b).
hydrocarbon groups (e.g., carbonyls, organic acids, phenols) listed in the speciated inventory
exceeded 1 Tg. The carbonyl group, which included formaldehyde, higher aldehydes,
acetone, and higher ketones, was the largest contributor of oxygenated hydrocarbons at
0.73 Tg.
Seasonal variations are available for the 1993 anthropogenic VOC emissions from
14 source categories in the United States (U.S. Environmental Protection Agency, 1994).
Very few seasonal variations occur for categories contributing approximately 85% of the total
annual VOC emissions. The only category of VOC emissions that showed significant
seasonal variation was off-highway sources, which comprises 13% of the total annual VOC
emissions. These off-highway sources contribute 31% of their VOC emissions in summer and
19% in winter. An earlier inventory for 1985 (U.S. Environmental Protection Agency, 1989),
which considered total point and area anthropogenic VOC emissions, indicated that very little
variation in VOC emissions occurred between the winter, spring, summer, and fall seasons.
The contribution of these VOC emissions also were shown to vary seasonally by region of the
United States.
Trends in Emissions
Emissions of nonmethane VOCs peaked in the early 1970s and have decreased
continually since then. Emissions of VOCs increased from 15.5 Tg in 1940 to 27.4 Tg in
3-71
-------
1970 and now are estimated to be back down to approximately the same level as in 1940
(U.S. Environmental Protection Agency, 1992a). Figure 3-22 illustrates these changes at
10-year intervals from 1940 to 1990. Until 1970, highway vehicles were the major source of
VOC emissions. As automobiles have been equipped with more and better emission controls,
however, emissions from the transportation sector have dropped below those from industrial
processes, the category that is now the leading contributor of VOC emissions to the
atmosphere. Transportation, industrial processes, and the miscellaneous burning and
solvent-use categories have accounted for 83 to 93% of VOC emissions over the past
50 years. Figure 3-23 shows the emission trends for these three categories. The
transportation-related emissions of VOCs are currently estimated to be at about the same level
as in 1940. Industrial process VOC emissions nearly tripled between 1940 and 1980,
followed by a small decline in more recent years. The miscellaneous category exhibited a
decrease in emissions from 4.5 Tg in 1940 to a 1990 level estimated at 2.8 Tg/year.
30-r
1940
1950
1960 1970
Years
1980
1990
Figure 3-22. Changes in emissions of volatile organic compounds from major manmade
sources in the United States, 10-year intervals, 1940 through 1990.
Source: U.S. Environmental Protection Agency (1992a).
Trends for the dominant VOC emissions categories from 1987 through 1991 are
shown in Table 3-12. Projections for the year 2000 forecast a 62% reduction in VOC
emissions from highway vehicles compared to 1990 levels. The major reduction in the
transportation area will contribute to a predicted overall 25% decrease in total national VOC
emissions between 1990 and 2000 (U.S. Environmental Protection Agency, 1992a).
Uncertainty in Estimates of Emissions from Manmade Sources
It has proven difficult to determine the accuracy of VOC emission estimates.
Within an area source such as an oil refinery, emission factors and activity levels are
3-72
-------
14
12--
g
.
s.
2--
Transportation
Industrial Processes
Miscellaneous
1940
1950
1960 1970
Years
1980
1990
Figure 3-23. Changes in emissions of volatile organic compounds from major manmade
sources, 1940 through 199O.a
The values for 1990 do not agree with those in Table 3-12 because different models were employed for
deriving the short- and long-term trends.
Source: U.S. Environmental Protection Agency (1992a).
Table 3-12. Recent Trends in Emissions of Volatile Organic Compounds
from Major Categories of Manmade Sources (Tg)
Year
1991
1990a
1989
1988
1987
Transportation
7.87
8.07
8.26
9.15
9.29
Industrial Processes
7.86
9.96
9.92
10.00
9.65
Waste Disposal
and Recycling
2.01
2.05
2.08
2.10
2.05
The values for 1990 do not agree with those in Figure 3-23 because different models were employed for
deriving the short- and long-term trends.
Source: U.S. Environmental Protection Agency (1993b).
3-73
-------
assigned for thousands of individual sources (e.g., valves, flanges, meters, processes), and
emission estimates for each of these sources are summed to produce the emissions total.
Because it would be impractical to determine an emission factor for each of these sources
within a refinery individually, average emission factors for the various source categories are
utilized. This can lead to substantial error if emissions from the individual sources deviate
from the assigned average factor. Even more troublesome are area sources that include a
large evaporative emissions component. These sources are dependent on environmental
factors such as temperature, which add to the difficulty in establishing reliable emission
estimates. Such sources fall into a miscellaneous solvent evaporation category that includes
emissions from processes such as dry cleaning, degreasing, printing, automobile body repair,
furniture manufacture, and motor vehicle manufacture.
Assigning accurate VOC emission estimates to the mobile source category has
proven troublesome, as well. Models are used that incorporate numerous input parameters,
each of which has some degree of uncertainty. For example, activity models are employed to
characterize the mobile source fleet. This includes the number of vehicles in various
categories (e.g., gasoline-fueled, diesel-fueled, catalyst-equipped, non-catalyst-equipped, etc.),
miles accumulated per year for each type of vehicle, and ages of the vehicles. Vehicle
registration statistics are employed for category assignment. Errors can arise because
registration data are not always up to date, and unregistered vehicles are completely omitted.
Military vehicles, foreign-owned automobiles, and old "junkers" that are on the highways but
not registered are included in the inventorying process. The activity models assume that
vehicles of the same age accumulate mileage at the same rate. This is most likely not
correct; there is a need to assess the uncertainty in this assumption through a systematic
collection of vehicle type, age, and mileage accumulation statistics.
Experiments carried out in tunnels have looked at the relationship between
measured VOC emission factors and those derived from automotive emission models. In a
study designed to verify automotive emission inventories for the South Coast Air Basin,
measurements in the Van Nuys Tunnel indicated that automotive VOC emissions were a
factor of 4 larger than predicted using emission models (Pierson et al., 1990). Results from
two tunnel studies conducted in 1992 (Robinson et al., 1996) show much better agreement
between VOC measurements and model predictions than those obtained in the Van Nuys
Tunnel. Comparisons were made for the Tuscarora Tunnel on the Pennsylvania Turnpike in
south-central Pennsylvania and at the Fort McHenry Tunnel under Baltimore Harbor. For
Tuscarora, MOBILE4.1 gives 131% of the average measured VOC values and MOBILES
gives 216% of the average measured VOCs. For Fort McHenry, MOBILE4.1 gives 53% of
the average measured VOC values and MOBILES gives 81% of the average measured VOCs.
Somewhat better agreement also was obtained from the Van Nuys Tunnel data when the
updated California emission model EMFACTEP was used (Robinson et al., 1996). At the
Gassier Tunnel in Vancouver, Canada, agreement within ±30% was obtained between the
VOC measurements and the Canadian version of the MOBILE models (Gertler et al., 1994).
The 1993 results from the Caldecott Tunnel in the San Francisco, CA, area show deviations
between VOC measurements and model predictions similar to those obtained in the Van Nuys
Tunnel, possibly because of similar urban/local fleets. Differences in test results between the
newer tunnel studies and those obtained in the early Van Nuys Tunnel study are likely due to
the better condition of the vehicles in the newer studies and the lack of power enrichment or
other transients because of the steady-speed driving. However, it is important to appreciate
that results from tunnel measurements do not necessarily predict equivalency of VOC
3-74
-------
measurements with model predictions under highly transient urban/local driving and fleet
conditions.
Recent developments in remote sensing have permitted more accurate measurement
of hydrocarbon exhaust emissions from on-road vehicles (Stedman et al., 1991). These
studies have demonstrated a highly skewed distribution, with the majority of VOC emissions
coming from about 20% of the automobiles. Emission factors developed from laboratory
dynamometer testing most likely do not properly account for the high-emitting vehicle
contribution (Pitchford and Johnson, 1993). In many cases, these high emitters are older cars
that are poorly maintained. In order to reduce this source of uncertainty, it may be necessary
to reassess the life spans assigned to vehicles. Vehicles manufactured more than 25 years
prior to 1993 are not included in the inventory. However, these older vehicles are likely to
be high emitters, and, if they are underrepresented in the model, emissions will be
underestimated. Activity models provide data in terms of national averages. This can
contribute to inaccuracies in emissions estimates if a particular region varies from the national
average in terms of vehicle types, age, or VMT.
Ambient measurements of VOCs and NOX have been employed in order to better
define uncertainty levels in VOC inventories. Some of the earliest work was carried out in
the Atlanta area in the 1980s. Using a simple model and measured ambient VOC and NOX
concentrations, it was shown that ambient NOX levels were consistent with the urban NOX
emission estimates. However, measured ambient VOC concentrations were as much as a
factor of 6 greater than predicted (Westberg and Lamb, 1985). Improvements in mobile
source emission models have resulted in somewhat higher emission estimates, so that the
discrepancy between model estimates and ambient data has been reduced to about a factor of
2.5 (Fujita et al., 1992; Cadle et al., 1993). It is clear that the relationship between emission
inventories and ambient concentrations of NOX and VOCs warrants further study. In addition
to improving the mobile source emission inventories, it will be necessary to place uncertainty
bounds on stationary source inventories. Whether stationary source emissions of VOCs are
underpredicted using current emission inventory methodology is not known (Finlayson-Pitts
and Pitts, Jr., 1993).
Biogenic Emissions
Vegetation emits significant quantities of reactive VOCs into the atmosphere.
Many of these biogenic VOCs may contribute to O3 production in urban (Chameides et al.,
1988) and rural (Trainer et al., 1987) environments. The VOC emissions of primary interest
are isoprene and the monoterpenes (e.g., oc-pinene, (3-pinene, myrcene, limonene, etc.), which
are hydrocarbons. Recent field measurements have shown that a variety of oxygenated
organics also are emitted from plants (Winer et al., 1992). A thorough discussion of biogenic
emissions and their implication for atmospheric chemistry has been published recently by
Fehsenfeld et al. (1992), who reviewed the techniques used to measure VOC emissions from
vegetation, laboratory emissions studies that have been used to relate emission rates to
temperature and light intensity, development of emission models, and the use of emission
models in the preparation of emission inventories.
Since the late 1970s, a number of regional and national biogenic emission
inventories have been reported (Zimmerman, 1979; Winer et al., 1983; Lamb et al., 1985,
1987, 1993). These inventories are based on algorithms that relate VOC emissions from a
particular vegetation class to ambient temperature, land-use, and, in the case of isoprene,
photosynthetically active radiation. Most biogenic VOC emissions from vegetation increase
3-75
-------
exponentially with temperature. Isoprene emissions are light-dependent, being minimal at
night and increasing with solar intensity during the day. Deciduous vegetation is the
dominant source of isoprene; whereas coniferous trees emit primarily monoterpenes. Other
things being equal, isoprene is emitted at a much higher rate than the monoterpenes. For
example, in a southern forest of mixed pine and hardwoods, the isoprene emission rate from
an oak tree is about 10 times larger than the flux of oc-pinene from an adjacent loblolly pine
during the midday period.
The most recent biogenic VOC emissions estimate for the United States totals
29 Tg/year (Lamb et al., 1993). This estimate includes 5.9 Tg isoprene, 4.4 Tg oc-pinene, 6.5
Tg other monoterpenes, and 12.3 Tg other VOCs. Table 3-13 provides a summary of the
contributions from the various vegetation categories based on an inventory of monthly
statewide data for eight land-cover types. In preparing this inventory, algorithms were
developed that related VOC emissions to temperature and light for each of the biomass
categories shown in the table. On a national scale, coniferous forests are the largest
vegetative contributor because of their extensive land coverage. The category "Other VOCs"
is the dominant biogenic hydrocarbon contributor to the national total. From the standpoint
of inventory accuracy, this is somewhat unfortunate because the identities of most of the
"Other VOCs" are uncertain. This classification has carried over from the extensive
field-measurement program conducted by Zimmerman (1979) and coworkers in the
mid-1970s. The category "Other VOCs", includes peaks that showed up in sample
chromatograms at retention times that could not be matched to known hydrocarbons. It is
likely that if the Zimmerman study were repeated today, most of the species making up this
"Other VOCs" category could be identified. Recent field studies have made use of GC/MS
techniques that were not available to Zimmerman in the 1970s.
Biogenic emissions, because of their dependence on temperature and vegetational
growth, vary by season. In addition, the southern tier of states is expected to produce more
biogenic emissions than those in the north because of higher average temperatures.
Table 3-14 shows a spatial and temporal breakdown of U.S. biogenic emissions. Summertime
emissions comprise 16.7 of the 29.1 Tg (or 57%) of the annual totals in all regions. The EPA
Regions IV and VI in the southeastern and southcentral United States, respectively, have the
highest summertime and annual biogenic VOC emission rates. Region IV contributes 16% of
the summertime and 18% of the annual biogenic VOC emissions in the United States,
whereas Region VI contributes 21% of the summertime and 23% of the annual biogenic
emissions in the United States. Compared to Regions IV and VI, regions to the north have
more rapid increases in biogenic VOC emissions in the spring and more rapid decreases in
biogenic VOC emissions in the fall.
Uncertainty in Estimates of Biogenic Emissions
Sources of error in the biogenic inventorying process arise from uncertainties in
emission measurements, determination of biomass densities, land-use characterization, and
measurement of light intensity and temperature. Within each of these categories, the error is
relatively small. However, when emission measurements are combined with temperature or
light intensity, or both, into a single algorithm, the uncertainty increases greatly. This results
from the fact that temperature and light are only surrogates for the real physiological
processes that control biogenic emissions. Emission rate and ambient temperature can be
highly correlated for data collected from one tree branch over a 24-h period; but, when these
data are combined with measurements from other branches and other trees the correlation is
3-76
-------
Table 3-13. Annual Biogenic Hydrocarbon Emission Inventory for the United States (Tg)a
Compound
Isoprene
oc-pinene
Other terpenes
Other VOCs
Total
Percent of Total
Oak
Forests
2.31
0.19
0.41
1.12
4.03
13.9
Other
Deciduous
Forests
1.01
0.23
0.44
0.88
2.56
8.8
Coniferous
Forests
0.61
2.07
3.08
2.72
8.48
29.2
Scrub-
lands
1.17
0.78
1.41
2.49
5.85
20.1
Land Use
Grass-
lands
0.49
0.13
0.24
0.45
1.31
4.5
Crop-
lands
0.2
0.85
0.81
4.51
6.37
21.9
Inland
Waters
0.02
0.06
0.06
0.07
0.21
0.7
Urban
Areas
0.08
0.04
0.06
0.08
0.26
0.9
U.S. Total
5.9
4.4
6.5
12.3
29.1
"See Appendix A for abbreviations and acronyms.
Source: Lamb et al. (1993).
-------
00
Table 3-14. Annual Biogenic Hydrocarbon Emission Inventory by Month and
by U.S. Environmental Protection Agency Region for United States Emissions (Tg)
U.S. Environmental Protection Agency Region
Month
1
2
3
4
5
6
7
8
9
10
11
12
Total
III
0.018
0.017
0.071
0.169
0.206
0.427
0.441
0.439
0.123
0.069
0.066
0.018
2.1
IV
0.092
0.139
0.428
0.460
0.475
0.874
0.903
0.903
0.461
0.286
0.162
0.080
5.3
V
0.004
0.004
0.067
0.189
0.240
0.550
0.568
0.568
0.137
0.066
0.063
0.004
2.5
VI
0.084
0.123
0.519
0.567
0.586
1.146
1.184
1.184
0.561
0.394
0.174
0.073
6.6
VII
0.007
0.007
0.078
0.211
0.226
0.508
0.524
0.524
0.136
0.026
0.025
0.007
2.3
VIII
0.022
0.020
0.108
0.303
0.362
0.809
0.836
0.820
0.280
0.290
0.110
0.022
4.0
IX
0.060
0.054
0.113
0.320
0.331
0.710
0.734
0.734
0.357
0.369
0.130
0.060
4.0
X
0.043
0.039
0.102
0.202
0.208
0.424
0.438
0.438
0.212
0.219
0.109
0.043
2.5
Total
0.3
0.4
1.5
2.4
2.6
5.5
5.6
5.6
2.3
1.7
0.8
0.3
29.1
Percent
of Total
1.1
1.4
5.1
8.3
9.1
18.7
19.3
19.3
7.8
5.9
2.9
1.1
Source: Lamb et al. (1993).
-------
not nearly as good. The uncertainty associated with the algorithms used to generate the U.S.
inventory described previously is estimated to be a factor of 3 (Lamb et al., 1987). Because
other sources of error in the inventorying process are much smaller, a factor of 3 is the
current best estimate of the overall uncertainty associated with biogenic VOC inventories.
However, this may be a lower limit if it is shown that oxygenated species are emitted in
significant quantities by vegetation. Emission measurement methods employed in the past
have not been adequate for quantifying polar, oxygenated organics.
Comparison ofManmade and Biogenic Emissions
The most recent anthropogenic and biogenic VOC emissions estimates for the
United States indicate that natural emissions (29 Tg) exceed manmade emissions (23 Tg).
During the summer months in the United States, anthropogenic emissions constitute 25% of
the annual anthropogenic VOC emissions, whereas biogenic emissions constitute 57% of the
annual biogenic emissions. On a teragram basis, anthropogenic VOC emissions during the
summer contribute 0.25 x 23 Tg = 5.75 Tg, whereas summer biogenic VOC emissions
contribute 0.57 x 29 Tg = 16.5 Tg. Therefore, on a national basis, the ratio of biogenic to
anthropogenic VOC emissions is approximately 2.9 for the United States. However, this ratio
varies with region in the summer months. These calculations depend on the assumption that
regional summertime anthropogenic VOC emissions are one-quarter of the annual VOC
emissions. With this assumption, the ratios of biogenic to anthropogenic VOC emissions for
three selected regions are as follows: 2.2 for Region IV, 1.6 for Region V, and 3.2 for
Region VI (Lamb et al., 1993; U.S. Environmental Protection Agency, 1994). However, in a
recent National Research Council review (1991), it was concluded that emissions from
manmade sources are currently underestimated by a significant amount (60 to 80%). Because
uncertainty in both biogenic and anthropogenic VOC emission inventories is large, it is not
possible to establish whether the contribution of emissions from natural or manmade sources
of VOCs is larger.
3.4.1.4 Relationship of Summertime Precursor Emissions and Ozone Production
Peak O3 levels are recorded in most regions of the country during the months of
June, July, and August. From the foregoing discussion, it is obvious that natural emissions of
NOX and VOCs peak during this same time frame. Biogenic emissions are very dependent on
temperature; and, as ambient temperatures rise during the summer months, NOX and VOC
emissions reach a maximum. Figure 3-24 clearly demonstrates this for biogenic VOC
emissions, and a plot of monthly biogenic NOX emissions would show a similar pattern. Well
over 50% of biogenic NOX and VOC emissions occur during the period of maximum
photochemical activity.
Seasonal changes in anthropogenic emissions of NOX are believed to be relatively
small. The transportation sector produces slightly less NOX during the warmer months, but
there is probably a small increase from the stationary source category because of higher
summertime power demands. Because these are the major U.S. sources of NOX and changes
in seasonal emissions tend to offset each other, there is no reason to expect that NOX
emissions will vary significantly by season on the national level. Evaporative emissions of
VOCs are enhanced during the warm summer months. Because evaporation is an important
component of anthropogenic VOC emissions, there is a summertime increase. In 1993, U.S.
VOC emissions during June, July, and August were estimated to exceed annual monthly
average VOC emissions by about 17% (U.S. Environmental Protection Agency, 1994). The
3-79
-------
234
Total NMHC Emissions (Tg)
Figure 3-24. Estimated biogenic emissions of volatile organic compounds in the
United States as a function of season.
Source: Fehsenfeld et al. (1992).
summertime anthropogenic VOC effect should be somewhat larger for southern regions. This
is a very small change, however, relative to the uncertainty associated with VOC emission
estimates. In an earlier discussion of the NAPAP inventory, VOC emissions from manmade
sources were considered to be almost independent of season.
Increases in O3 precursor emissions during the peak O3 season will have a
tendency to enhance O3 production. Ozone production in rural areas is usually NOX-limited
(Fehsenfeld et al., 1992). Thus, enhanced summertime emissions of NOX from soils and
lightning will add NOX to the atmosphere in rural regions, which in turn will lead to the
production of more O3. Larger summertime emissions of VOCs will enhance O3 production in
urban areas. Biogenic VOC sources in the vicinity of urban areas can contribute significant
quantities of reactive hydrocarbons to the urban O3 precursor mix (Cardelino and Chameides,
1990).
3.4.2 Concentrations of Precursor Substances in Ambient Air
The volatile organic compounds, excluding CH4, often are referred to as
NMOCs. The class of NMOCs most frequently analyzed in air are the nonmethane
hydrocarbons (NMHCs). The NMHC measurements often provide an acceptable
approximation of the NMOCs. The NMHCs and the NOX within urban areas tend to have
morning concentration peaks. These result from vehicular traffic in combination with limited
mixing depths. Later in the morning into the afternoon hours, concentrations of NMHCs and
NOX decrease, but to varying extents (Purdue et al., 1992), because of chemical reactions and
increases in mixing depths and consequent increases in dilution volumes. Photochemical
3-80
-------
atmospheric reactions also can rapidly convert NO to NO2, and hydrocarbons to carbonyls,
PANs, and other products (Sections 3.2.4, 3.4.2.1, and 4.9). Late afternoon and early
evening peaks might be expected in NMHC and NOX concentrations because of increased
vehicular traffic at urban locations, but such increases often are not discernible (Purdue et al.,
1992). This effect probably results from the presence of substantial mixing depths in the
warmer months that persist through these hours in many urban locations.
Because of the emphasis on early morning inputs of NMOCs and NOX for
models such as the Empirical Kinetics Modeling Approach (EKMA), most of the
measurements available emphasize the 6 a.m. to 9 a.m. period. The variations in the
concentrations of NMOCs and NOX, their ratios, and the composition of NMOCs are important
factors in the generation of O3 and other photochemical products.
3.4.2.1 Nonmethane Organic Compounds
In earlier measurements based on GC analyses made during a number of different
studies in urban areas between 1969 and 1983, the mean 6 a.m. to 9 a.m. NMHC
concentrations were reported to range from 0.324 to 3.388 ppm carbon (ppmC) (U.S.
Environmental Protection Agency, 1986a). The highest NMHC concentrations were those
measured at sites in Los Angeles.
A program for analysis of NMOCs and NOX in the months of June through
September was conducted in a considerable number of U.S. cities during the 1980s. The
results obtained from measurements made during the 6 a.m. to 9 a.m. period at sites in
22 cities in 1984 and 19 cities in 1985 have been subjected to statistical analysis and
interpretation (Baugues, 1986). The total NMOC measurements throughout the June through
September periods in these cities were obtained by the cryogenic preconcentration-direct flame
ionization detection method (PDFID) (McElroy et al., 1986). In addition, during about 15%
of the 6 a.m. to 9 a.m. periods, canister samples were collected for subsequent GC analysis
(Seila et al., 1989). In 1984, the lowest median NMOC value obtained was 0.39 ppm C from
measurements in Charlotte, NC, whereas the highest median NMOC value obtained was
1.27 ppmC from measurements in Memphis, TN. In 1985, the lowest median NMOC value
obtained was 0.38 ppmC from measurements in Boston, MA, whereas the highest median
NMOC value obtained was 1.63 ppmC in Beaumont, TX. The overall median values from all
urban sites were approximately 0.72 ppmC in 1984 and 0.60 ppmC in 1985 (Baugues, 1986).
The GC analyses made on samples collected in 1984, 1985, and 1986 have been reported (Seila
et al., 1989). The more abundant individual hydrocarbons include Q-Cg alkanes, C2-C5
alkenes, C6-C9 aromatics, and acetylene. Based on the 48 most abundant concentrations, the
overall median concentrations by class of hydrocarbon (NMHCs) were as follows: paraffins,
0.266 ppmC, 60% of total; aromatics, 0.166ppmC, 26% of total; olefins, 0.047 ppmC, 11%
of total; and acetylene, 0.013 ppmC, 3% of total (Seila et al., 1989). Additional individual
NMHCs (totaling about 0.100 ppmC) were detected at concentrations DO.002 ppm C each.
Most of these compounds were identified by class but not by structure.
Detailed hydrocarbon analyses for C2-C10 NMHCs were obtained during the
17 intensive days of the Southern California Air Quality Study (SCAQS) in 1987 (Lonneman et
al., 1989; Rasmussen, 1989; Stockberger et al., 1989). The average percentage ambient
composition from eight Southern California sites during 11 intensive sampling days of the
summer of 1987 by class of NMHCs were as follows: paraffins, 53.4; aromatics, 27.2;
olefins, 12.1; carbonyls, 7.7 (Main and Lurmann, 1993).
3-81
-------
In Atlanta, during the summer of 1990, hydrocarbon concentrations were measured
at six sites with automated GCs. Results were reported on 54 hydrocarbons, with 24-h average
concentrations ranging from 0.186 to 0.397 ppmC (Purdue et al., 1992).
A comparison of NMHC measurements made by GC analyses over a period of
years in Los Angeles and in the New York City area has been reported (Lonneman and Seila,
1993). In the Los Angeles area, the NMHC concentrations averaged 2.81 ppmC in 1968,
compared to 1.02 ppmC in 1987. In the New York City area, the NMHC concentrations
averaged about 1.1 ppmC in 1969, compared to 0.62 ppmC from 1986 to 1988. In both the
Los Angeles and New York areas, there were significant decreases in NMHC concentrations as
well as compositional changes in NMHCs during these years, with increases observed in the
percentage of paraffin hydrocarbons and decreases in the percentage of aromatic hydrocarbons
and acetylene (Lonneman and Seila, 1993).
Aldehydes and ketones occur in urban air as O3-oxidant precursors from emissions
such as vehicular exhaust and as products of reactions of OH radicals with NMHCs, reactions
of alkenes with O3, and, at night, reactions with NO3 radicals. Early morning aldehyde
concentrations have been predicted to result to a greater extent from atmospheric reactions of
alkenes than from emission of vehicular exhaust (Altshuller, 1993). During the day, aldehydes
and ketones are rapidly produced from reactions of OH radicals with aliphatic and aromatic
hydrocarbons and of alkenes with O3. Carbonyl concentrations tend to increase through the
daytime hours (Grosjean, 1982, 1988; Grosjean et al., 1993b).
Measurements of ambient air concentrations of carbonyls indicate the total loading
of aldehydes and ketones from all processes. Ambient urban air concentrations of HCHO and
total aldehydes were tabulated for the 1960 to 1981 period (Altshuller, 1983a). Subsequent
studies by 2,4-dinitrophenylhydrazine high-performance liquid chromatography
(DNPH-HPLC) techniques (Section 3.5.2.3) have shown consistently that HCHO and
acetaldehyde are the most abundant aldehydes; however, a number of other carbonyls
(including propanal, acrolein, acetone, butanal, crotonaldehyde, methyl ethyl ketone, pentanal,
hexanal, benzaldehyde, and tolualdehyde) also have been measured (Fung, 1989; Grosjean
1982, 1988, 1991; Kalabokas et al., 1988; Zweidinger et al., 1988). The ratios of HCHO to
acetaldehyde concentrations (in parts per billion volume) can vary from less than 0.5 in cities
in Brazil, where there is high use of ethanol fuels, up to 4.0 to 5.0 at a few urban sites
(Grosjean et al., 1993). However, at most urban sites, the ratios of HCHO to acetaldehyde
concentrations occur in the 1.0 to 3.0 range.
A compilation of the maximum, average range of HCHO concentrations from many
studies in Southern California carried out between 1960 and 1989 is available (Grosjean,
1991). A downward trend in HCHO concentrations occurs, probably because of decreased
production from precursor alkenes and decreased emission in vehicular exhaust (Sigsby et al.,
1987; Dodge, 1990). For example, the maximum HCHO concentrations decreased from above
100 ppbv in the 1960s to the 10- to 30-ppbv range during the last decade (Grosjean, 1991). In
other U.S. cities in the early 1980s, the maximum HCHO concentrations ranged from 5 to
45 ppb (Salas and Singh, 1986).
Several studies have reported concurrent morning hydrocarbon and carbonyl
concentrations in downtown Los Angeles (Grosjean and Fung, 1984); Raleigh, NC
(Zweidinger et al., 1988); and Atlanta (Shreffler, 1992; Grosjean et al., 1993b). The average
percentage of carbonyls relative to total NMHCs were reported as follows: Los Angeles, 3%;
Raleigh, 2%; and Atlanta, D2% (formaldehyde + acetaldehyde concentrations) at two different
3-82
-------
sampling sites. In SCAQS, carbonyls were measured at eight sites in the summer and five in
the fall of 1987 (Fung, 1989; Fujita et al., 1992). The average percentage of C, to
C6 carbonyls relative to NMHCs was 7.6% in the summer and 3.7% in the fall.
Compilations of NMHC concentrations of nonurban and remote locations are
available (U.S. Environmental Protection Agency, 1986a; Altshuller, 1989a). Total NMHC
concentrations reported ranged from less than 0.01 to 0.14 ppmC. At remote locations over
the Pacific, NMHC concentrations generally were less than 0.01 ppmC. Over both continental
and oceanic locations there can be contributions from biogenic sources of NMHCs.
Interest in the contribution of biogenic hydrocarbons has existed for many years,
and earlier work has been reviewed (Altshuller, 1983b). Photochemical modeling in the
United States predicts significant effects of biogenic hydrocarbons on O3 production
(Chameides et al., 1988; Roselle et al., 1991). Similar modeling of the effect of biogenic
hydrocarbons on O3 production within urban plumes over southeastern England predicted a 2-
to 8-ppb increase in plume and background O3 concentrations (MacKenzie et al., 1991).
Because of lower emissions of biogenic and lower overall NMOC/NOX ratios, O3 production
over southeastern England is predicted to be limited by the availability of anthropogenic
hydrocarbons.
Compilations of results of earlier measurements of isoprene and terpene
concentrations are available (Altshuller, 1983b; U.S. Environmental Protection Agency,
1986a). Average concentrations of isoprene ranged from 0.001 to 0.020 ppmC and terpenes
from 0.001 to 0.030 ppmC. When concurrent measurements of biogenic and anthropogenic
NMHCs were available, the biogenic NMHCs usually constituted much less than 10% of the
total NMHCs (Altshuller, 1983b).
Among more recent studies are two investigations of terpene and isoprene emissions
in the central valley of California and in Louisiana (Arey et al., 1991b; Khalil and Rasmussen,
1992). Both studies reported a large number of individual terpenes, measured by using
enclosure methods. When ambient air measurements were made, most of the terpenes
measured in the enclosures were not detectable (Khalil and Rasmussen, 1992). In ambient air,
isoprene was the predominate hydrocarbon, accounting on average for 70% of the biogenic
species and 36% of NMOCs. It has been concluded that the bag enclosure method can lead to
large overestimates in biogenic emissions (Khalil and Rasmussen, 1992).
In two other recent studies in deciduous forests, the isoprene oxidation products
were measured as well as isoprene itself (Pierotti et al., 1990; Martin et al., 1991). Both
studies report the ambient concentrations of methacrolein and methyl vinyl ketone. In an
investigation in a central Pennsylvania deciduous forest in the summer of 1988, average
midday concentrations of isoprene were in the 0.005- to 0.010-ppmC range; whereas the
corresponding concentrations of methacrolein and methyl vinyl ketone were in the 0.001- to
0.002-ppmC range (Martin et al., 1991). In the study conducted in California forests with
samples collected between noon and 4:00 p.m. in late spring and summer, the upper quartile of
isoprene concentrations was within the 0.010- to 0.025-ppmC range, whereas methacrolein
concentrations were within the 0.001- to 0.003-ppmC range, and methyl vinyl ketone
concentrations were within the 0.0005- to 0.0015-ppmC range (Pierotti et al., 1991).
Higher-molecular-weight semivolatile carbonyls have been measured in a number of
rural-remote areas (Juttner, 1986; Yokouchi et al., 1990; Nondek et al., 1992; Ciccioli et al.,
1993). The compounds identified include C5-C12 aliphatic aldehydes, aliphatic ketones, and
aromatic aldehydes. Comparisons of the measurement of these carbonyls relative to aromatic
3-83
-------
hydrocarbons in two studies indicated higher carbonyl concentrations and much lower aromatic
hydrocarbon concentrations in the rural-remote sites compared to the urban areas (Yokouchi et
al., 1990; Ciccioli et al., 1993). Widely varying natural sources have been associated with
these carbonyls, including emissions from forest species (Nordek et al., 1992) and short
vegetation (Ciccioli et al., 1993) and as secondary products of natural emissions of terpenes
(Ciccioli et al., 1993) or oleic acid (Yokouchi et al., 1990). Among other oxygenates reported
to be of natural origin are higher-molecular-weight alcohols (Juttner, 1986; Nordek et al.,
1992; Goldan et al., 1993). These oxygenates contribute to the "Other VOCs" category in the
biogenic emissions inventory (Section 3.4.1.3).
In an urban-scale study in Atlanta during the summer of 1990 (as part of the
Southern Oxidant Study), isoprene concentrations rose in late morning and into the afternoon,
with early evening peaks observed at residential and rural-residential sites (Purdue et al.,
1992). A similar diurnal profile for isoprene was observed at a Pennsylvania forest site
(Martin et al., 1991). The median concentration at the sampling sites in Atlanta early in the
evening ranged from 0.006 to 0.020 ppmC. The isoprene as a percentage of total NMHCs in
the early evening ranged among the sites from 2 to 12% (Shreffler, 1992).
3.4.2.2 Nitrogen Oxides
Measurements of NOX were obtained with continuous NOX analyzers at sites in
22 and 19 U.S. cities during the months of June through September of 1984 and 1985,
respectively. These results have been evaluated and the 6 a.m. to 9 a.m. values tabulated
(Baugues, 1986). In 1984, the lowest median NOX concentration of 0.010 ppm was obtained
from measurements in West Orange, TX; whereas the highest median NOX concentration of
0.088 ppm was obtained from measurements in Memphis. In 1985, the lowest median NOX
concentration of 0.005 ppm was obtained from measurements in West Orange, whereas the
highest median NOX concentration of 0.100 ppm was obtained from measurements in
Cleveland, OH. The median NOX concentration values for sites in most of these cities in 1984
and 1985 ranged between 0.02 and 0.08 ppm. Because of high vehicular emission rates and
shallow mixing depths, the median 6 a.m. to 9 a.m. concentration values in many of these
cities exceeded the annual average NOX values of 0.02 to 0.03 ppm in U.S. metropolitan areas
between 1980 and 1989 (U.S. Environmental Protection Agency, 199la). In the 1990 Atlanta
study, the average summer NOX concentration values at the six study sites ranged from
0.011 to 0.026 ppm (Purdue et al., 1992).
At nonurban sites, NOX concentrations have been reported as mean 24-h seasonal or
annual NOX values. The available results have been compiled for work reported through 1983
(Altshuller, 1986). The average seasonal or annual NOX concentrations ranged from less than
0.005 to 0.015 ppm. At remote sites in the earlier investigations, monthly average NOX
concentrations were less than 0.001 ppm. In more recent work, the statistics on NOX
concentrations have been reported for several relatively remote U.S. sites (Fehsenfeld et al.,
1988). The 24-h average NOX concentrations and the range in the central 90% of values were
as follows: Point Arena, CA, spring 1985, 0.0004 ppm, 0.0007 to 0.001 ppm; Niwot Ridge,
CO, summer 1985, 0.0005 ppm, 0.0001 to 0.002 ppm; and Scotia, PA, summer 1986,
0.002 ppm, 0.0007 to 0.009 ppm. It should be noted that each of these sites can be subject to
anthropogenic influences, thus accounting for the higher NOX values. For example, at Niwot
Ridge, CO, with upslope flow from the Denver-Boulder, CO, urban area, higher NOX
3-84
-------
concentrations are measured. Nitrogen oxide concentrations at or below 0.0001 ppm occur at
other remote surface locations (Fehsenfeld et al., 1988).
3.4.2.3 Ratios of Concentrations of Nonmethane Organic Compounds to Nitrogen
Oxides
The ratios of 6 a.m. to 9 a.m. NMOC/NOX have been obtained from the
measurements in the U.S. cities discussed above (Baugues, 1986). In 1984, the lowest median
NMOC/NOX ratio of 9.1 was obtained in Cincinnati, OH, and the highest median NMOC/NOX
ratio of 37.7 was obtained in Texas City, TX. In 1985, the lowest median NMOC/NOX ratio
of 6.5 was obtained in Philadelphia, PA, whereas the highest median NMOC/NOX ratio of
53.2 was obtained in Beaumont, TX. The range in daily 6 a.m. to 9 a.m. NMOC/NOX ratios
within a given city is large, with 10th percentile to 90th percentile NMOC/NOX ratios varying
usually by factors of 2 to 4 and, at several sites, by factors of 5 to 10 (Baugues, 1986). There
appears to be a tendency for higher NMOC/NOX ratios in the cities included in the southeastern
(9) and southwestern (15) United States than in the northeastern (7) and midwestern United
States (7) (Altshuller, 1989b). The NMOC-to-NOx ratios at rural sites tend to be higher than
the mean NMOC-to-NOx ratios in urban locations, with mean values at several rural sites
ranging between 20 and 40 (Altshuller, 1989b).
In SCAQS, the ambient NMOC (NMHCs + carbonyl)/NOx ratios averaged 8.8 in
the summer and 6.9 in the fall of 1987 (Fujita et al., 1992). However, the 6 intensive days in
the fall between November 11 and December 11 were not characterized by elevated
O3 concentrations (Zeldin, 1993). These ambient ratios were 2 to 2.5 times higher than the
corresponding emission inventory ratios. Discrepancies as large or larger have been discussed
previously for urban and rural NMHC/NOX ambient-to-emission ratios in the eastern United
States (Altshuller, 1989b).
A trend analysis of NMHC/NOX ratios in the South Coast Air Basin is available for
the 1976 to 1990 period (Fujita et al., 1992). The ratios were consistently higher in the
summer than in the fall. These ratios started decreasing slowly during the 1980s, from
maximum ratios of about 12 in the summer and 9 in the fall to 8.5 in the summer and 7 in the
fall by 1990. The ambient-to-emission inventory ratios over this period ranged from as high as
3.4 in the summer to 1.7 in the winter (Fujita, 1993).
Interest in the 6 a.m. to 9 a.m. NMOC/NOX ratios is associated with their use in the
EKMA type of trajectory model (Section 3.6.1.2). The analysis at 10 eastern and midwestern
sites of upper-quartile O3 days relative to other O3 days indicated a significant difference
(p D 0.10) by the two-sample Wilcoxon Rank Sum test at four of the 10 sites with NMOC/NOX
ratios (Wolff and Korsog, 1992). However, the correlation of NMOC/NOX ratios with
maximum 1-h O3 concentrations was very weak. It was concluded that the use of the 6 a.m. to
9 a.m. NMOC/NOX ratio in EKMA will not provide sufficient information to distinguish
among NMOC, NOX, or combined VOC-NOX strategies as optimum strategies for urban areas.
A compilation of VOC/NOX ratios between 1981 and 1988 in 10 cities in the
northeastern and midwestern United States presents ratios ranging from 5.8 to 11.5, but
generally below 10 (Wolff, 1993). Trends between 1986 and 1991 in VOC/NOX ratios in four
of these cities; New York; Newark, NJ; Philadelphia; and Washington, DC, show downward
trends towards VOC/NOX ratios between 4 and 6 (Zalewsky et al., 1993; Wolff, 1993). For
Philadelphia, and the other sites, the downward trend in VOC/NOX is associated with
decreasing VOC concentrations with little change in NOX (Zalewsky et al., 1993; Wolff,
3-85
-------
1993). It has been pointed out that, in the National Academy of Science report (National
Research Council, 1991), hydrocarbon control is considered more effective for VOC/NOX
ratios of about 10 or less, whereas NOX control is considered more effective for VOC/NOX
ratios of 20 or more. Based on these results, it may be concluded that, in many cities in the
northeastern and midwestern United States, continued VOC control, rather than NOX control,
will be more effective in reducing O3 (Wolff, 1993). It also has been concluded by Wolff
(1993) that models with greater spatial resolution than the Regional Oxidant Model (ROM),
such as the UAM, are more applicable than ROM for determining appropriate O3 control
strategies in urban areas.
3.4.3 Source Apportionment and Reconciliation
3.4.3.1 Source Apportionment
Source apportionment refers to determining the quantitative contributions of sources
to ambient air pollutant concentrations. In principle, it includes two fundamentally different
approaches: (1) source-oriented and (2) receptor-oriented. In the source-oriented approach, a
mathematical dispersion model is applied to an emissions inventory and meteorological data to
produce an estimate of ambient pollutant concentrations that can be expected at a specified
point in space and time. In contrast, the receptor-oriented approach depends on simultaneous
ambient concentration measurements of a variety of pollutant species and a knowledge of the
relative amounts of the species (source profiles) that are present in the emissions of the sources
that are potential contributors. A mathematical receptor model operates on the source profile
and ambient species concentration information to deconvolute the ambient concentrations into
their source contributions, without the need of emissions inventory or meteorological
information. Indeed, the desire to avoid the latter two kinds of information, whose acquisition
is often problematical, has been an important motivation in the development of the
receptor-oriented approach.
Although source apportionment, in its general sense, embraces both approaches, in
recent years, it has come to be regarded as synonymous with the receptor-oriented approach
(receptor modeling). The equivalence of source apportionment and receptor modeling is
assumed in the following discussion. The most recent review of the field of receptor modeling
has been given by Gordon (1988).
Because tropospheric O3 is a secondary pollutant, the natural role of receptor
modeling is in determining the quantitative source contributions of the VOC precursors of O3.
Historically, receptor modeling was first developed in the 1970s for the apportionment of
ambient aerosol, and aerosol applications since then have been more extensive than VOC
applications. The aerosol and VOC areas of receptor modeling application have more
similarities than differences, however, so that much of the mathematical apparatus that has
been developed for aerosol problems is readily adaptable to VOCs.
For reasons that will become apparent, the separation of emissions sources into
anthropogenic and biogenic classes is a natural division for VOC receptor modeling and is used
in the following.
Manmade Sources of Volatile Organic Compounds
A principal approach for receptor modeling of anthropogenic VOC sources is that
of "mass balance". In this approach, a particular linear combination of source profiles is
sought that best approximates (in a linear least-squares sense) the profile of VOC species
3-86
-------
concentrations measured in an ambient sample. Here a VOC source profile is defined as the
set of numbers giving the fractional amounts (abundances) of individual species in the
emissions from the source. The profile may be normalized to the sum of the abundances of all
VOC species emitted by the source or to a sum over some arbitrary subset of species. For the
linear combination of profiles that gives the best fit, the coefficients are the source strengths (in
the same units as the measured ambient concentrations) associated with each of the included
source profiles.
Early efforts to use various versions of the mass balance approach include Ehrenfeld
(1974), Mayrsohn and Crabtree (1976), and Mayrsohn et al. (1977) in Los Angeles and Nelson
et al. (1983) in Sydney, Australia.
Of these studies, the work of Mayrsohn et al. (1977) is the most comprehensive:
900 samples from eight sites collected during June to September, 1974. The average results
were automotive exhaust, 53%; whole gasoline evaporation, 12%; gasoline headspace vapor,
10%; commercial natural gas, 5%; geogenic natural gas, 19%; liquefied natural gas, 1%. The
percentages are for NMHCs through C10 (i.e., not all of the total VOCs).
Together, the estimates for the first three vehicle-related sources account for 75 %
of the ambient NMHCs, which is the approximate percentage estimated in the other studies
listed. Geogenic natural gas is obviously not anthropogenic but is included here for
completeness. Its strength of 19% is striking; however, it seems unlikely that a contribution
this large would be typical of other locales lacking a petroleum-related geology. In any case,
accounting for the urban atmospheric concentrations of ethane and propane (the main NMHC
constituents of natural gas) has remained an unsatisfactorily resolved problem, so the 19%
result for geogenic natural gas has to be regarded skeptically.
Although dated, these early studies are of more than just historical interest. In one
respect, they are superior to more recent studies in their recognition of two distinctly different
kinds of gasoline evaporation: (1) headspace vapor, which represents the partial evaporation
of gasoline in situations such as storage tank evaporation or vehicle diurnal evaporation,
characterized by an enrichment of high-volatility species; and (2) whole gasoline emissions,
which can arise from spillage, leakage, and vehicle hot-soak emissions, and has a composition
resembling liquid gasoline itself. The implications of gasoline evaporation are discussed
below.
In the mid-1980s, a useful degree of standardization was incorporated into the mass
balance approach by the introduction of EPA's chemical mass balance (CMB) software. The
current version, CMB7 (Watson et al., 1990), embodies a comprehensive treatment of error
(including uncertainty in both ambient data and source profiles) and many diagnostics
(including profile collinearity) and has been used frequently in recent VOC receptor modeling
studies.
Recent studies include Wadden et al. (1986) in Tokyo, Japan; O'Shea and Scheff
(1988) in Chicago, IL; Aronian et al. (1989) in Chicago; Sweet and Vermette (1992) in
Chicago and East St. Louis, IL; Harley et al. (1992) in Los Angeles; Kenski et al. (1993) in
Chicago, Beaumont, Detroit, Atlanta, and Washington, DC; Spicer et al. (1993) in Columbus,
OH; and Lewis et al. (1993) in Atlanta.
The source categories covered by these studies taken together include vehicle
exhaust, gasoline evaporation (whole gasoline and headspace vapor), industrial emissions
(refineries, coke ovens, and chemical plants), architectural coatings, dry cleaning, wastewater
treatment, auto painting, industrial solvents/degreasers, graphic arts (printing), and natural
3-87
-------
gas. Each study gives estimates for the percentage contributions to measured ambient VOCs
(or related quantity) for a selected subset of these source categories. The one exception is the
work of Sweet and Vermette (1992) that estimates the percentage source contributions to
individual species, rather than to total VOCs. Such species apportionment is always available
from the CMB calculations, but often is not reported explicitly.
Usually, the source profiles used were generic; that is, from compilations of source
measurements taken elsewhere (U.S. Environmental Protection Agency, 1993d). The work of
Lewis et al. (1993) is unique in the use of profiles extracted from the ambient air data
themselves.
Generally, for these urban-based studies, vehicle exhaust is found to be the
dominant contributor to ambient VOCs. Exceptions are the Tokyo results of Wadden et al.
(1986) that show an unreasonably small average contribution of 7% and the Beaumont results
of Kenski et al. (1993), 14%. For all the rest, the average vehicle exhaust results fall in the
range of 45 + 15%.
The results for gasoline evaporation contribution estimates are much less
satisfactory. This is because the recent studies, with the exceptions of Harley et al. (1992) and
Lewis et al. (1993), included a gasoline headspace vapor profile but not a whole gasoline
profile in their calculations. The latter two studies suggest that this omission is a serious error.
For example, Harley et al. (1992) found a remarkably large whole gasoline contribution
(nearly the same as that of vehicle exhaust), and Lewis et al. (1993) find a whole gasoline
contribution that is about 20% that of vehicle exhaust. Both, however, find a whole gasoline
contribution about four times greater than the headspace contribution. Because vehicle exhaust
and whole gasoline profiles are quite similar (except for the very light species that are absent in
gasoline but present in exhaust as combustion products), excluding the whole gasoline profile
will tend to overestimate the exhaust contribution. Although this error may not greatly affect
the total mobile-source-related emissions estimate, it is misleading with regard to implied
control strategies.
Beyond the ubiquitous vehicle-related contributions, other anthropogenic source
contribution estimates tend to be smaller or locale-specific.
Biogenic Sources of Volatile Organic Compounds
The possible role of biogenic VOC emissions in O3 formation is being considered
much more seriously now (Chameides et al., 1988) than was the case a decade ago. Because
of the severe experimental problems in accurately measuring biogenic emissions directly,
receptor modeling approaches are of considerable interest. Compared with anthropogenic
sources, however, the application of receptor modeling methodology to biogenic sources has
been very limited. The principal reason is that it has not been possible to find VOC species
that are simultaneously distinctive components of biogenic emissions, emitted in an
approximately fixed proportion to the total VOC biogenic emissions, and relatively unreactive.
Without these conditions, the construction of a credible stable biogenic source profile is not
possible, and, consequently, the CMB approach is unusable.
In this situation, a crude form of receptor modeling has been used in which the
ambient concentration of a VOC species, whose only source is thought to be biogenic, is
divided by the estimated abundance of the species in the total VOC biogenic emissions.
Typical candidates include isoprene (deciduous emission) and the terpenes D- and D-pinene
(coniferous emission), D-caranene, and limonene. Because these are all highly reactive, any
3-88
-------
such estimate can be regarded only as a lower limit of the contribution that biogenic emissions
make to total ambient VOC, if the loss resulting from atmospheric transformation is not taken
into account. As an example, Lewis et al. (1993) used isoprene, the most prominent biogenic
species measured in downtown Atlanta during the summer of 1990, to infer a lower limit of
2% (24-h average) for the biogenic percentage of total ambient VOC at that location. Isoprene
emissions have a strong diurnal dependence. Lower limits for biogenic emissions at other
hours, inferred from average isoprene concentrations, were 1% at 8:00 a.m., 5% at noon, 6%
at 4:00 p.m., and 2% at 9:00 p.m.
The recent review article by Fehsenfeld et al. (1992) lists other prominent biogenic
species and calls attention to the newly recognized importance of alcohols, such as methanol
(CH3OH), as biogenic primary emissions. Goldan et al. (1993) reported the C5 alcohol,
2-methyl-3-buten-2-ol to be the most abundant VOC of biogenic origin present in a
predominantly lodgepole pine forest in Colorado. Ciccioli et al. (1993) present data from sites
in Germany and Italy showing substantial contributions from various aldehydes and argue that
their dominant source is biogenic primary emissions, rather than photochemical oxidation
products.
A more sophisticated form of biogenic receptor modeling involves the radiocarbon
isotope 14C. The approach depends on the fact that 14C constitutes a nearly fixed fraction
(approximately 10"12) of all carbon present throughout the biosphere. In contrast, the 14C in
dead organic material older than 40,000 years, certainly the case for fossil fuels, has been
reduced by at least 99% through radioactive decay. This leads to a simple estimate of the
biogenic fraction of a carbon-containing sample given by fs/f0 , where fs is the 14C fraction in
the sample, and f0 is the 14C fraction in living material. Besides its conceptual simplicity, the
approach is appealing for VOC apportionment because 14C retains its identity in the reaction
products that may result from atmospheric transformation of reactive VOC. The method
appears to be reliable for particulate phase organics (Lewis et al., 1988, 1991) but is still under
development for VOC applications (Klouda et al., 1993).
3.4.3.2 Source Reconciliation
Source reconciliation refers to the comparison of measured ambient VOC
concentrations with emissions inventory estimates of VOC source emission rates for the
purpose of validating the inventories. Because concentrations and emission rates are specified
in different units, the comparisons are done in terms of percentages: the percentage
contribution of a source to ambient total VOCs as estimated by receptor modeling versus the
emission rate of the source as a percentage of the total VOC emission rate of the inventory.
Nearly all the receptor modeling studies listed above have included such a
percentage comparison. Typically, the agreement is quite good for vehicle exhaust, generally
the dominant VOC source in urban airsheds. Gasoline evaporation comparisons are much less
consistent, at least partly for the reasons already indicated. Typically, there is at least
qualitative agreement for the other anthropogenic sources: they are small in the inventory, and
the receptor-estimated contributions are small. An interesting exception is refinery emissions
in Chicago (Scheff and Wadden, 1993), for which the receptor estimate was 7%, five times
greater than the inventory estimate. Another is the significant (5 to 20%) natural gas/propane
contribution estimated in Los Angeles, Columbus, and Atlanta but not reflected in their
inventories. The few biogenic source estimates provided by receptor modeling are generally
smaller than those given in emissions inventories, at least partly because of the previously
3-89
-------
referred to reactivity problem. Credible 14C measurements on VOC samples would be
extremely helpful in validating the magnitude of the biogenic component of emissions
inventories.
Lewis et al. (1993) has noted that comparisons based on percentages are quite
insensitive for dominant source components, and the comparisons are more dependent on how
"total VOC" is defined than is often appreciated (the definition varies for the studies listed).
Thus, unfortunately, the generally good agreement (receptor versus inventory estimates) found
for vehicle exhaust does not translate into a definitive judgment on the current concern that this
source component may be significantly underestimated in existing inventories. For example, if
the emission rate of vehicle exhaust in a typical inventory were arbitrarily doubled, the
resulting change in the percentage of this component in the inventory is well within the range
of what can be produced in the receptor estimate by merely choosing a different definition of
total VOC from plausible alternatives. Such alternatives relate to questions such as which
subset of hydrocarbons are summed? and whether unidentified chromatographic components
are included in the sum? In the future, this situation can be improved by more consistency in
the total VOC definition and by transforming the receptor modeling results from a
concentration-based representation to an emission-rate one. This unavoidably involves
introducing some limited meteorological information (Lewis and Conner, 1991).
3.5 Analytical Methods for Oxidants and Their Precursors
3.5.1 Sampling and Analysis of Ozone and Other Oxidants
3.5.1.1 Ozone
Introduction
The measurement of O3 in the atmosphere has been a subject of research for decades
because of the importance of this compound in atmospheric chemistry and because of its
potential and demonstrated effects on human health and welfare.
Because of the importance of O3 in the air of populated regions, widespread
O3 monitoring networks have been operated for many years, and the development of
measurement and calibration approaches for O3 has been reviewed extensively (e.g., U.S.
Environmental Protection Agency, 1986a). This section focuses on the measurement of O3 in
the ambient atmosphere at ground level and summarizes the current state of ambient
O3 measurement and calibration. No attempt is made here to cover the full history of
development of these methods because that has been documented elsewhere (e.g., U.S.
Environmental Protection Agency, 1978, 1986a). Instead, this section concentrates on those
methods currently used and on new developments and novel approaches to O3 measurement.
Although no method is totally specific for O3, current methods for O3 must be
distinguished from earlier methods that measured "total oxidants". The wet chemical methods
used earlier for total oxidants have been replaced for essentially all ambient measurements by
two more specific instrumental methods based on the principles of chemiluminescence and UV
absorption spectrometry. These two approaches are described below. In addition, recent
developments in spectroscopic measurements, in other chemical approaches, and in passive
sampling devices for O3 are described.
Chemiluminescence Methods
3-90
-------
Gas-Phase Chemiluminescence. The most common chemiluminescence method
for O3 is direct gas-phase reaction of O3 with an olefin to produce electronically excited
products, which decay with the emission of light. This approach was first used nearly 30 years
ago for chemical analysis by Nederbragt (Nederbragt et al., 1965), and development of a
portable monitor (Warren and Babcock, 1970) and application to atmospheric measurements
(Stevens and Hodgeson, 1970) followed soon after. Typically, an O3 monitor based on this
approach functions by mixing a constant flow of about 1 L/min of sample air with a small
constant flow (D50 cm3/min) of ethylene. Mixing occurs in a small inert reaction chamber
fitted with a sealed window through which light can pass to the photocathode of a
photomultiplier tube. Electronically excited formaldehyde molecules, generated by a small
fraction of the O3-ethylene reactions, produce a broad band of emission centered at 430 nm.
The emission intensity is linearly proportional to the O3 concentration over the range of 0.001
ppm to at least 1 ppm. Calibration of the monitor with a known O3 source provides the
relationship between monitor response and O3 concentration. Detection limits of 0.005 ppm
and a response time of less than 30 s are easily attained, and are typical of currently available
commercial instruments.
Although no interference has been found from common atmospheric pollutants, a
positive interference from atmospheric water vapor has been reported (California Air
Resources Board, 1976; Kleindienst et al., 1993, and references therein) and has recently been
confirmed (Kleindienst et al., 1993; Hudgens et al., 1994). The recent results indicate a
positive interference of about 3% per percent H2O by volume at 25 DC, based on tests at O3
concentrations of 0.085 to 0.32 ppm, and at H2O concentrations of 1 to 3% (i.e., dew point
temperatures of 9 to 24 DC). It has been estimated that the interference of water in ethylene
chemiluminescent measurements at 30 DC and 60% relative humidity could be as high as
13 ppbv of O3, or 11 % of the O3 reading at 120 ppbv (Kleindienst et al., 1993). Calibration
with known O3 concentrations in air of temperature and humidity, similar to that of the sample
air, can minimize this source of error.
A separate potential problem with the ethylene chemiluminescent method is leakage
of the pure ethylene reagent gas. Because O3 and hydrocarbon measurements are often
co-located for monitoring purposes, leakage of ethylene could cause difficulty in obtaining
valid measurements of total nonmethane hydrocarbons (TNMHC) in ambient air.
The measurement principle set forth by EPA for compliance monitoring for O3 is
the chemiluminescence method using C2H4 (Federal Register, 1971). Methods of testing and
the required performance specifications that commercial O3 monitors must meet to be
designated a reference or equivalent method are documented (Federal Register, 1975).
A monitor may be designated a reference method if it employs gas-phase chemiluminescence
with C2H4 as the measuring principle and achieves the required performance specifications.
An equivalent method must show a consistent relationship with the reference method and must
meet the required performance specifications. Table 3-15 shows those specifications for
O3 monitors. Note that ethylene chemiluminescence monitors typically have response times far
superior to those required in Table 3-15.
The list of commercial O3 monitors designated as reference or equivalent methods
by EPA is shown in Table 3-16 (updated as of August 1, 1994). Details on three monitors not
described in the 1986 EPA criteria document for O3 and other oxidants are presented in Table
3-17. All of the reference methods are C2H4 chemiluminescence instruments, as required by
the definition of a reference method. The equivalent methods are based on either gas-solid
3-91
-------
chemiluminescence or UV-absorption measurements. Those methods are described below. A
gas-liquid chemiluminescence analyzer for O3, which may be submitted for EPA equivalency
in the near future, also is described below.
Gas-Solid Chemiluminescence. The reaction of O3 with Rhodamine-B adsorbed
on activated silica gel produces chemiluminescence in the red region of the visible spectrum.
This was the first chemiluminescence method ever developed for ambient O3 measurement
3-92
-------
Table 3-15. Performance Specifications for Automated Methods of Ozone Analysis
Performance Parameter Units Specification
Range ppm 0 to 0.5
Noise ppm 0.005
Lower detectable limit ppm 0.01
Interference equivalent
Each interference ppm +0.02
Total interference ppm 0.06
Zero drift, 12 and 24 h ppm +0.02
Span drift, 24 h
20% of upper range limit % +20.0
80% of upper range limit % +5.0
Lag time min 20
Rise time min 15
Fall time min 15
Precision
20% of upper range limit ppm 0.01
80% of upper range limit ppm 0.01
Source: Federal Register (1975); Code of Federal Regulations (1994).
(Regener, 1960, 1964). The emitted light intensity is linearly related to the O3 concentration,
and the detection limit can be as low as 0.001 ppm. No direct interferences from other gas-
phase pollutants are known; however, decay of the sensitivity because of surface aging can
occur (Hodgeson et al., 1970). Addition of gallic acid to the surface stabilizes the response
characteristics, apparently by allowing direct reaction of O3 with the gallic acid, rather than
with the Rhodamine-B (Bersis and Vassiliou, 1966). A commercial analyzer (Phillips Model
PW9771) based on this approach has been designated an equivalent method for ambient O3 (see
Table 3-16), but gas-solid chemiluminescence currently is used rarely for ambient
measurements.
Gas-Liquid Chemiluminescence. A recently developed commercial monitor uses
the chemiluminescent reaction of O3 with the dye eosin-Y in solution (Topham et al., 1992).
The monitor functions by exposing a fabric wick, wetted with the eosin-Y solution, to a flow
of sample air within view of a red-sensitive photomultiplier tube. The monitor, designated the
LOZ-3, is compact, portable, and requires no reagent gases. The LOZ-3 provides very fast
response: a lag time of 2 s, a rise time of 3 s, and a fall time of 2 s, all relative to a step
change of 400 ppbv O3, are reported (Topham et al., 1992). Instrument noise at zero and at
382 ppbv ozone is 0.05 ppbv or less, calculated as the standard deviation of 25 successive
2-min averages. The precision of the LOZ-3 is reported to be 0.80 ppbv at 100 ppbv O3 and
1.87 ppbv at 400 ppbv O3, both calculated as one standard deviation of six repeated
measurements at these levels (Topham et al., 1992). The instrument provides linear response
up to 200 ppbv, with a gradually decreasing slope of the response curve above that
3-93
-------
Table 3-16. Reference
Designated by the U.S.
and Equivalent Methods for Ozone
Environmental Protection Agencf
Principal Method
Designation
Number
Method
Code
Reference Methods
(Ethylene Chemiluminescence)
Beckman 950A
Bendix 8002
CSI 2000
McMillan 1100-1
McMillan 1100-2
McMillan 1100-3
Meloy OA325-2R
Meloy OA350-2R
Monitor Labs 8410E
RFOA-
RFOA-
RFOA-
RFOA-
RFOA-
RFOA-
RFOA-
RFOA-
RFOA-
0577-020
0176-007
0279-036
1076-014
1076-015
1076-016
1075-003
1075-004
1176-017
020
007
036
514
515
016
003
004
017
Equivalent Methods
(UV Absorption)
Advanced Pollution Instrument 400
Dasibi 1003-AH,-PC,-RS
Dasibi 1008-AH,-PC,-RS
Environics 300
Lear-Siegler ML9810
Monitor Labs 8810
PCI Ozone Corporation LC-12
Thermo Electron 49
Equivalent Methods (Gas/Solid CL)
Philips PW9771
EQOA-0992-087
EQOA-0577-019
EAOA-0383-056
EQOA-0990-078
EQOA-0193-091
EQOA-0881-053
EQOA-0382-055
EQOA-0880-047
EQOA-0777-023
087
019
056
078
091
053
055
047
023
aAs of August 1, 1994; see Appendix A for abbreviations and acronyms.
level. Temperature and pressure sensitivity are corrected by internal circuitry (Topham et al.,
1992). An initial large positive interference from SO2 that becomes smaller and negative as the
eosin solution ages is reported, and a positive interference from CO2 is also present. Topham
et al. (1992) report that a pretreatment technique applied to the eosin reagent solution
minimizes both of these interferences. Several of the performance characteristics of the LOZ-3
are impressive, but verification of the reported interference levels and the effectiveness of
temperature and pressure corrections appears to be needed.
3-94
-------
Table 3-17. List of Designated Reference and Equivalent Methods for Ozone
Federal Register
Designation
Number
Identification
Source
Manual
or Auto
Ref. or
Equiv.
Vol.
Page
Notice Date
EQOA-0990-078
u>
-------
Table 3-17 (cont'd). List of Designated Reference and Equivalent Methods for Ozonfe
Federal Register
Designation
Number
Identification
Source
Manual
or Auto
Ref. or
Equiv.
Vol.
Page
Notice Date
EQOA-0193-091
u>
-------
This method is undergoing testing and is likely to be submitted for EPA certification as an
equivalent method.
Ultraviolet Photometry
This method is based on the fact that O3 has a reasonably strong absorption band
with a maximum near 254 nm, coinciding with the strong emission line of a low-pressure
mercury lamp. The molar absorption coefficient at the mercury line is well known, the
accepted value being 134 (+2) IVT'cnr1 in base 10 units at 0 DC and 1 atmosphere pressure
(Hampson et al., 1973). Ultraviolet absorption has frequently been used to measure O3 in
laboratory chemical and kinetics studies. Ultraviolet photometry also was used for some of the
first atmospheric O3 measurements, but the early instruments suffered from poor precision
because of the small absorbances being measured (U.S. Department of Health, Education, and
Welfare, 1970).
Modern digital electronics have now solved the precision problems resulting from
measurement of small absorbances, and several commercial O3 monitors now employ
UV photometry. Several instruments based on this principle have been designated by EPA as
equivalent methods for ambient O3 (Tables 3-16 and 3-17). Ultraviolet photometry is now the
predominant method for assessing compliance with the NAAQS for O3. The commercial
monitors use pathlengths of 1 m or less, and operate in a sequential single-beam mode.
Transmission of 254-nm light through the sample air is averaged over a short period of time
(as short as a few seconds) and is compared to a subsequent transmission measurement on the
same air stream from which O3 has been selectively removed by a manganese dioxide (MnOj)
scrubber. The electronic comparison of the two signals can be converted directly into a digital
readout of the O3 concentration. The method is in principle absolute, because the absorption
coefficient and pathlength are accurately known, and the measured absorbance can be
converted directly to a concentration.
Commercial UV photometers for ambient O3 measurements have detection limits of
approximately 0.005 ppm. Time response depends on the averaging time used, but is typically
< 1 min. Long-term precision can be within +5%. The method has the advantage of
requiring no gas supplies, and commercial instruments are compact and reasonably portable.
Sample air flow control is not critical, within the limitations of the MnO2 scrubber. Because
the measurement is absolute, UV photometry also is used to assay O3 calibration standards as
in Section 3.5.1.1. Ambient air monitors using UV photometry are generally calibrated with
standard O3 mixtures to account for losses of O3 in sampling lines.
A potential disadvantage of UV photometry is that any atmospheric constituent that
absorbs 254-nm light and is removed fully or partially by the MnO2 scrubber will be a positive
interference in O3 measurements. Potential interferents include aromatic hydrocarbons,
mercury vapor, and SO2. A recent study (Kleindienst et al., 1993) demonstrated that toluene
and possibly aromatic reaction products, such as benzaldehyde, produce positive interferences
in UV photometric O3 measurements. This result was found using photochemically reactive
mixtures of toluene and NOX, at concentrations at a factor of 2 to 5 higher than those expected
in polluted urban air. Consideration of the relative absorption coefficients of O3 and the
aromatics indicated that, at higher humidities, toluene can cause an interference of 0.1 ppbv O3
per ppbv of toluene, whereas benzaldehyde may cause an interference as high as 5 ppbv O3 per
3-97
-------
ppbv benzaldehyde (Kleindienst et al., 1993). This interference may be humidity dependent.
In earlier work at very low humidities, no interference was observed with toluene, and only a
very small interference was observed with benzaldehyde (Grosjean and Harrison, 1985b).
However, even at very low humidities, these investigators observed significant interferences
from styrene, cresols, and nitrocresols. Evaluation of aromatic interference is limited by a
lack of appropriate absorption spectra in the 250-nm range and by a lack of ambient
measurements of most of the aromatic photochemical reaction products. The use of C2H4
chemiluminescence monitors in areas where aromatic concentrations are substantial has been
suggested (Kleindienst et al., 1993).
The same study found no consistent effect of ambient water vapor on measured
O3 concentrations using UV photometry, in contrast to the effect noted using C2H4
chemiluminescence (Kleindienst et al., 1993). However, short-term disturbances in UV
photometric O3 readings were observed when the humidity of the sample air was changed
substantially within a few seconds. This finding corroborates the observations of Meyer et al.
(199la) in an earlier study that indicated microscopic irregularities in the UV cell windows as
the cause of such disturbances. This effect should be absent in UV photometric measurements
of ambient O3 at the ground but could be important in other applications, such as measuring
vertical O3 profiles from an aircraft (Kleindienst et al., 1993).
A different approach to evaluating potential interferences in O3 measurements was
taken by Leston and Ollison (1993). These investigators examined ambient O3 data from
instruments of different measurement principles co-located at monitoring sites. The focus of
their study was the O3 "design value", the fourth highest daily maximum hourly value from a
monitoring station within an urban area, which is established in the 1990 Clean Air Act
Amendments (CAAA) (U.S. Congress, 1990) as the basis for classification of the area relative
to attainment of the NAAQS for O3. Leston and Ollison (1993) examined hourly
O3 concentration data from co-located UV and QH^ chemiluminescence instruments, from
1989 and 1990 at a site in Madison, CT, and from shorter periods at sites in East Hartford,
CT, and Mobile, AL. They also examined 11 winter days of simultaneous O3 data from UV
and Luminox LOZ-3 instruments, from Long Beach, CA. Leston and Ollison (1993) reported
positive biases in the UV data of 20 to 40 ppbv O3 during "hot, humid, hazy conditions typical
of design value days." They proposed that most O3 data and all design values are biased high
by known and suspected interferences, and that those interferences are exacerbated by water
vapor. Leston and Ollison (1993) argue that the interference in UV measurements from
benzene derivatives (e.g., styrene, cresols, benzaldehyde, nitro-aromatics) is poorly accounted
for. For example, of these compounds, only styrene is measured in the Photochemical
Aerometric Monitoring Stations (PAMS) VOC monitoring network (Leston and Ollison,
1993).
An experimental study by Hudgens et al. (1994) attempted to address the issues
raised by Leston and Ollison (1993) by evaluating several aspects of both UV and
chemiluminescence O3 measurements. This study confirmed the positive interference of water
vapor in the chemiluminescence method as 3% per percent water by volume (Kleindienst et al.,
1993) and also confirmed that no comparable interference exists with the UV method.
However, Hudgens et al. (1994) also showed that some UV instruments give noisier response
when operated under conditions in which condensation of moisture may occur in the sampling
lines, as in an air conditioned enclosure during hot, humid weather. Hudgens et al. (1994)
also tested several aromatic hydrocarbons for both absorbance at 254 nm and behavior in the
3-98
-------
O3 scrubber of UV instruments. Both positive and negative potential interferences were found;
the former by adsorption of UV-absorbing aromatics in the scrubber, and the latter by release
of those stored compounds on an increase in sample humidity. Transient O3 breakthrough also
was said to occur under humid conditions (Hudgens et al., 1994). The combined effects of
adsorbed material and sample humidity may contribute to the anomalous behavior reported for
a few scrubbers from field instruments, but that behavior could not be reproduced with new
scrubbers, even after continuous sampling of a smog chamber mixture for up to 13 weeks
(Hudgens et al., 1994). The aromatic compounds present at highest concentrations in ambient
air (e.g., benzene, toluene, xylenes, benzaldehyde) are relatively weak UV absorbers and are
not efficiently removed by the O3 scrubber. As a result, those compounds are not significant
interferences in the UV method (Hudgens et al., 1994). However, less common aromatics
(e.g., styrene, nitrotoluene) were found to absorb 254-nm light as effectively as does O3 and to
be efficiently adsorbed in the O3 scrubber (Hudgens et al., 1994). The importance of such
compounds as interferents in the UV method will depend on their ambient concentrations.
Interferences of the magnitude suggested by Leston and Ollison (1993) clearly
would have serious implications for monitoring of ambient O3. It is difficult to estimate
whether interferences in the UV method could be as high as suggested, in part, because data
are lacking on the ambient levels of potential interferents. Many potential interferents are
photochemically reactive, and it is questionable whether such compounds could co-exist with
ozone in sufficient quantities to constitute a significant interference. The results of Hudgens et
al. (1994) also suggest that periodic replacement of the O3 scrubber may minimize any
interferences in the UV method. In any case, full evaluation of interferences in UV and
ethylene chemiluminescence methods may require simultaneous measurements of O3, humidity,
temperature, and speciated organic compounds, and perhaps of other meteorological
parameters and potential interferents.
Spectroscopic Methods for Ozone
Spectroscopic methods have the potential to provide direct, sensitive, and specific
measurements representative of broad areas, rather than of single monitoring sites. This
potential has led to investigation of Spectroscopic approaches, primarily differential optical
absorption spectrometry (DOAS), for O3 measurement. Differential optical absorption
spectrometry measures the absorption through an atmospheric path (typically 0.5 to 1.5 km) of
two closely spaced wavelengths of light from an artificial source. One wavelength is chosen to
match an absorption line of the compound of interest, and the other is close to but off that line,
and is used to account for atmospheric effects. Platt and Perner (1980) reported measurements
of several atmospheric species, including O3, by DOAS, and various investigators have applied
the technique since then (Stevens et al., 1993, and references therein). Stevens et al. (1993)
described testing of a commercial DOAS instrument in North Carolina in the fall of 1989.
Ozone was measured using wavelengths between 260 and 290 nm, over a 557-m path. A
detection limit for O3 of 1.5 ppbv was reported, based on a 1-min averaging time (Stevens et
al., 1993). Comparison of DOAS results to those from a UV absorption instrument showed
(DOAS O3) = 0.90 x (UV O3) D 2.5 ppbv, with a correlation coefficient (r2) of 0.89, at ozone
levels up to 50 ppbv. The sensitivity, multiple analytical capability, stability, and speed of
response of the DOAS method are attractive, although further intercomparisons and
interference tests are recommended (Stevens et al., 1993).
3-99
-------
Personal and Passive Samplers for Ozone
A passive sampler is one that depends on diffusion of the analyte in air to a
collecting or indicating medium. In general, passive samplers are not adequate for
compliance-monitoring purposes because of limitations in specificity and averaging time.
However, passive sampling devices (PSDs) for O3 are of value as a means of obtaining
personal human exposure data for O3 and as a means of obtaining long-term O3 measurements
in areas where the use of instrumental methods is not feasible. Estimation of long-term
population exposure and ecological monitoring for vegetation effects of O3 in remote areas are
examples of the latter application. Passive sampling devices have the advantages of simplicity,
small size, and low cost, but also may present disadvantages, such as poor precision, loss of
effectiveness during use or storage, and interference from other atmospheric constituents.
New designs for PSDs have been implemented to overcome some of these limitations and to
make them more useful for short-term ambient and indoor studies, personal exposure
assessments, and validation of exposure models. Passive samplers for measuring O3 at ambient
concentrations are now commercially available.
The Ogawa PSD for O3 (Ogawa, Inc., Pompano Beach, FL) contains 0.1 mL of a
solution of NaNO2 and NaCO3 in glycerine on glass fiber filter paper. The nitrite ion reacts
with O3 to form nitrate. Following exposure, the PSDs are analyzed by extraction of the
nitrate with deionized water, followed by ion chromatographic (1C) analysis. In a comparative
ambient O3 study over 24 weeks, this PSD demonstrated agreement within about 10% with the
weekly real-time measurements taken by a UV O3 monitor (Mulik et al., 1991). Extension of
these measurements to a full year produced similar results (Mulik et al., 1991). The standard
deviation of weekly average measurements by three collocated PSD samplers ranged from
about +1 to +6 ppb, at weekly average O3 levels of 12 to 45 ppb (Mulik et al. 1991). The
Ogawa PSD also was used in a study of personal exposure to indoor and outdoor O3, showing
a correlation of r = 0.91, and relative errors of 15% (daytime) and 25% (nighttime) relative to
UV photometric data (Liu et al., 1992).
Another PSD for O3 has been developed that is based on the use of a colorant that
fades when exposed to O3 (Grosjean and Hisham, 1992; Grosjean and Williams, 1992). The
plastic, badge-type PSD contains a diffusion barrier and a colorant-coated filter as the O3 trap.
The colorant used is indigo carmine (5,5'-disulfonate sodium salt of indigo, D max = 608 nm).
With a plastic grid or Teflon® filter as the diffusion barrier, detection limits of 30 ppb D day
and 120 ppb D day, respectively, are achieved. Interferences from NO2, HCHO, and PAN are
15, 4, and 16%, respectively, of the ambient interferent concentrations. For sampling ambient
O3 in most locations, these interferences are probably negligible (Grosjean and Hisham, 1992).
Following sampling, the color change is measured by reflectance spectroscopy and no chemical
analysis is required. The reported shelf life is 3 mo prior to O3 exposure and 12 mo after O3
exposure (Grosjean and Hisham, 1992).
Field tests of the indigo carmine PSD were conducted at five forest locations in
California in the summer of 1990 (Grosjean and Williams, 1992). During these tests, ambient
O3 ranged up to 250 ppbv; 3-day average O3 values at the sites ranged from 40 to 88 ppbv.
The precision of the measurements was +12% based on 42 sets of collocated samplers, over
sampling durations of 3 to 30 days. The color change in the PSD was highly correlated (r =
0.99) with O3 dose as measured by UV photometry. No effect of ambient temperature or
3-100
-------
humidity variations was observed, and the total interference caused by other pollutants (NO2,
PAN, aldehydes) was less than 5%.
A third PSD for O3 also has been developed recently; it is based on color formation
from the reaction of O3 with an aromatic amine (Kirollos and Attar, 1991). The
ChromoSense™ direct-read passive dosimeter is a credit-card-sized device that changes color
proportionally to the integrated dose of exposure of the specific toxic material for which it was
designed (U.S. Patent 4,772,560). The dosimeter consists of an outer polyester pouch that
encloses a polymeric plate with a sorbent and membrane. A filtering layer is coated on the
membrane to reduce the sensitivity of the detection process to NO2. The chromophoric layer,
consisting of an aromatic amine that can react with O3 and form color, is encapsulated so as to
create a very high surface area. A polymeric barrier separates the chromophore from a UV-
absorbing layer to reduce their interaction. The UV absorber (in a polymeric matrix) helps
stabilize the chromophore toward intense light exposure when the device is used outdoors.
The transparent polymeric plate keeps the wafer flat and allows uninterrupted optical viewing
of the color of the reference and the sample area. An electronic reading device measures color
on both the exposed (sample) and unexposed (reference) areas, and displays a digital reading
that is proportional to the log of the O3 dose. Visible color is formed at doses as low as 20 ppb
Dh. No interference from NO2 is observed at NO2 concentrations up to 350 ppb, and only a
small effect of ambient humidity has been reported (Kirollos and Attar, 1991). No data on
precision have yet been reported.
A popularization of a PSD for O3 has been achieved in the form of the EcoBadge®,
which employs color formation by reaction of O3 with an undisclosed reagent in a filter paper
(Vistanomics, Inc., Glendale, CA). The EcoBadge is available through several scientific
equipment catalogs, primarily as a tool for classroom instruction on environmental issues. The
badge is said to indicate both a 1-h and an 8-h average O3 concentration. Comparison of color
development to a standard chart indicates O3 concentrations up to about 350 ppbv, with a limit
of resolution of about 20 ppbv or more. The badge is stated to be unaffected by air velocity,
humidity, or temperature, with only a slight interference from NO2. Test data apparently have
not been published. The EcoBadge has been used in middle and high school programs
promoting science and math and is included in the curriculum of the Global Thinking Project
(1993), an international telecommunications and education network.
Calibration Methods for Ozone
Because it is an unstable molecule and cannot be stored, O3 must be generated at the
time of use to produce calibration mixtures. Electrical discharges in air or oxygen readily
produce O3, but at concentrations far too high for calibration of ambient monitors.
Radiochemical methods are expensive and require the use of radioactive sources, with
associated safety requirements. For calibration purposes, low levels of O3 nearly always are
generated by photolysis of oxygen at wavelengths < 200 nm. Placing a mercury lamp near a
quartz tube through which air is flowing produces small amounts of O3 in the airstream.
Commercial O3 sources based on this approach typically adjust the lamp current to control the
amount of light transmitted, and thus the O3 produced.
Once a stable, low concentration of O3 has been produced from a photolytic
generator, that O3 output must be established by measurement with an absolute reference
method. The original reference calibration procedure promulgated by EPA in 1971 (Federal
Register, 1971) was an iodometric procedure, employing 1% aqueous neutral buffered
3-101
-------
potassium iodide (NBKI). A large number of studies conducted in the 1970s revealed several
deficiencies with potassium iodide (KI) methods, the most notable of which were poor
precision or interlaboratory comparability and a positive bias of NBKI measurements relative
to simultaneous absolute UV absorption measurements.
Following investigations of problems with the NBKI method, EPA evaluated four
potential reference calibration procedures and selected UV photometry on the basis of superior
accuracy and precision and simplicity of use (Rehme et al., 1981). In 1979, UV photometry
was designated the reference calibration procedure by EPA (Federal Register, 1979a).
The measurement principle of UV O3 photometers used as reference standards is
identical to that of O3 photometers used for ambient measurements (see Section 3.5.1.1).
A laboratory photometer used as a reference standard will typically contain a long-path cell (1
to 5 m) and employ sophisticated digital techniques for making effective double-beam
measurements of small absorbances at low O3 concentrations.
A primary reference standard is a UV photometer that meets the requirements set
forth in the 1979 revision designating UV photometry as the reference method (Federal
Register, 1979a). Commercially available O3 photometers that meet those requirements may
function as primary standards. The EPA and the National Institute of Standards and
Technology (NIST, formerly National Bureau of Standards [NBS]) have established a
nationwide network of standard reference photometers (SRPs) that are used to verify local
primary standards and transfer standards. A secondary or transfer standard is a device or
method that can be calibrated against the primary standard and then moved to another location
for calibration of O3 monitors. Commercial UV photometers for O3 often are used as
secondary or transfer standards, as are commercial photolytic ozone generators and apparatus
for the gas-phase titration of O3 with NO.
The latter method, gas-phase titration (GPT) of O3 with NO (NO + O3 D
NO2 + O2), is a direct and absolute means of determining O3, provided NO is in excess so that
no side reactions occur. Under such conditions, GPT has the advantage that measurement of
the NO or O3 consumed or the NO2 produced gives a simultaneous measurement of the other
two species. All three modes have been used, and this method is often used for calibration of
NO/NOX analyzers. Gas-phase titration has been compared to UV photometry in several
studies. The most detailed study is that of Fried and Hodgeson (1982), who used an NBS
primary standard UV photometer, highly accurate flow measurements, photoacoustic detection
of NO2, and NBS (now NIST) Standard Reference Materials as sources of NO and NO2. That
study showed that decreases in O3 as measured by the UV method averaged 3.6% lower than
the corresponding decrease in NO and increase in NO2 measured independently. Because of
uncertainty about the origin of the small bias relative to UV photometry, GPT is used as a
transfer standard but not as a primary reference standard.
3.5.1.2 Peroxyacetyl Nitrate and Its Homologues
During laboratory organic photooxidation studies, Stephens et al. (1956a,b)
determined the presence of a number of alkyl nitrates and an unidentified species called
"Compound X". The presence of Compound X in the atmosphere of Los Angeles was
confirmed by Scott et al. (1957). In later work (Stephens et al., 1961), its structure was
determined and Compound X was named peroxyacetyl nitrate, or PAN. Since the discovery of
PAN, much effort has been directed toward its atmospheric measurement. In the following
subsections PAN measurement and calibration techniques are described. The discussion on
3-102
-------
measurement techniques includes a summary description and identifies limits of detection,
specificity (interferences), reproducibility, and accuracy of each method. The relative merits
of each method also are presented. The subsection on calibration techniques includes those
methods most often employed during ambient air measurement studies.
Measurement Methods
Two methods generally have been employed to make atmospheric measurements of
PAN. These methods are infrared spectroscopy (IR) and GC. Infrared spectroscopy permits
the sampling and analysis to be conducted in real time. Because PAN is very reactive in the
gas phase and exhibits surface adsorptive effects, the minimal contact time offered by IR makes
this method very attractive. However, IR instrumentation is expensive and complex and
requires a good deal of space. On the other hand, GC is inexpensive and requires minimal
space and operator training. A GC can be set up to automatically sample and analyze air for
PAN and PANs. Application of these methods for obtaining ambient concentrations of PAN
and other organic nitrates recently has been reviewed by Roberts (1990).
Infrared Spectroscopy. Conventional long-path infrared spectroscopy and FTIR
have been used to detect and measure atmospheric PAN. Sensitivity is enhanced by the use of
FTIR. The most frequently used IR bands have been assigned, and the absorptivities reported
in the literature (Stephens, 1964; Bruckmann and Willner, 1983; Holdren and Spicer, 1984;
Niki et al., 1985; Tsalkani and Toupance, 1989) permit the quantitative analysis of PAN
without calibration standards. Tuazon et al. (1978) have described an FTIR system operable at
pathlengths up to 2 km for ambient measurements of PAN and other trace constituents. This
system employed an eight-mirror multiple reflection cell with a 22.5-m base path. The spectral
windows available at pathlengths of 1 km were 760 to 1,300, 2,000 to 2,230 cm'1. Thus, PAN
could be detected by the bands at 793 and 1,162 cm"1. The 793-cnr1 band is characteristic of
peroxynitrates, whereas the 1,162-cnr1 band is reportedly caused by PAN only (Stephens,
1969; Hanst et al., 1982). Tuazon et al. (1981a,b) reported ambient measurements with this
system during a smog episode in Claremont, CA, in 1978. Maximum daily PAN
concentrations ranged from 6 to 37 ppb over a 5-day episode. A detection limit for PAN was
3 ppb at a pathlength of Dl km. Hanst et al. (1982) modified the FTIR system used by Tuazon
by changing it from an eight-mirror to a three-mirror cell configuration and by considerably
reducing the cell volume. A detection limit for PAN was increased to 1 ppb at a similar
pathlength.
The limited sensitivity (Dl ppb) and the complexity of the above FTIR systems
generally have limited their field use to urban areas such as Los Angeles. More recently,
cryogenic sampling and matrix-isolation FTIR have been used to measure PAN in 15-L
integrated samples of ambient air. The matrix isolation technique has a theoretical level of
detection of D50 ppt (Griffith and Schuster, 1987).
Gas Chromatography-Electron Capture Detection. Peroxyacetyl nitrate is
normally measured by using a GC coupled to an electron capture detector (GC/ECD). The
method was originally described by Darley et al. (1963) and subsequently has been refined and
employed by scientists over the years. Key features of the method remain unchanged. The
column and detector temperatures are kept relatively low (D50 and 100 DC, respectively) to
minimize PAN thermal decomposition. Short columns of either glass or Teflon® generally are
used (1 to 5 ft in length). Finally, column packing normally includes a Carbowax stationary
phase coated onto a deactivated solid support. Using packed columns, detection limits of
3-103
-------
10 ppt have been reported using direct sampling with a 20-mL sample loop (Vierkorn-Rudolph
et al., 1985). Detection limits were further extended to 1 to 5 ppt using cryogenic enrichment
of samples (Vierkorn-Rudolph et al., 1985; Singh and Salas, 1983). These studies have found
only slight overall losses of PAN (10 to 20%) associated with cryosampling, provided samples
are warmed only to room temperature during desorption.
Recently, improved precision and sensitivity have been reported using fused-silica
capillary columns instead of packed columns (Helmig et al., 1989; Roberts et al., 1989).
Signal-to-noise enhancement of 20 has been claimed (Roberts et al., 1989).
Gas Chromatography-Alternate Detection. As noted earlier, PAN is readily
reduced to NO in the gas phase. To separate PAN, NO, and NO2, Meyrahn et al. (1987)
coupled a GC with a molybdenum converter and used a chemiluminescent analyzer to measure
PAN as NO. Using a 10-mL sample loop, a detection limit of 10 ppb was reported.
A luminol-based detector also has shown sensitivity to PAN. Burkhardt et al.
(1988) used GC and a commercially available luminol-based instrument (i.e., Scintrex LMA-3
Luminox) to detect both NO2 and PAN. Using a sampling interval of 40 s, linear response was
claimed from 0.2 to 170 ppb NO2 and from 1 to 65 ppb PAN. Although the PAN calibration
was nonlinear below 1 ppb, a detection level of 0.12 ppb was reported. Drummond et al.
(1989) slightly modified the above approach by converting the PAN from the GC column to
NO2 and measuring the resulting NO2 with a luminol-based instrument.
Peroxyacetyl Nitrate Stability
Peroxyacetyl nitrate is an unstable gas and is subject to surface-related
decomposition as well as thermal instability. Peroxyacetyl nitrate exists in a temperature-
sensitive equilibrium with the peroxyacetyl radical and NO2 (Cox and Roffey, 1977).
Increased temperature favors the peroxyacetyl radical and NO2 at the expense of PAN. Added
NO2 should force the equilibrium toward PAN and enhance its stability. In the presence of
NO, peroxyacetyl radicals react rapidly to form NO2 and acetoxy radicals, which decompose in
O2 to radicals that also convert NO to NO2. As a result, the presence of NO acts to reduce
PAN stability and enhance its decay rate (Lonneman et al., 1982). Stephens (1969) reported
that appreciable PAN loss in a metal sampling valve was traced to decomposition on a silver-
soldered joint. Meyrahn et al. (1987) reported that PAN decayed according to first-order
kinetics at a rate of 2 to 4%/h in glass vessels, and they suggested first-order decay as the basis
for a proposed method of in-field PAN calibration. In contrast, Holdren and Spicer (1984)
found that without NO2 added, 20 ppb PAN decayed in Tedlar bags according to first-order
kinetics at a rate of 40%/h. The addition of 100 ppb NO2 acted to stabilize the PAN (20 ppb)
in the Tedlar bags.
A humidity-related difference in GC/ECD response has been reported (Holdren and
Rasmussen, 1976). Low responses observed at humidities below 30% and PAN concentrations
of 10 and 100 ppb, but not 1,000 ppb, were attributed to sample-column interactions. A
humidity effect was alluded to by Nieboer and Van Ham (1976) but details were not given. No
humidity effect was observed by Lonneman (1977). Watanabe and Stephens (1978) conducted
experiments at 140 ppb and did not conclude that the reduced response was from faults in the
detector or the instrument. They concluded that there was no column-related effect, and they
observed surface-related sorption by PAN at 140 ppb in dry acid-washed glass flasks. They
recommended that moist air be used to prepare PAN calibration mixtures to avoid potential
surface-mediated effects.
3-104
-------
Another surface-related effect has been reported for PAN analyses of remote marine
air (Singh and Viezee, 1988). Peroxyacetyl nitrate concentrations were found to increase by
20 to 170 ppt, an average factor of 3, when the sample was stored in a glass vessel for 1 to
2 min prior to analysis. This effect remains to be explained.
Preparation and Calibration
Because PAN is unstable, the preparation of reliable calibration standards is
difficult. The more promising methods are described here. The original method used the
photolysis of ethyl nitrite in pure oxygen (Stephens, 1969). When pure PAN is desired, the
reaction mixture must be purified, usually by chromatography, to remove the major
by-products, acetaldehyde and methyl and ethyl nitrates (Stephens et al., 1965). For GC
calibration, purification is unnecessary; the PAN concentration in the reactant matrix is
established from the IR absorption spectrum and subsequently diluted to the parts-per-billion
working range needed for calibration purposes (Stephens and Price, 1973).
Static mixtures of molecular chlorine, acetaldehyde, and NO2 in the ratio of 2:4:4
can be photolyzed in the presence of a slight excess NO2 to give a near-stoichiometric yield of
PAN (Gay et al., 1976). This method was adapted by Singh and Salas (1983) and later by
Grosjean et al. (1984), using photolytic reactors to provide continuous PAN calibration units at
concentrations between 2 and 400 ppb. In the former approach, the PAN concentration is
established by measuring the change in acetaldehyde concentration across the reactor. In the
latter approach, the PAN concentration is established by measuring the acetate in an alkaline
bubbler where PAN is hydrolyzed.
A static technique involving the photolysis of acetone in the presence of NO2 and air
at 250 nm has been reported to produce a constant concentration of PAN (Meyrahn et al.,
1987; Warneck and Zerbach, 1992). A Penray mercury lamp is inserted into a mixture of
10 ppm NO2 and 1 % acetone and irradiated for 3 min to yield 8.9 + 0.3 ppm PAN.
Peroxyacetyl nitrate can be synthesized in the condensed phase by the nitration of
peracetic acid in C6H14 (Helmig et al., 1989), heptane (Nielsen et al., 1982), C8H18 (Holdren
and Spicer, 1984), or n-tridecane (Gaffney et al., 1984). Purification of PAN in the liquid
phase is needed using the first two methods. The resulting PAN-organic solution can be stored
at D20 to D80 DC with losses of less than 3.6%/mo and can be injected directly into a vessel
containing air to produce a calibration mixture. The PAN concentration is normally
established by FTIR analysis of the solution or the resulting PAN-air mixture.
Peroxyacetyl nitrate readily disassociates to NO, and chemiluminescence NOX
analyzers have near-quantitative response to PAN. Thus, under some circumstances,
chemiluminescent NOX response can be used for PAN calibration. One method uses the
difference in NOX signal measured upstream and downstream of an alkaline bubbler (Grosjean
and Harrison, 1985a). Joos et al. (1986) coupled a chemiluminescence NOX analyzer with a
GC system to permit calibration of the BCD response by reference to the chemiluminescence
NOX analyzer that has been calibrated by traditional methods.
As noted previously, NO in the presence of PAN is converted to NO2.
Approximately four molecules of NO can react per molecule of PAN. Lonneman et al. (1982)
devised a PAN calibration procedure based on the reaction of PAN with NO in the presence of
benzaldehyde, which is added to control unwanted radical chemistry and to improve precision.
Using this approach and an initial NO-to-PAN ratio between 10 and 20 to 1, the change in NO
concentration is monitored with a chemiluminescence NO analyzer, the change in PAN GC-
3-105
-------
BCD response is monitored, and the resulting ratio (i.e., DNO/DPAN) is divided by the
stoichiometric factor of 4.7 to arrive at a calibration factor for the BCD.
Peroxyacetyl nitrate and n-propyl nitrate (NPN) have similar BCD responses.
Serial dilution of the more stable compound, NPN, has been used for field operations
(Vierkorn-Rudolph et al., 1985). This approach is not recommended for primary calibration,
however, because it does not permit verification of quantitative delivery of PAN to the detector
(Stephens and Price, 1973).
3.5.1.3 Gaseous Hydrogen Peroxide
Although O3 has long been considered to be the primary oxidant affecting air
quality, atmospheric chemists recently have identified H2O2, a photochemical reaction product,
as another oxidant that also may play a significant role in diminishing air quality. In order to
assess the role of atmospheric H2O2, good measurement methods are needed. Early
measurements in the 1970s reported H2O2 concentrations ranging from 10 to 180 ppb (Gay and
Bufalini, 1972 a,b; Kok et al., 1978 a,b). However, these measurements are in error because
of artifact formation of H2O2 from reactions of absorbed gaseous O3 (Zika and Saltzman, 1982;
Heikes et al., 1982; Heikes, 1984). Modeling results also indicate that H2O2 atmospheric
concentrations should be on the order of 1 ppb (Chameides and Tan, 1981; Logan et al.,
1981).
In the following section, the discussion focuses on those sampling and analytical
methods most frequently used within the last decade to determine atmospheric levels of H2O2.
The measurement techniques are described and limits of detection, specificity (interferences),
reproducibility, and accuracy are discussed.
Measurement Methods
In situ measurement methods that have been employed for determining gaseous
H2O2 include both FTIR and tunable diode laser absorption spectrometry (TDLAS). Four
methods involving sample collection via wet chemical means and subsequent analysis via
chemiluminescent or fluorescent detection also have been used frequently: (1) luminol,
(2) peroxyoxalate, (3) enzyme-catalyzed (peroxidase), and (4) benzoic acid-fenton reagent
methods. Application of most of these methods for obtaining ambient concentrations of H2O2
recently has been reviewed by Sakugawa et al. (1990) and Gunz and Hoffmann (1990).
In Situ Methods. Fourier transform infrared spectroscopy was employed in the
early 1980s for atmospheric measurements (Tuazon et al., 1980; Hanst et al., 1982). Even
though the FTIR is very specific for H2O2, it saw limited use because of the high detection
level of D50 ppb when using a 1-km path length. The TDLAS also has very high specificity
for H2O2 and was subsequently evaluated and shown to have a much improved detection limit
of 0.1 ppb when using scan-averaging times of several minutes (Slemr et al., 1986; MacKay
and Schiff, 1987a; Schiff et al., 1987).
Wet Chemical Methods. Numerous wet chemical techniques for measuring H2O2
have been reported. However, discussion in this section is limited to the four approaches most
frequently used by researchers.
Luminol Method. Hydrogen peroxide concentrations in the atmosphere have been
determined by the chemiluminescent response obtained from the catalyzed oxidation of luminol
(5-amino-2,3-dihydro-l,4-phthalazinedione) by H2O2. Copper2+ (Armstrong and Humphreys,
1965; Kok et al., 1978 a,b; Das et al., 1982) and hemin, a blood component (Yoshizumi et al.,
3-106
-------
1984), have been reported as catalysts for the luminol-based H2O2 oxidation. Method
sensitivity of DO.01 ppb has been achieved. Interference from O3, SO2, metal ions, and high
pH have been reported along with ways to mitigate these effects (Heikes et al., 1982; Zika and
Saltzman, 1982; Ibusuki, 1983; Lazrus et al., 1985; Aoyanagi and Mitsushima, 1985; Hoshino
and Hinze, 1987).
Peroxyoxalate Method. The peroxyoxalate chemiluminescence method also has
been employed by a number of researchers (Rauhut et al., 1967; Scott et al., 1980; Klockow
and Jacob, 1986). Hydrogen peroxide reacts with fe(2,4,5-trichloro-6-phenyl)-oxalate to form
a high-energy dioxetanedione (Stauff and Jaeschke, 1972). The chemiluminescence is
transmitted to the fluorophore, perylene, which emits light on return to the ground state.
Method sensitivity of DO.01 ppb is achieved, and no interferences are observed from O3 and
metal ions. A signal depression has been reported for trace levels of nitrite (> 10~5 M), sulfite
(> 10'4 M), and formaldehyde (> 10'3 M) (Klockow and Jackob, 1986).
Enzyme-Catalyzed Method (Peroxidase). This general method involves three
components: (1) a substrate that is oxidizable; (2) the enzyme, horseradish peroxidase (HRP);
and (3) H2O2. The production or decay of the fluorescence intensity of the substrate or
reaction product is measured as it is oxidized by H2O2, catalyzed by HRP. Some of the more
widely used chromogenic substrates have been scopoletin (6-methoxy-7-hydroxy-l,2-
benzopyrone) (Andreae, 1955; Perschke and Broda, 1961), 3-(p-hydroxphenyl)propionic acid
(HPPA) (Zaitsu and Okhura, 1980), leuco crystal violet (LCV) (Mottola et al., 1970), and
p-hydroxyphenylacetic acid (POPHA) (Guilbault et al., 1968).
Of the chromogens used, POPHA is one of the better indicating substrates.
Hydrogen peroxide oxidizes the peroxidase and is itself reduced by electron transfer from
POPHA. The POPHA radicals form a dimer that is highly fluorescent. Because the chemical
reaction is sensitive to both H2O2 and organic peroxides, a dual-channel system with a H2O2
removal step (use of catalase) is used to distinguish H2O2 from organic peroxides (Lazrus et
al., 1985; Wei and Weihan, 1987; Dasgupta and Hwang, 1985; Kok et al., 1986).
The peroxidase-POPHA-fluorescence technique has been used by several groups to
measure gas-phase H2O2 concentrations (Lazrus et al., 1986; Tanner et al., 1986; Heikes et al.,
1987; Van Valin et al., 1987; Dasgupta et al., 1988; Olszyna et al., 1988; Meagher et al.,
1990). Method detection levels range from 0.01 to 0.1 ppb. However, artifact formation does
occur as a result of the reaction of dissolved O3 in the collection devices (Staehelin and
Hoigne, 1982; Heikes, 1984; Gay et al., 1988). To overcome the O3 interference, researchers
have used NO to eliminate O3 (Tanner, 1985; Tanner et al., 1986; Shen et al., 1988).
Fenton Reagent-Isomeric Hydroxybenzoic Acids Method. This technique involves
the formation of aqueous OH radicals from the reaction of Fenton reagent (Fe2+) complex)
with gaseous H2O2. The OH radicals, in turn, react with benzoic acid (hydroxyl radical
scavenger) to form isomeric hydroxybenzoic acids (OHBA). The OHBA fluoresces weakly at
the pH necessary to carry out the above reactions. Fluorescence is enhanced by adding NaOH
to the product stream (Lee et al., 1990) or by using a low pH A13+ fluorescence enhancing
reagent (Lee et al., 1993).
Comparison of Methods
The above techniques have been shown to measure H2O2 in the atmosphere with
detection levels of DO. 1 ppb. Kleindienst et al. (1988) compared several of these techniques
using three sources of H2O2: (1) zero air in the presence and absence of common interferences,
3-107
-------
(2) steady-state irradiations of hydrocarbon-NOx mixtures, and (3) ambient air. The
measurements were conducted simultaneously from a common manifold. For pure samples in
zero air, agreement within 23% was achieved among methods over a concentration range of
0.06 to 128 ppb. A negative SO2 interference was caused with the luminol technique. During
the irradiation experiment, significant concentrations of organic peroxides were generated, and
the agreement among techniques for H2O2 was very poor. For ambient measurements, the
methods agreed reasonably well with an average deviation of 30% from the mean values.
Atmospheric intercomparison studies also have been conducted as part of the
Carbon Species Methods Comparison Study (Calif, 1986). The results of the study indicated
that the wet chemical methods still suffer from sampling artifacts and interferences from other
atmospheric constituents (Dasgupta et al., 1990; MacKay et al., 1990; Kok et al., 1990;
Sakugawa et al., 1990; Tanner and Shen, 1990). Lee et al. (1991) showed that substantial loss
of airborne H2O2 can occur when air is drawn through Teflon® tubing of inlet sampling
devices. In addition to reducing the H2O2 in incoming ambient air, this line loss also
compromises the use of aqueous standards to calibrate a gas-phase monitoring system. More
recently, Lee et al. (1994) have demonstrated a surfaceless inlet system to eliminate line loss
problems. It is clear from the above studies that further comparisons of techniques are needed
to resolve questions of errors and to provide improved measurement techniques.
Calibration Methods
The most frequently used method for generating aqueous standards is simply the
serial dilution of commercial grade 30% H2O2/water. The dilute solutions of H2O2 as low as
10"4 have been found to be stable for several weeks if kept in the dark (Armstrong and
Humphreys, 1965). The stock H2O2 solution is standardized by iodometry (Allen et al., 1952;
Hochanadel, 1952; Cohen et al., 1967) or, more recently, by using a standardized
permanganate solution (Lee et al., 1991).
Gaseous H2O2 standards are not as easily prepared, and stability problems require
the use of standard mixtures immediately. One method makes use of the injection of microliter
quantities of 30% H2O2 solution into a metered stream of air that flows into a Teflon® bag.
The amount of H2O2 in the gas phase is determined by the iodometric titration method (Cohen
and Purcell, 1967). Gas-phase H2O2 standards also have been generated by equilibrating N2
with an aqueous H2O2 solution of known concentration that is maintained at constant
temperature. Equilibrium vapor pressures and corresponding gas-phase concentrations are
calculated using Henry's law constant (Lee et al., 1991).
3.5.2 Sampling and Analysis of Volatile Organic Compounds
3.5.2.1 Introduction
The term volatile organic compounds generally refers to gaseous organic
compounds that have a vapor pressure greater than 0.15 mm and, generally, have a carbon
content ranging from C, through C12. As discussed in Sections 3.2 and 3.4, VOCs are emitted
from a variety of sources and play a critical role in the photochemical formation of O3 in the
atmosphere.
The U.S. Environmental Protection Agency revised the ambient air quality
surveillance regulations in Title 40, Part 58, of the Code of Federal Regulations to include,
among other activities, the monitoring of VOCs. The revisions require states to establish VOC
air monitoring stations in nonattainment areas as part of their existing State Implementation
3-108
-------
Plan (SIP) monitoring networks. Authority for requiring the enhanced monitoring is provided
for in Title I, Section 182, of the CAAA of 1990 (U.S. Congress, 1990). Several states have
begun acquiring VOC data on 55 O3 precursors at these PAMS, using methodology discussed
in an EPA technical assistance document (U.S. Environmental Protection Agency, 1991c).
The term nonmethane organic compounds also is used frequently and refers to a
subset of VOCs, because it excludes the compound CH4. Numerous sampling, analytical, and
calibration methods have been employed to determine NMOCs in ambient air. Some of the
analytical methods utilize detection techniques that are highly selective and sensitive to specific
functional groups or atoms of a compound (e.g., formyl group of aldehydes, halogen), whereas
others respond in a more universal manner (i.e., to the number of carbon atoms present in the
organic molecule). In this overview of the most pertinent measurement methods, NMOCs
have been arranged into three major classifications: (1) NMHCs, (2) carbonyl species, and (3)
polar volatile organic compounds (PVOCs). Measurement and calibration procedures are
discussed for each classification.
3.5.2.2 Nonmethane Hydrocarbons
Nonmethane hydrocarbons constitute the major portion of NMOC in ambient air.
Traditionally, NMHCs have been measured by methods that employ a flame ionization detector
(FID) as the sensing element. This detector was originally developed for GC and employs a
sensitive electrometer that measures a change in ion intensity resulting from the combustion of
air containing organic compounds. Ion formation is essentially proportional to the number of
carbon atoms present in the organic molecule (Sevcik, 1975). Thus, aliphatic, aromatic,
alkenic, and acetylenic compounds all respond similarly to give relative responses of 1.00 +
0.10 for each carbon atom present in the molecule (e.g., 1 ppm hexane = 6 ppmC; 1 ppm
benzene = 6 ppmC; 1 ppm propane = 3 ppmC). Carbon atoms bound to oxygen, nitrogen, or
halogens give reduced relative responses (Dietz, 1967). Consequently, the FID, which is
primarily used as a hydrocarbon measuring method, more correctly should be viewed as an
organic carbon analyzer.
In the following sections, discussion focuses on the various methods utilizing this
detector to measure total nonmethane organics. Methods in which no compound speciation is
obtained are covered first. Methods for determining individual organic compounds then are
discussed.
Nonspeciation Measurement Methods
The original EPA reference method for NMOC, which was promulgated in 1971,
involves GC separation of CH4 from the remaining organics in an air sample (Federal Register,
1971). A second sample is injected directly to the FID without CH4 separation. Subtraction of
the first value from the second produces a nonmethane organic concentration.
A number of studies of commercial analyzers employing the Federal Reference
Method have been reported (Reckner, 1974; McElroy and Thompson, 1975; Harrison et al.,
1977; Sexton et al., 1982). These studies indicated overall poor performance of the
commercial instruments when either calibration or ambient mixtures containing NMOC
concentrations < 1 ppmC were used. The major problems associated with using these NMOC
instruments have been reported in an EPA technical assistance document (Sexton et al., 1981).
The technical assistance document also suggests ways to reduce the effects of existing
problems. Other nonspeciation approaches to the measurement of nonmethane organics also
3-109
-------
have been investigated. These approaches have been discussed in the 1986 EPA air quality
criteria document (U.S. Environmental Protection Agency, 1986a). Again, these approaches
also are subject to the same shortcomings as the EPA reference method (i.e., poor performance
below 1 ppmC of NMHC).
More recently, a method has been developed for measuring NMOC directly and
involves the cryogenic preconcentration of nonmethane organic compounds and the
measurement of the revolatilized NMOCs using FID (Cox et al., 1982; Jayanty et al., 1982).
This methodology has been formalized and is referred to as Method TO-12 and is published in
a compendium of methods for air toxics (Winberry et al., 1988). The EPA recommends this
methodology for measuring total NMOC and has incorporated it into the Technical Assistance
Document for Sampling and Analysis of Ozone Precursors (U.S. Environmental Protection
Agency, 1991c).
A brief summary of the method is as follows. A whole air sample is drawn through
a glass bead trap that is cooled to approximately D185 DC using liquid argon. The cryogenic
trap collects and concentrates the NMOC, while allowing the CH4, nitrogen, oxygen, etc., to
pass through the trap without retention. After a known volume of air has been drawn through
the trap, carrier gas is diverted to the trap first to remove residual air and CH4. When the
residual gases have been flushed from the trap, the cryogen is removed and the temperature of
the trap is ramped to approximately 100 DC. The revolatilized compounds pass directly to a
FID (no analytical column). The corresponding signal is integrated over time (several minutes)
to obtain a total FID response from the NMOC species. Water vapor, which also is
preconcentrated, causes a positive shift in the FID signal. The effect of this shift is minimized
by optimizing the peak integration parameters.
The sensitivity and precision of Method TO-12 are proportional to the sample
volume. However, ice formation in the trap limits sampling volumes to D500 cc. The
detection level is 0.02 ppmC (with a signal-to-noise ratio [S/N] of 3), and the precision at
1 ppmC and above has been determined to be D5 %. The instrument response has been shown
to be linear over a range of 0 to 10 ppmC. Propane gas certified by NIST is normally used as
the calibrant. Accuracy at the method quantitation level (S/N = 10) is +20%.
Speciation Measurement Methods
The primary measurement technique utilized for NMOC speciation is GC. Coupled
with FID, this analytical method permits the separation and identification of many of the
organic species present in ambient air.
Separation of compounds is accomplished by means of both packed and capillary
GC columns. If high resolution is not required and large sample volumes are to be injected,
packed columns are employed. The traditional packed column may contain either a solid
polymeric adsorbent (gas-solid chromatography) or an inert support, coated with a liquid (gas-
liquid chromatography). Packed columns containing an adsorbent substrate normally are
required to separate C2 and C3 compounds. The second type of column can be a support-
coated or wall-coated open tubular capillary column. The latter column has been used widely
for environmental analysis because of its superior resolution and broader applicability. The
wall-coated capillary column consists of a liquid stationary phase coated or bonded (cross-
linked) to the specially treated glass or fused-silica tubing. Fused-silica tubing is most
commonly used because of its physical durability and flexibility. When a complex mixture is
introduced into a GC column, the carrier gas (mobile phase) moves the sample through the
3-110
-------
packed or coated capillary column (stationary phase). The chromatographic process occurs as
a result of repeated sorption-desorption of the sample components (solute) as they move along
the stationary phase. Separation occurs as a result of the different affinities that the solute
components have for the stationary phase.
As described in the previous O3 criteria document (U.S. Environmental Protection
Agency, 1986a), the GC-FID technique has been used by numerous researchers to obtain
ambient NMOC data. Singh (1980) drew on the cumulative experience of these researchers to
prepare a guidance document for state and local air pollution agencies interested in obtaining
speciation data. In general, most researchers have employed two gas chromatographic units to
carry out analyses of NMOC species in ambient air. The more volatile VOCs (C2 through C5)
generally are measured on one unit using packed-column technology, whereas the other GC
separates the less volatile organics using a capillary column. In typical chromatograms of
urban air, all major peaks are identified and, on a mass basis, represent from 65 to 90% of the
measurable nonmethane organic burden.
Identification of GC peaks is based on matching retention times of unknown
compounds with those of standard mixtures. Subsequent verification of the individual species
is normally accomplished with gas chromatographic-mass spectrometric (GC/MS) techniques.
Compound-specific detection systems, such as electron capture, flame photometry, and
spectroscopic techniques, also have been employed to confirm peak identifications. The peak
matching process is far from being a trivial task. Ambient air chromatograms are often very
complex (> 200 peaks/run) and require a good deal of manual labor to assure that the peak
matching process is being carried out correctly by the resident peak identification/
quantification software. Efforts to improve on the accuracy of peak assignment and diminish
the labor hours normally associated with the objective recently have been reported. Silvestre et
al. (1988) developed an off-line spreadsheet program that is menu-driven and used to identify
and edit a chromatogram containing 200 peaks within 15 min. The accuracy of peak
assignment was typically better than 95%. Mason et al. (1992) developed a novel algorithm
that is embedded within the Harwell MatchFinder software package and demonstrated the
potential of the algorithm for enhancing peak identification in complex chromatograms. The
authors indicate that the software could be used to batch-process large volumes of
chromatographic data. A commercial software package from Meta Four Software, Inc., was
recently employed during the Atlanta Ozone Precursor Monitoring Study to batch process
chromatographic data from over 6,000 GC runs (Purdue et al., 1992). This software was also
used to validate peak identities from two GC databases and was shown to improve peak
identities from the originally processed data by 10 to 20% (Holdren et al., 1993).
Because the organic components of the ambient atmosphere are present at parts-per-
billion levels or lower, sample preconcentration is necessary to provide sufficient material for
the GC-FID system. The two primary techniques utilized for this purpose are the use of solid
adsorbents and cryogenic collection. The more commonly used sorbent materials are divided
into three categories: (1) organic polymeric adsorbents, (2) inorganic adsorbents, and (3)
carbon adsorbents. Primary organic polymeric adsorbents used for NMOC analyses include
the materials Tenax®-GC and XAD-2®. These materials have a low retention of water vapor,
and, hence, large volumes of air can be collected. These materials do not, however, efficiently
capture highly volatile compounds such as C2 to C5 hydrocarbons, nor certain polar compounds
such as CH3OH and C3H6O. Primary inorganic adsorbents are silica gel, alumina, and
molecular sieves. These materials are more polar than the organic polymeric adsorbents and
3-111
-------
are thus more efficient for the collection of the more volatile and polar compounds.
Unfortunately, water also is collected efficiently, which in many instances leads to rapid
deactivation of the adsorbent. Carbon adsorbents are less polar than the inorganic adsorbents
and, as a result, water adsorption by carbon adsorbents is a less significant problem. The
carbon-based materials also tend to exhibit much stronger adsorption properties than organic
polymeric adsorbents; thus, lighter-molecular-weight species are more easily retained. These
same adsorption effects result, however, in irreversible adsorption of many compounds.
Furthermore, the very high thermal desorption temperatures required (350 to 400 DC) limit the
use of carbon adsorbents and also may lead to degradation of labile compounds. The
commonly available classes of carbon adsorbents include various conventional activated
carbons, carbon molecular sieves (Spherocarb®, Carbosphere®, Carbosieve®), and
carbonaceous polymeric adsorbents (Ambersorb® XE-340, XE-347, SE-348).
Although a number of researchers have employed solid adsorbents for the
characterization of selected organic species in air, only a few attempts have been made to
identify and quantitate the range of organic compounds from C2 and above. Westberg et al.
(1982) evaluated several carbon and organic polymeric adsorbents and found that Tenax®-GC
exhibited good collection and recovery efficiencies for DC6 organics; the remaining adsorbents
tested (XAD-4®, XE-340®) were found unacceptable for the lighter organic fraction. The
XAD-4® retained DC2 organic gases, but it was impossible to desorb these species completely
without partially decomposing the XAD-4®. Good collection and recovery efficiencies were
provided by XE-340® only for organics of C4 and above. Ogle et al. (1982) used a
combination of adsorbents in series and designed an automated GC-FID system for analyzing
C2 through C10 hydrocarbons. Tenax-GC® was utilized for C6 and above; whereas Carbosieve
S® trapped C3 through C5 organics. Silica gel followed these adsorbents and effectively
removed water vapor while passing the C2 hydrocarbons onto a molecular-sieve, 5A adsorbent.
More recently, Levaggi et al. (1992) used a combination of adsorbents in series for analyzing
C2 through C10 hydrocarbons. Tenax GR, Carbotrap, and Carbosieve S-III were evaluated. At
room temperature collection, excellent recovery efficiencies were obtained for all species
except acetylene (breakthrough begins after 220 cc). Smith et al. (1991b) evaluated a
commercially available GC system (Chrompack, Inc.) and found that a Carbotrap C,
Carbopack B, and Carbosieve S-III combination was effective for all C2 and above species, if
the trap temperature was maintained at D30 DC during collection (600 cc). The above
researchers also caution that artifact peaks do occur during thermal desorption and recommend
closely screening the resulting data.
The preferred method for obtaining NMOC data is cryogenic preconcentration
(Singh, 1980). Sample preconcentration is accomplished by directing air through a packed trap
immersed in either liquid oxygen (B.P. D183 DC) or liquid argon (B.P. D186 DC). For the
detection of about 1 ppbC of an individual compound, a 250-cc air sample normally is
processed. The collection trap generally is filled with deactivated 60/80 mesh glass beads
(Westberg et al., 1974), although coated chromatographic supports also have been used
(Lonneman et al., 1974). Both of the above cryogens are sufficiently warm to allow air to pass
completely through the trap, yet cold enough to collect trace organics efficiently. The use of
cryogenic preconcentration for collection of VOCs, in general, was automated to allow
sequential hourly updates of GC data (McClenny et al., 1984), leading to the initial
configuration of what are now referred to as automated GCs ("auto GCs") for O3 precursor
monitoring. The cryogenic collection procedure also condenses water vapor. An air volume
3-112
-------
of 250 cc at 50% relative humidity and 25 DC contains approximately 2.5 mg of water that
appears as ice in the collection trap. The collected ice at times will plug the trap and stop the
sample flow; furthermore, water transferred to the capillary column during the thermal
desorption step occasionally causes plugging and other deleterious column effects.
To circumvent water condensation problems, Pleil et al. (1987) have characterized the use of a
Nafion® tube drying device to remove water vapor selectively during the sample collection
step. Although hydrocarbon species are not affected, polar organics are partially removed
when the drying device is used. Burns et al. (1983) also showed that partial loss or
rearrangement of monoterpenes, or both (e.g., D-pinene, limonene), occurs when the Nafion®
tube is used to reduce water vapor.
The EPA has recently provided technical guidance for measuring VOCs that is
based on the above studies as well as emerging and developing technology (U.S.
Environmental Protection Agency, 1991c). Guidance for the use of auto GC sampling and
analysis for VOCs has been derived from experience gained from application of this
technology during an O3 precursor study conducted by EPA in Atlanta during the summer of
1990 (Purdue et al., 1992). For that study, an auto GC system developed and manufactured by
Chrompack, Inc., and modified for O3 precursor monitoring (McClenny et al., 1991b) was
used to obtain hourly VOC measurements. The GC system was equipped with a
preconcentration adsorption trap, a cryofocusing secondary trap, and a single analytical
column. The study was focused on the identification and quantitation of 55 O3 precursor
compounds, and resulted in accounting for 65 to 80% of the total NMOC mass. Sample
volumes of 600 cc were used and a detection level of 0.1 ppb C was reported. External
auditing indicated accuracy of +30% at challenge concentrations of 2 ppbC (17-component
audit mixture).
The study also revealed several weaknesses. First of all, excessive amounts of
liquid cryogen were consumed in carrying out the measurements. The inferior quality of the
cryogen containers and poor delivery schedules resulted in reduced data capture. Secondly,
because of the single-column approach, numerous target species either co-eluted or were
poorly resolved. Thirdly, several significant artifact peaks co-eluted with the target species
and, therefore, biased the reported concentrations of those species, as well as the total NMOC
(by summation of peaks). Additionally, Shreffler (1993) reported results from analyses of
canister samples collocated with the automated field GC systems. In general, good agreement
between the systems was found when comparing the sum of the 55 identified O3 precursors.
However, regression analysis indicated that the average total NMOC concentrations found
from the analyses of canister samples were 50% higher than those measured with the field GC
systems.
Based on these operational deficiencies, EPA has challenged commercial GC
instrument makers with improving the current state of the art. One result has been the
evolution of systems that require no liquid cryogen for operation, yet provide sufficient gas
chromatographic resolution of target species (McClenny, 1993; Holdren et al., 1993).
A recent comparison study of auto GCs at Research Triangle Park with five participating
vendors has indicated that the newer auto GC designs use cryogens more efficiently (Purdue,
1993).
In addition to direct sampling via preconcentration with sorbents and cryogenic
techniques, collection of whole air samples is frequently used to obtain NMOC data. Rigid
devices such as syringes, glass bulbs, or metal containers and nonrigid devices such as Tedlar®
3-113
-------
and Teflon® plastic bags are often utilized during sampling. The primary purpose of whole-air
collection is to store an air sample temporarily until subsequent laboratory analysis is
performed. The major problem with this approach is assuring the integrity of the sample
contents prior to analysis. The advantages and disadvantages of the whole air collection
devices were summarized in the 1986 air quality criteria document (U.S. Environmental
Protection Agency, 1986a).
The canister-based method is the preferred means for collecting VOCs and is
described as part of the "EPA Compendium of Methods for the Determination of Toxic
Organic Compounds in Ambient Air" (Compendium Method TO-14). McClenny et al.
(199la) recently reviewed the canister-based method and discussed basic facts about the
canisters, described canister cleaning procedures, contrasted the canister collection system
versus solid adsorbents, and discussed the storage stability of VOCs in canisters. Although
storage stability studies have indicated that many target VOCs can be stored with integrity over
time periods of at least 7 days, there are still many VOCs for which there are no stability data
(Pate et al., 1992; Oliver et al., 1986; Holdren and Smith, 1987; Westberg et al., 1982, 1984;
Gholson et al., 1990). Coutant (1993) has developed a computer-based model for predicting
adsorption behavior and vapor-phase losses in multicomponent systems, based on the potential
for physical adsorption as well as the potential for dissolution in condensed water for canister
samples collected at high humidities. At present, the database for the model contains relevant
physicochemical data for 78 compounds (including water), and provisions for inclusion of up
to 120 additional compounds are incorporated in the software.
Calibration Methods
Calibration procedures for NMOC instrumentation require the generation of dilute
mixtures at concentrations expected to be found in ambient air. Methods for generating such
mixtures are classified as static or dynamic systems.
As described in the previous O3 criteria document (U.S. Environmental Protection
Agency, 1986a), static systems generally are preferred for quantitating NMOCs. The most
commonly used static system is a compressed-gas cylinder containing the appropriate
concentration of the compound of interest. These cylinder gases also may be diluted with
hydrocarbon-free air to provide multi-point calibrations. Cylinders of calibration gases and
hydrocarbon-free air are available commercially. Also, some standard gases such as propane
and benzene, as well as a 17-component ppb mixture, are available from NIST as certified
standard reference materials (SRMs). Commercial mixtures generally are referenced against
these NIST standards. In its recent technical assistance document for sampling and analysis of
O3 precursors, EPA recommended propane (or benzene)-in-air standards for calibration (U.S.
Environmental Protection Agency, 1991c). Some commercially available propane cylinders
have been found to contain other hydrocarbons (Cox et al., 1982), so that all calibration data
should be referenced to NIST standards.
Because of the uniform carbon response of a GC-FID system (+10%) to
hydrocarbons (Dietz, 1967), a common response factor is assigned to both identified and
unknown compounds obtained from the speciation systems. If these compounds are
oxygenated species, an underestimation of the actual concentrations will be reported. Dynamic
calibration systems are employed when better accuracy is needed for these oxygenated
hydrocarbon species. Dynamic systems normally are employed to generate in situ
3-114
-------
concentrations of the individual compound of concern and include devices such as permeation
and diffusion tubes and syringe delivery systems.
3.5.2.3 Carbonyl Species
Historically, the major problem in measuring concentrations of carbonyls in
ambient air has been to find an appropriate monitoring technique that is sensitive to low
concentrations and specific for the various homologues. Early techniques for measuring
HCHO, the most abundant aldehyde, were subject to some interferences and lacked sensitivity
at low parts-per-billion concentrations (Altshuller and Leng, 1963; Altshuller et al., 1961;
Altshuller and McPherson, 1963). However, spectroscopic methods such as FTIR and DO AS
also lack sensitivity for HCHO in the low parts-per-billion concentration range. The 1986 air
quality criteria document described two methods frequently used: (1) the chromotropic acid
(CA) method for HCHO and (2) the 3-methyl-2-benzothiazolone hydrazone (MBTH) technique
for total aldehydes (U.S. Environmental Protection Agency, 1986a). However, spectroscopic
methods, on-line colorimetric methods, and the HPLC method employing DNPH derivatization
are the preferred methods currently used for measuring atmospheric levels of carbonyl species.
Spectroscopic Methods
Three spectroscopic methods have been used to make measurements for atmospheric
levels of HCHO and were recently intercompared at an urban site in California (Lawson et al.,
1990). The FTIR method used gold-coated 30-cm-diameter mirrors and a total optical path of
1,150 m. The 2,781.0-cm"1 "Q-branch" adsorption peak was used to measure HCHO. The
limit of detection was 3 ppb, and the measurement errors were within +3 ppb. The DO AS
method was operated at an 800-m pathlength, and an absorption peak at 339 nm was used to
measure HCHO; NO2 and HNO2 spectral features were subtracted. The limit of detection was
4.5 ppb, and the experimental error was +30%. A TOLAS method was operated at a
pathlength of 150 m. Laser diodes were mounted in a closed-cycle helium cryocooler with a
stabilizing heater circuit for constant temperature control. Radiation from the diode was
collected and focused into the sampling by reflective optics. Formaldehyde absorption was
measured at 1,740 cm"1. The limit of detection was 0.1 ppb and the measurement errors were
within +20%. Additional information on FTIR and DO AS has been reported by Winer et al.
(1987), Atkinson et al. (1988), and Biermann et al. (1988). A more complete description of
TOLAS is given by MacKay and Schiff (1987a).
On-line Fluorescence Method
A wet chemical method based on the derivatization of HCHO in aqueous solution to
form a fluorescent product was developed by Kelly et al. (1990) and recently tested (Kelly and
Fortune, 1994). The detection of fluorescent product was made more sensitive by using
intense 254 nm light from a mercury lamp for excitation. This procedure allowed the use of a
simple and efficient glass coil scrubber for collection of gaseous HCHO. A detection limit of
0.2 ppb was obtained with a response time of 1 min. The instrument is portable and highly
selective for HCHO.
High-Performance Liquid Chromatography-2,4-Dinitrophenylhydrazine Method
The preferred and most current method for measuring aldehydes in ambient air is
one involving derivatization of the aldehydes concurrent with sample collection, followed by
3-115
-------
analysis using HPLC. This method takes advantage of the reaction of carbonyl compounds
with DNPH to form a 2,4-dinitrophenylhydrazone:
1CDO D NH2NHC6H3(NO2) D H2O D RRDCDNNHC&H3(NC (3.82)
Because DNPH is a weak nucleophile, the reaction is carried out in the presence of acid in
order to increase protonation of the carbonyl.
In this method, atmospheric sampling initially was conducted with micro-impingers
containing an organic solvent and an aqueous, acidified DNPH reagent (Papa and Turner,
1972; Katz, 1976; Smith and Drummond, 1979; Fung and Grosjean, 1981). After sampling
was completed, the hydrazone derivatives were extracted, and the extract was washed with
deionized water to remove the remaining acid and unreacted DNPH reagent. The organic
layer was then evaporated to dryness, subsequently dissolved in a small volume of solvent, and
analyzed by reversed-phase liquid chromatographic techniques employing a UV detection
system (360 nm).
An improved procedure subsequently was reported that is much simpler than the
above aqueous impinger method (Lipari and Swarin, 1982; Kuntz et al., 1980; Tanner and
Meng, 1984). This scheme utilizes a midget impinger containing a C2H3N solution of DNPH
and an acid catalyst. After sampling, an aliquot of the original collection solution is injected
directly into the liquid chromatograph. This approach eliminates the extraction step and
several sample-handling procedures associated with the DNPH-aqueous solution and provides
much better recovery efficiencies. This method has been formalized by EPA as Compendium
Method TO-5 (Winberry et al., 1988). The TO-5 Method has been further modified to include
the use of a DNPH-impregnated solid adsorbent, rather than DNPH impinger solutions, as the
collection medium. This modification and associated sampling conditions are referred to as
EPA Method TO-11. The methodology can be used easily for long-term (1 to 24 h) sampling
of ambient air. Sampling rates of 500 to 1,200 cc/min can be achieved and detection levels of
1 ppbv can be attained with sampled volumes of 100 L. The method currently calls for the use
of SepPak® silica gel material as the sorbent material. However, researchers have noted that
O3 present in ambient air reacted more easily with carbonyl compounds collected on DNPH-
coated silica gel cartridges than on DNPH-coated Clg-bonded silica material. To eliminate this
negative bias, these researchers used an O3 scrubber (Arnts and Tejada, 1989). Smith et al.
(1989) noted that artifact peaks occurred when O3 was bubbled through impingers containing
DNPH solution. These artifacts were identified as DNPH-O3 reaction products and were
shown to cause positive interferences unless they were chromatographically resolved from the
HCHO-hydrazone derivative. Although the TO-11 Method has been included in EPA's
Technical Assistance Document for Sampling and Analysis of Ozone Precursors (U.S.
Environmental Protection Agency, 1991c), a Kl-coated denuder tube also has been
recommended to remove O3 upstream of the DNPH-coated cartridges.
Although both the Sep-Pak® Clg and silica gel cartridges have been used by
researchers (either with or without O3 scrubbers), there is still considerable uncertainty as to
which type of cartridge gives the most reliable carbonyl results. Experiments are currently
underway at several laboratories to investigate the effect of O3 on the performance of
DNPH-coated cartridges. However, results have not yet been reported in the literature.
3-116
-------
Calibration of Carbonyl Measurements
Because they are reactive compounds, it is extremely difficult to make stable
calibration mixtures of carbonyl species in pressurized gas cylinders. Although gas-phase
3-117
-------
standards are available commercially, the vendors do not guarantee long-term stability and
accuracy.
Formaldehyde standards generally are prepared by one of several methods. The
first method utilizes dilute commercial formalin (37% HCHO, w/w). Calibration is
accomplished by the direct spiking into sampling impingers of the diluted mixture or by
evaporation into known test volumes, followed by impinger collection. Formaldehyde also
can be prepared by heating known amounts of paraformaldehyde, passing the effluent gases
through a methanol-liquid nitrogen slush trap to remove impurities, and collecting the
remaining HCHO. Paraformaldehyde permeation tubes also have been used (Tanner and
Meng, 1984).
For the higher molecular weight carbonyl species, liquid solutions can be
evaporated, or pure vapor can be generated in dynamic gas-flow systems (permeation tubes,
diffusion tubes, syringe delivery systems, etc.). These test atmospheres are then passed
through the appropriate collection system and analyzed. A comparison of these data, with the
direct spiking of liquid carbonyl species into the particular collection system, provides a
measure of the overall collection efficiency.
3.5.2.4 Polar Volatile Organic Compounds
The VOCs discussed earlier in this chapter (Section 3.5.2.2) have included
aliphatic, aromatic, alkenic, and acetylenic hydrocarbons. These compounds are relatively
nonpolar, nonreactive species, and measurement methods have been easily applied in
determining ambient concentrations.
Recently, attention also has been directed toward the more reactive oxygen- and
nitrogen-containing organic compounds, in part by the inclusion of many of these compounds
on a list of 189 hazardous air pollutants specified in the 1990 CAAA (U.S. Congress, 1990).
Many of these compounds are emitted directly from a variety of industrial processes, mobile
sources, and consumer products, and some also are formed in the atmosphere by
photochemical oxidation of hydrocarbons. However, as indicated earlier in this document,
very few ambient data exist for these species. These compounds have been referred to
collectively as PVOCs, although it is their reactivity and water solubility, more than simple
polarity, that make their measurement difficult with existing methodology.
Two approaches have been utilized in developing analytical methods for PVOCs.
One approach has incorporated the use of cryogenic trapping techniques similar to those
discussed earlier for the nonpolar hydrocarbon species; the second approach has utilized
adsorbent material for sample preconcentration. To be effective for sensitive parts-per-billion
measurement of PVOCs, both approaches require some type of water management system to
mitigate the adverse effects that water has on the chromatography and detector sensitivity and
reliability. Several researchers have reported the use of cryogenic trapping with
two-dimensional chromatography to selectively remove water vapor from the analytical
process (Pierotti, 1990; Cardin and Lin, 1991). Although this column "heart cutting"
technique has been successful for selected compounds, additional studies are needed to
determine its potential use for the wide range of PVOCs. Ogle et al. (1992) developed a
novel water management system based on the condensation of moisture from the saturated
carrier gas stream during thermal desorption of a cryogenic trap. The moisture management
system was found to be effective for reducing the amount of water delivered to the column
during laboratory analyses of spiked mixtures. However, the system has not yet been
extended to field monitoring. Gordon and Miller (1989) have used cryogenic trapping and
3-116
-------
GC/MS spectrometry techniques to demonstrate the potential of chemical ionization (CI)
within an ion trap to detect PVOCs. The water vapor present in the sample served as the CI
reagent gas and appeared to be an effective reagent gas; however, deleterious chromatography
results were also encountered. The authors concluded that further laboratory work is needed
before this methodology can be applied to ambient air monitoring. Martin et al. (1991) also
reported the use of cryogenic trapping with a GC-FID system to measure ambient levels of
isoprene and two of its oxidation products, methacrolein and methylvinyl ketone (detection
level of 0.5 ppb). The water vapor was selectively removed by using a potassium carbonate
(K2CO3) trap ahead of the cryogenic trap. Frequent replacement of the K2CO3 trap was
required.
The use of solid adsorbents for sample preconcentration of PVOCs has been
reported by Kelly et al. (1993). The analytical method was used extensively at two field sites
that formerly were used in EPA's Toxic Air Monitoring Study (TAMS). The analytical
method consisted of GC separation of PVOCs with quantification by a ion-trap mass
spectrometer. A two-stage adsorbent trap containing Carbopack B and Carbosieve S-III
(Supelco catalog number 2-0321) was used to separate water vapor from the PVOCs. The
optimum room temperature trapping and drying procedure consisted of a 320-cc sample
(100 cc/min) followed by a dry nitrogen purge of 1,300 cc (100 cc/min). The trap was then
backflushed and thermally desorbed with helium at 220 °C. A 5-min, 260 °C trap bakeout
followed each collection-analysis cycle. The target list contained 14 PVOCs, including
alcohols, ethers, esters, and nitrile species. Individual detection limits ranged from 0.2 to
1 ppb.
3.5.3 Sampling and Analysis of Nitrogen Oxides
3.5.3.1 Introduction
The measurement of NOX in ambient air is of interest because of the role that
certain of those compounds play as precursors to O3 and because NO2 has been shown to
impact health effects. Most of the NOX emitted from combustion sources are NO and NO2.
Collectively these two compounds are called NOX. They contribute to O3 formation by means
of reactions discussed in Section 3.2. As a result, measurement of NOX is important in efforts
to understand and control O3 and NO2 in ambient air.
The atmospheric photochemistry that produces O3 also results in conversion of NO
and NO2 to products such as HNO3, nitrous acid (HNO2), organic nitrates such as PAN
(CH3C(O)O2NO2), and other species. The total of all of these labile nitrogen species in air,
NOX included, is termed NOy. Such compounds may be labile via photolysis (e.g., HNO2) or
thermal decomposition (e.g., PAN), and may be toxic, irritating, or acidic. The organic
nitrates can occur in the atmosphere as reservoirs for NO2. However, in general, they do not
play the same critical role that NO2 and NO play as O3 precursors. For that reason, this
section focuses on measurement methods for NO and NO2, as the primary O3 precursors of
NOX. Nitrogen oxides other than NOX may be important, however, as interferents in efforts to
measure NO and NO2. These non-NOx species are considered in this section in that regard.
Measurements of NOX may involve measurements of NO, of NO2, or of the sum of
NOX. Nitrogen dioxide, but not NO, is a criteria air pollutant, and, thus, reference and
equivalent methods are specified for NO2 measurements. In this section, the current state of
measurement methods for NO and NO2 will be summarized separately. Such methods in
some cases rely on measurements of total NOX, or at least an approximation of NOX. This
3-117
-------
discussion focuses on current methods and on promising new technologies, but no attempt is
made here to cover the extensive history of the development of these methods. More detailed
discussions of such methods may be found elsewhere (U.S. Environmental Protection Agency,
1993c; National Aeronautics and Space Administration, 1983). Wet chemical methods are no
longer commonly used and are not discussed here; a review of such methods is given by
Purdue and Hauser (1980).
3.5.3.2 Measurement of Nitric Oxide
Gas-Phase Chemiluminescence Methods
By far the most common method of NO measurement is gas-phase CL with O3. In
this method, excess O3 is added to air containing NO in a darkened, internally reflective
chamber monitored by a photomultiplier tube. A small portion of the NO reactions with
O3 produce electronically excited NO2 molecules that decay by emission of light of
wavelengths longer than 600 nm. The emitted light is detected by a red-sensitive
photomultiplier tube, through an optical filter that prevents passage of wavelengths shorter
than 600 nm. This optical filtering minimizes interference from CL produced by O3 reactions
with other species (e.g., hydrocarbons). The excited NO2 is readily quenched in air, so that,
in typical instruments, air and O3 are mixed at reduced pressure (i.e., at least 20 in. of Hg
vacuum). The intensity of the emitted light is linearly proportional to the NO content of the
sample air over several orders of magnitude in concentration.
Commercial CL instruments for continuous measurement of NO are available from
several manufacturers. The chemiluminescence approach is also an EPA-designated
measurement principle for measuring ambient NO2; it requires a means of converting NO2 to
NO for detection. The complexities of this conversion are discussed in Section 3.5.3.3, on
NO2 methods. The commercial NO monitors typically are claimed to have detection limits of
a few parts per billion by volume in air, with response time of a few minutes. Field
evaluations of several commercial instruments have indicated that minimum levels of
detection for NO2 are 5 to 13 ppbv (Michie et al., 1983; Holland and McElroy, 1986).
However, more recent evaluations have indicated better performance. Rickman et al. (1989)
reported detection limits of 0.5 to 1 ppbv, and precision of ±0.3 ppbv, from laboratory and
field evaluations of two commercial instruments operated on their 50 ppbv full-scale ranges.
Commercial NO analyzers are portable and quite reliable and now are commonly used in
ambient air monitoring networks.
Commercial NO analyzers may not have sensitivity sufficient for surface
measurements in urban, rural, or remote areas, or for airborne measurements. As a result,
several investigators have devised modifications to commercial instruments to improve their
sensitivity and response time (Delany et al., 1982; Tanner et al., 1983; Dickerson et al., 1984;
Kelly et al., 1986). Those modifications include operating at low pressure and high sample
flow rate; using a larger, more reflective reaction chamber that promotes mixing of the
reactants close to the photomultiplier tube; increasing the O3 supply; for example, by use of
oxygen in the O3 source; cooling of the photomultiplier to reduce noise; adopting
photon-counting techniques for light detection; and adding a prereactor to obtain a more
stable and appropriate background signal. Commercial instruments modified in these ways
are generally reported to have detection limits of 0.1 ppbv or less, with response times of 30 s
or less.
Research-grade NO instruments specially designed for ultra-high sensitivity also
have been built for use in remote ground-level or airborne applications (e.g., Ridley and
3-118
-------
Hewlett, 1974; Kley and McFarland, 1980; Kelly et al., 1980; Helas et al., 1981; Drummond
et al., 1985; Torres, 1985; Kondo et al., 1987; Parrish et al., 1990). These instruments
typically have detection limits of 10 ppt (i.e., 0.01 ppbv) or less, with response times from a
few seconds to 1 min.
A number of studies indicate that the CL method is essentially specific for NO.
Operation at reduced pressure prevents interference resulting from quenching by water vapor
(Michie et al., 1983; Drummond et al., 1985). In air sampling, no significant interferences
have been found in NO detection from sulfur-, chlorine-, and nitrogen-containing species
(Joshi and Bufalini, 1978; Sickles and Wright, 1979; Grosjean and Harrison, 1985b; Fahey
et al., 1985). However, H2S and possibly other sulfur-containing compounds from seawater
have been reported to give false NO signals (Zafiriou and True, 1986). This effect should not
be important for ambient air measurements. Fahey et al. (1985) and Drummond et al. (1985)
also reported no significant NO interference from a variety of other nitrogen-containing
species, including NO2, HNO3, PAN, N2O5, NH3, HCN, N2O, and HO2NO2; as well as no
interference from CH4, propylene, and H2O2.
Several ambient air intercomparisons have been done of CL NO instruments
(Walega et al., 1984; Hoell et al., 1987; Fehsenfeld et al., 1987; Gregory et al., 1990a).
These studies have focused on high-sensitivity research instruments, rather than the
commercial instruments used for widespread ambient air measurements. These studies have
shown excellent agreement among the CL NO instruments, even at NO levels in the low ppt
range (Hoell et al., 1987; Gregory et al., 1990a). These results support the validity of the CL
approach for NO. Good agreement also has been found between CL measurements and
spectroscopic NO measurements in these studies (see Section 3.5.3.2).
Spectroscopic Methods for Nitric Oxide
Direct spectroscopic methods for NO include two-photon laser-induced
fluorescence (TPLIF), TOLAS, and two-tone frequency-modulated spectroscopy (TTFMS).
The primary characteristics of these methods are their very high sensitivity and selectivity for
NO. For example, a detection limit of 10 ppt has been quoted for TPLIF with a 30-s
integration time, with no significant interferences from atmospheric species (Davis et al.,
1987). An accuracy of ±16% as a 90% confidence limit has been calculated for NO
measurement by TPLIF from an aircraft (Davis et al., 1987). The TOLAS method is
similarly highly selective for NO and achieves a detection limit of 0.5 ppbv (Schiff et al.,
1983). The response time of the TOLAS instrument is about 1 min for NO, and is limited by
stabilization of concentrations with the large surface area of the multi-pass White cell. The
newest method is TTFMS, which appears in laboratory studies to be very sensitive, fast, and
selective. With a 100-m path length in a 20-torr multiple-pass cell, and a 1-min averaging
time, the detection limit of NO is estimated to be 4 ppt (Hansen, 1989).
Spectroscopic methods have compared well with the CL method for NO in
ambient measurements. Walega et al. (1984) reported good agreement between CL and
TOLAS results for NO in laboratory air, in ambient air, and in downtown Los Angeles air.
Gregory et al. (1990a) reported comparisons of TPLIF and CL NO methods in airborne
measurements. Agreement at levels below 20 ppt was within the expected accuracy and
precision of the instruments (i.e., within 15 to 20 ppt).
The major drawbacks of these spectroscopic methods are their complexity, size,
and cost. Although possessing remarkable characteristics, these methods are restricted to
3-119
-------
research applications. The TTFMS approach, in fact, is still in the laboratory development
stage.
Passive Samplers
At present, no passive sampler exists that directly measures NO. Instead, passive
samplers developed for NO2 have been adapted for NO measurement, using an oxidizing
material that converts NO to NO2. Palmes tubes (Palmes and Tomczyk, 1979) have been
adapted for NO measurement by using two tubes in parallel. One tube collects NO2 on a
triethanolamine (TEA)-coated grid, whereas the other collects NO2 on a TEA grid, plus NO
oxidized by a chromic acid-coated surface. The grids are then extracted and analyzed for
NO2" ion. Nitric oxide is determined by difference between the two results, after accounting
for the different diffusivities of NO and NO2. The sampling rates depend on temperature and
air velocity. The tubes cannot be used for periods longer than 24 h and are intended for use
at ppm NO levels important in the workplace (e.g., 2 to 200 ppm -h). Applicability to
ambient NO levels has not been demonstrated.
A more sensitive passive sampler for NO has been reported (Yanagisawa and
Nishimura, 1982) that uses the same TEA chemistry, with CrO3 as the NO oxidizer.
A detection limit of 70 ppbv-h has been reported. As with any currently available passive
sampler, the disadvantages of the method are the potential for interferences, relatively poor
precision, and low sensitivity for ambient air measurements.
3.5.3.3 Measurements for Nitrogen Dioxide
Gas-Phase Chemiluminescence Methods
In 1976, the gas-phase CL approach described above for NO detection was
designated as the measurement principle on which EPA reference methods for ambient NO2
must be based. The CL method thus filled the vacancy left by withdrawal of the Jacobs-
Hochheiser method, because of technical problems, in 1973. To be designated as a reference
method, an NO2 detection method must use the CL approach and be calibrated by the
specified methods (gas-phase titration of NO with O3, or use of an NO2 permeation device).
In addition the instrument must meet the performance specifications shown in Table 3-18.
An equivalent method, either manual or automated, must meet certain requirements for
comparability with a reference method when measuring simultaneously in a real atmosphere.
Those comparability requirements are shown in Table 3-19. An automated equivalent method
must also meet the performance requirements shown in Table 3-18.
The selection of the O3-CL method as the reference measurement principle for
ambient NO2 was the result of comparison tests of CL and wet chemical methods.
Chemiluminescence analyzers were found superior to the wet chemical methods in response
time, zero and span drift, and overall operation, although agreement among all the methods
tested was good, at the NO2 spike levels provided (Purdue and Hauser, 1980). Table 3-20
lists the methods currently designated (as of August 1, 1994) by EPA as reference and
equivalent methods for ambient NO2. Three wet chemical methods are shown as equivalent
methods, but these rarely are used for ambient air measurements.
The O3 CL method does not measure NO2 directly, because the CL is produced by
reaction of NO with O3. As a result, NO2 must first be reduced to NO for detection.
In principle, such a reduction should readily result in measurement of NO + NO2 (i.e.,
3-120
-------
Table 3-18. Performance Specifications for Nitrogen Dioxide Automated
Methods3
Performance Parameter Units NO2
Range ppm 0-0.5
Noise
0% Upper range limit ppm 0.005
80% Upper range limit ppm 0.005
Lower detectable limit ppm 0.01
Interference equivalent
Each interferant (SO2,NO,NH3,H2O) ppm ±0.02
Total interferant ppm <0.04
Zero drift, 12 and 24 hours ppm ±0.02
Span drift, 24 hours
20% Upper range limit % ±20.0
80% Upper range limit % ±5.0
Lag time min 20
Rise time min 15
Fall time min 15
Precision
20% Upper range limit ppm 0.02
80% Upper range limit ppm 0.03
aSee Appendix A for abbreviations and acronyms.
Source: Code of Federal Regulations (1987), Ambient Air Monitoring Reference and Equivalent Methods,
C.F.R. Title 40, Part 53.
Table 3-19. Comparability Test Specifications for Nitrogen Dioxide
Nitrogen Dioxide Maximum Discrepancy
Concentration Range (ppm) Specification (ppm)
Low 0.02 to 0.08 0~02
Medium 0.10 to 0.20 0.02
High 0.25 to 0.35 0.03
NOX), and allow indirect measurement of NO2 by difference between NO and NOX responses,
measured either sequentially, or simultaneously by separate detectors. In practice, however,
selective measurement of NOX by this approach has proven very difficult.
Several methods have been employed to convert NO2 to NO, including catalytic
reduction with heated molybdenum or stainless steel, reaction with CO over a gold catalyst
surface, reaction with ferrous sulfate (FeSO4) at room temperature, reaction with carbon at
200 °C, and photolysis of NO2 at wavelengths of about 320 to 400 nm (Kelly et al., 1986). It
has been found in many separate investigations that the heated converters reduce NO2 to
3-121
-------
Table 3-20. Reference and Equivalent Methods for Nitrogen
Designated by U.S. Environmental Protection Agency
Method
Reference Methods (Continuous CL Analyzers)
Advanced Pollution Instrumentation 200
Beckman 952A
Bendix 8101-B
Bendix 8101-C
CSI 1600
Dasibi 2108
Lear Siegler ML9841
Meloy NA53OR
Monitor Labs 8440E
Monitor Labs 8840
Monitor Labs 8841
Philips PW9762/02
Thermo Electron 14B/E
Thermo Electron 14D/E
Thermo Environmental 42
Designation
Number
RFNA-069 1-082
RFNA-0 179-034
RFNA-0479-038
RFNA-0777-022
RFNA-0977-025
RFNA- 11 92-089
RFNA- 1292-090
RFNA- 1078-031
RFNA-0677-021
RFNA-0280-042
RFNA-099 1-083
RFNA-0879-040
RFNA-0 179-03 5
RFNA-0279-037
RFNA- 1289-074
Dioxide
a
Method
Code
082
034
038
022
025
089
090
031
021
042
083
040
035
037
074
Equivalent Methods (Wet Chemical)
Sodium arsenite
Sodium arsenite/Technicon II
TGS-ANSAb
EQN-1277-026
EQN-1277-027
EQN-1277-028
084
084
098
aAs of August 1, 1994.
bTriethanolamine-guaiacol-sulfite with 8-amino-l-naphthalene-sulfonic acid ammonium salt.
NO effectively, but also reduce other NOy species as well (e.g., Winer et al., 1974; Cox,
1974; Joseph and Spicer, 1978; Grosjean and Harrison, 1985b; Fahey et al., 1985).
Efficiencies of conversion near 100% are reported in these studies for NO2 and for NOy
species such as HNO3, HNO2, PAN, and organic nitrates. This finding is particularly
important for widespread monitoring networks that use commercial instruments, because such
instruments without exception use heated catalytic converters (typically molybdenum). Thus,
such instruments measure not NO and NOX, but more nearly NO and total NOy. Although
NOX is the predominant NOy species during early morning hours, other NOy species constitute
a substantial percent of the NOy later in the day, especially in rural areas. The NO2 value
inferred from such measurements may be significantly in error (see below), and may in turn
affect the results of models of ambient O3. The completeness of the measured NOy value is
also questionable because, for example, HNO3 is readily lost to surfaces, and, in ambient
sampling, may be removed within the sampling system before reaching the heated converter.
3-122
-------
Other conversion methods for NO2 have been tried in an effort to achieve higher
selectivity. Ferrous sulfate has been used for ambient NO2 measurements using high-
sensitivity research grade CL instruments (e.g., Kelly et al., 1980; Helas et al., 1981;
Dickerson et al., 1984). This material is an efficient reducer of NO2, but also has been found
to convert a portion of PAN, and possibly a portion of HNO2 and organic nitrates (Fehsenfeld
et al., 1987). Memory effects and reduction in efficiency can occur because of humidity
effects (Fehsenfeld et al., 1987). As a result of these characteristics, use of FeSO4 has given
high readings in comparison with spectroscopic instruments and the photolytic NO2 converter,
and its use likely results in overestimating ambient NOX by a significant amount (Fehsenfeld
et al., 1987; Ridley et al., 1988a; Gregory et al., 1990b). Ferrous sulfate has never been used
in commercial NOX instruments and is no longer used in research measurements.
The most specific method for converting NO2 to NO is photolysis (Kley and
McFarland, 1980). In the most common approach, ambient NO2 is photolyzed to NO by light
of 350 to 410 nm from a xenon arc lamp. The method does not produce NO from the major
potential interferents present in air (i.e., HNO3, PAN, and organic nitrates), but less abundant
NOy species such as FDSTO2 or HO2NO2 may interfere. A detailed description of steps to
minimize such interferences is given by Ridley et al. (1988b). As currently used, the
photolytic converter appears to be essentially specific for NO2. However, it does not provide
complete conversion of NO2. Conversion efficiencies are 50 to 60% with a new lamp but
may decline to 20% over the course of several weeks (Parrish et al., 1990). Thus, the
conversion efficiency must be calibrated repeatedly. This approach has not been implemented
with commercial NO detectors but has been implemented with research-grade CL NO
instruments for studies of NOX and NOy chemistry at a variety of locations (e.g., Buhr et al.,
1990; Parrish et al., 1990; Trainer et al., 1991; Parrish et al., 1992, 1993). The photolytic
method compares well with other techniques, including spectroscopic methods, even at NO2
levels as low as 0.05 ppbv (Gregory et al., 1990b). Further improvement of the photolytic
converter approach is continuing. Bradshaw et al. (1994) reported on plans to minimize wall
effects in the photolytic converter and to use a metal halogen lamp in place of the xenon arc
lamp. The metal halogen lamp emits strongly in the proper wavelength region and is much
less expensive than the xenon arc lamp, allowing more frequent replacement of the lamp and
consequently higher long-term photolytic efficiency.
As noted above, the commercial CL analyzers used for most ambient air NO and
NOX measurements actually measure NO and NOy. The magnitude of the resulting
overestimation of NO2, determined by difference, obviously depends on the portion of NOy
that is NOX. The smaller the portion of NOy that is NOX, the greater will be the error in the
NO2 determined by difference. In rural/remote areas, where NOX has undergone extensive
conversion to other products during transport from a source region, NOX may contribute a
small fraction of NOy. In urban areas, close to sources, NOX may comprise nearly all of NOy.
For example, in measurements at Point Arena, Parrish et al. (1992) report NOx/NOy ratios
averaging 0.3 in air of marine origin and 0.75 in air subject to continental influence. Buhr et
al. (1990) and Parrish et al. (1993) reported measurements at rural sites in eastern North
America that indicate NOx/NOy ratios ranging from about 0.25 to 0.75, varying with the time
of day, with the lowest ratios occurring during daytime, photochemically active periods.
Clearly, although the commercial CL instruments are designated as reference methods for
NO2, the great majority of existing ambient air data for NO2 or NOX are biased high, due to
the inclusion of some portion of other NOy species. The magnitude of this bias may not be
large in urban areas, but, in any case, it is essentially unknown at this time.
3-123
-------
Luminol Chemiluminescence Method
This approach is based on the CL reaction of gaseous NO2 with the surface of an
aqueous solution of luminol (5-amino-2,3-dihydro-l,4-phthalazinedione). Emission occurs
primarily between 380 and 520 nm. In commercial instruments, luminol solution flows down
a fabric wick that lies vertically on a clear window viewed by a photomultiplier tube.
Nitrogen dioxide in sample air passing over the wick produces light, the intensity of which is
proportional to the NO2 concentration. Commercial instruments using this approach are
compact, light, and relatively inexpensive and can provide detection limits as low as
0.01 ppbv, with response times below 30 s. The instrument has the advantage of detecting
NO2 directly. However, several difficulties have had to be dealt with in developing the
method.
Original reports of the approach (Maeda et al., 1980) indicated positive
interferences from O3 and SO2 and a negative one from CO2. Reformulation of the luminol
reagent solution has minimized, although not fully eliminated, those interferences (Wendel
et al., 1983; Schiff et al., 1986). Reported effects include a slight negative response from
NO, and sensitivity to PAN, HNO2, and O3 (Wendel et al., 1983; Schiff et al., 1986; Rickman
et al., 1989; Kelly et al., 1990; Spicer et al., 1991). Response to NO2 may be nonlinear at
low concentrations (Kelly et al., 1990), although recent reformulation of the reagent
apparently has reduced this behavior (Busness, 1992). Evaluation of the luminol NO2 monitor
indicates that great care must be taken in using and calibrating the instrument in order to
achieve good precision and accuracy in ambient measurements (Kelly et al., 1990). The
monitor has been used widely as a research tool, but has not been used widely in ambient air
monitoring nor has it been designated an equivalent method for NO2.
An O3 scrubber is available to eliminate the O3 interference noted above, but it
also was found to remove a portion of the NO2 (Kelly et al., 1990). The luminol approach
also has been modified to measure NO, by using a CrO3 converter that oxidizes NO to NO2
for detection. Thus NO is detected by difference. This method has the potential for
measurement of total NOX; however, evaluations of the CrO3 converter are still underway at
several laboratories. Given the known interferences in the luminol approach, careful
evaluation of this method must be completed before it gains acceptance as an NO
measurement method.
An adaptation of the commercial luminol NO2 detector has been reported to
provide measurements of total NOy, NO2, and NOX (Drummond et al., 1992). This adaptation,
called the LNC-3M, uses a commercial luminol instrument for NO2 detection, with a
CrO3 converter for NOX detection. The NOX measurement must be corrected for the few
percent of the ambient NO2 that is lost in the CrO3 converter (Drummond et al., 1992). The
NOy measurement is achieved using a stainless steel converter maintained at 400 °C. Tests
indicate that this converter provides a more complete conversion of alkyl nitrates, and
consequently a more complete measurement of NOy, than is provided by either the heated
molybdenum converters used in commercial O3 CL NOX detectors or the gold converters with
CO addition used in research instruments (Drummond et al., 1992). The LNC-3M adds a
small amount of NO2 to the sample to eliminate the nonlinearity at low concentrations, and
uses a zeroing scrubber that greatly reduces the interference from PAN. However, this
scrubber must be replaced weekly when it is in continuous use (Drummond et al., 1992).
Spectroscopic Methods
3-124
-------
Several spectroscopic approaches to NO2 detection have been developed; TDLAS,
TTFMS, DO AS, and differential absorption lidar (DIAL) are absorption methods that have
been used. The TDLAS method is probably the most commonly used spectroscopic NO2
method. It can provide high selectivity for NO2, with a detection limit of 0.1 ppbv, accuracy
of ±15%, and a response time on the order of 1 min because of the White cell (Mackay and
Schiff, 1987b). The DOAS method is an open-path, long-pathlength system. The detection
limit for NO2 with a 0.8-km pathlength and 12-min averaging time has been reported as
4 ppbv, with measurement accuracy reported as ±10% (Biermann et al., 1988). However,
recent improvements have resulted in a commercial DOAS instrument capable of an NO2
detection limit of 0.6 ppbv, based on a 557-m path and a 1-min averaging time (Stevens et
al., 1993). The detection limit for NO2 by the DIAL technique has been reported as 10 ppbv
with a 6-km pathlength (Staehr et al., 1985). The novel TTFMS method noted above for NO
is reported to have an NO2 detection limit of 0.3 ppt, but is not fully proven for ambient
measurements.
Fluorescence methods also have been used for NO2, including photofragmentation
TPLIF (PF/TPLIF) (Davis, 1988). This method uses two cells in which NO is measured by
TPLIF. In one of the cells, an excimer laser emitting at 353 nm photolyzes NO2 to NO for
detection. Thus NO2 is ultimately measured, by difference, as NO, but the NO is formed
directly by photolysis of NO2. With a 2-min integration time, an NO2 detection limit of
12 ppt is reported. The method is highly selective for NO2, because an interferant would
have to photolyze to produce NO. Several potential atmospheric species have been ruled out
in this regard (Davis, 1988).
The drawbacks of most of these methods are, as noted earlier, complexity, size,
and cost. At present, these factors outweigh the obvious advantages of the sensitivity and
selectivity of these spectroscopic methods and largely have restricted the use of these
NO2 methods to specific research applications or as reference methods in intercomparisons.
In such intercomparisons, absorption measurements have been used most commonly. The
TDLAS method has been used in ground-level comparisons with O3 CL and luminol
instruments to provide specific NO2 measurements (Walega et al., 1984; Sickles et al., 1990;
Fehsenfeld et al., 1990) and in an airborne comparison with PF/TPLIF and O3 CL instruments
(Gregory et al., 1990b). A finding of these studies was that the TDLAS consistently read
higher than other established methods at very low NO2 levels (i.e., <0.4 ppbv) (Fehsenfeld et
al., 1990; Gregory et al., 1990b).
The spectroscopic NO2 method most fully developed beyond the research stage is
the DOAS technique. Stevens et al. (1993) report testing of a commercial DOAS instrument
in North Carolina over 17 days in the fall of 1989. The DOAS measured NO2 using
wavelengths between 400 and 460 nm and achieved a detection limit of 0.6 ppbv, as noted
above. Simultaneous measurements of O2, SO2, HCHO, and HNO2 also were provided by the
DOAS instrument. Comparison of the DOAS NO2 results to those from a commercial CL
detector showed (DOAS NO2) = 1.14 x (CL NO2) ±2.7 ppbv, with an r2 of 0.93, at NO2
levels up to 50 ppbv (Stevens et al., 1993). The sensitivity, stability, response time, and
multicomponent capability are the primary advantages of the DOAS approach. Further
intercomparisons and interference tests are recommended (Stevens et al., 1993).
Passive Samplers
3-125
-------
Passive samplers are attractive, inexpensive, and simple means to obtain long-term
or personal exposure data for NO2 or NOX. The simplest passive sampler for NO2 is the
nitration plate, which is essentially an open dish containing filter paper impregnated with
TEA. Nitrogen dioxide diffuses to the paper and is extracted later as NO2" for analysis.
No diffusion barrier exists in this approach or in a similar approach using a candle-shaped
absorber (Kosmus, 1985). Consequently, results are very much subject to ambient conditions
and give, at best, a qualitative indication of NO2 or NOX.
Addition of a diffusion barrier to the nitration plate concept has led to badge-type
passive samplers for NO2 (e.g., Mulik and Williams, 1986, 1987; Mulik et al., 1989, 1991).
In general, such devices use perforated screens, plates, or filters as diffusion barriers on the
chemically reactive material, which may be exposed on one or both sides, depending on the
application. Extraction of the sorbent then allows measurement of the NO2 collected,
typically as NO2" ion. Such a device using TEA as the active material gave very good
agreement relative to a CL analyzer in laboratory tests with NO2 at 10 to 250 ppbv (Mulik
and Williams, 1987). However, interferences from PAN and HNO2 (the latter in both outdoor
and indoor air) are expected (Sickles and Michie, 1987). Comparison of ambient NO2 results
in the 5 to 25 |ig/m3 range (i.e., about 2.5 to 12.5 ppbv) from the passive device to those
from TDLAS showed good agreement on average values, but a correlation coefficient (r) of
only 0.47 on daily values (Mulik et al., 1989).
Badge-type personal samplers for NO2 also have been developed by Yanagisawa
and Nishimura (1982) (YN) and by Cadoff and Hodgeson (1983) (CH). Triethanolamine is
used as the active collecting medium in both samplers, and both use colorimetry as the
analytical method for detection of NO2". The samplers differ in that the YN device uses
TEA-coated on a cellulose filter, with a Teflon® filter as a diffusion barrier; whereas the CH
sampler uses TEA-coated on a glass fiber filter, with a polycarbonate filter as a diffusion
barrier. Detection limits are reported to be 0.07 ppm-h (Yanagisawa and Nishimura, 1982)
and 0.06 ppm-h (Cadoff and Hodgeson, 1983) for the YN and CH samplers, respectively.
Interferences from PAN and HNO2 are expected (Sickles and Michie, 1987); likewise, the
devices are sensitive to the speed of ambient air movement.
Palmes tubes have been developed for NO2 measurement and adapted to NO
measurement as described above. The device has been used for workplace and personal
exposure monitoring (Wallace and Ott, 1982) and for ambient air measurements (U.S.
Environmental Protection Agency, 1993c). A detection limit of 0.03 ppm-h can be achieved
if 1C is used to determine the extracted NO2" (Mulik and Williams, 1986). Adsorption of NO2
to the tube walls may raise this limit considerably (Miller, 1988), but this effect can be
counteracted by use of stainless steel tubes. The device is sensitive to temperature and wind
speed, and PAN and HNO2 are likely interferences (Sickles and Michie, 1987). In a
comparison with two commercially produced NO2 passive samplers, the Palmes tube showed
reasonable accuracy and precision at loadings of 1 to 80 ppm-h. However, the commercial
devices were designed for use at relatively high loadings; therefore, this comparison does not
support the use of Palmes tubes for ambient air monitoring.
3.5.3.4 Calibration Methods
Calibration of NO measurement methods is done using standard cylinders of NO in
nitrogen. Typical NO concentrations in such cylinders are 1 to 50 ppmv. Dilution of such
standards with clean air using mass flow controllers can accurately provide NO concentrations
in the ambient (i.e., 1 to 100 ppbv) range for calibration. Nitric oxide standards are available
3-126
-------
as SRMs from NIST and as commercially available Certified Reference Standards.
Commercially available certified NO standards have been shown to be stable and accurate for
the specified concentrations.
Standard cylinders of NO2 in nitrogen or air are sometimes used for NO2
calibration. These standards are commercially available and are readily diluted to parts-per-
billion-volume levels in the same manner as for NO standards. However, instability of the
NO2 levels in such standards has been reported, and caution must be used in relying on NO2
standards as the primary means of calibration.
Two calibration methods for NO2 are specified in the Code of Federal Regulations
(1987) for calibration of ambient NO2 measurements: (1) permeation tubes and (2) gas-phase
titration.
An NO2 permeation tube is an inert enclosure, generally of Teflon®, glass and
Teflon®, or stainless steel and Teflon®, that contains liquid NO2. As long as liquid NO2 is
present, NO2 will permeate through the Teflon® at a rate that depends on the temperature of
the tube. Maintaining the permeation tube at a constant temperature (i.e., ±0.1 °C) results in
permeation of NO2 at a constant rate. Dilution of the emitted NO2 with a flow of dry air or
N2 results in known low NO2 concentrations for calibration. Nitrogen dioxide permeation
tubes are supplied as SRMs by NIST, and tubes are commercially available with a wide range
of permeation rates. Permeation tubes are small, simple, reliable, and relatively inexpensive,
although constant temperature ovens and dilution systems are required to obtain good results.
Nitrogen dioxide permeation tubes are susceptible to moisture, and changes in permeation rate
or emission of other species (HNO3, HNO2, NO) may occur if they are not kept dry. As with
NO2 cylinder standards, the NO2 permeation tube requires care as a calibration method for
NO2.
Gas-phase titration uses the rapid reaction of NO with O3 to produce NO2 with
1:1 stoichiometry. In practice, excess NO generated from a standard cylinder containing
50 to 100 ppmv NO is reacted with O3 from a stable source. The resultant decrease in NO
concentration, usually measured on the NO channel of a CL NOX analyzer, equals the
concentration of NO2 generated. Varying amounts of NO2 can be produced by varying the
amount of O3.
3.6 Ozone Air Quality Models
To plan control strategies to achieve compliance with the NAAQS for O3 at some
future date, it is necessary to predict how O3 concentrations change in response to prescribed
changes in source emissions of precursor species (NOX and VOCs). This assessment requires
an air quality model, which in the case of O3 prediction is often called a photochemical air
quality model. The model, in effect, is used to determine the emission reductions needed to
achieve the O3 air quality standard. For at least a decade, EPA has offered guidelines on the
selection of air quality modeling techniques for use in SIP revisions, new source reviews, and
studies aimed at the prevention of significant deterioration of air quality.
It is worth noting the interrelated nature of O3 and other air quality issues. Ozone,
PM10, visibility, and acid deposition are all connected as a result of similar sources and
complex chemical mechanisms. Consequently, strategies for O3 abatement that involve
reductions of VOC and NOX emissions also will impact particulate matter, visibility, and acid
deposition.
3-127
-------
Ozone air quality models provide the ability to address "what if questions, such
as, what if emissions of VOCs or NOX are reduced? The model can be used as an experiment
that cannot be run in the atmosphere. Sensitivity questions can be asked, such as, how
important is emissions change A relative to emissions change B? or what is the effect of an
X% uncertainty in a particular chemical reaction rate constant on the predicted O3 levels?
Models are the ultimate integrators of our knowledge of the comprehensive
chemistry and physics of the atmosphere. As such, they are an indispensable tool for
understanding the complex interactions of transport, transformation, and removal in the
atmosphere. Models are useful in the design of field measurement programs and are essential
in the interpretation of data from such programs.
Models can be verified by the demonstration of agreement between observations
and predictions, but confirmation is inherently partial. Verification of mathematical models of
natural systems is always incomplete because complete information on the natural system is
never available. Furthermore, model results always include some degree of nonuniqueness
because model inputs and parameters are never precisely known. Ozone air quality model
applications are most reliable in the domain and conditions where model predictions have
been evaluated by extensive, valid data and the comparisons of observations and predictions
fall within accepted guidelines.
Historically, the primary measure of model performance has been degree of
agreement between observed and predicted O3 concentrations during simulated episodes,
although it now is recognized that comparisons of observations and predictions for other
compounds, such as organics and NOy components, are also important in assessing model
performance.
The purpose of Section 3.6 is to review briefly the main elements of O3 air quality
models, to describe several of the current models, to discuss the performance evaluation of
these models, and to present examples of the use of the models for determining VOC and
NOX control strategies.
3.6.1 Definitions, Description, and Uses
Air quality models are mathematical descriptions of the atmospheric transport,
diffusion, removal, and chemical reactions of pollutants. These models operate on sets of
input data that characterize the emissions, topography, and meteorology of a region and
produce outputs that describe air quality in that region. Mathematical models for
photochemical air pollution first were developed in the early 1970s and have been improved,
applied, and evaluated since that time. Much of the history of the field is described in
reviews by Tesche (1983), Seinfeld (1988), and Roth et al. (1990).
Photochemical air quality models include treatments of the important physical and
chemical processes that contribute to O3 formation in and downwind of urban areas.
In particular, such models contain a representation of the following phenomena (Roth et al.,
1990):
• Precursor emissions. The spatial and temporal characteristics of reactive
hydrocarbon, CO, and NOX emissions sources must be supplied as inputs to the
model. Hydrocarbon emissions generally are apportioned into groups (e.g.,
alkanes, alkenes, aromatics, etc.) according to the speciation requirements of the
chemical kinetic mechanism embedded in the model.
3-128
-------
• Pollutant transport. Once the O3 precursors are emitted into the atmosphere,
they are transported by the wind. When O3 is formed, it also is subject to
transport by the wind. Grid-based models require the preparation of three-
dimensional, time-varying fields of wind speed and direction. These values
must be specified for each grid cell. Cloud venting and mixing processes that
are important on the regional scale also can be included in the pollutant
transport description.
• Turbulent diffusion. Ozone and its precursors also are subject to turbulence-
related dispersion processes that take place on a subgrid scale. These turbulent
diffusion effects usually are represented in grid-based models by the so-called
gradient transport hypothesis, where the pollutant flux is assumed to be
proportional to the spatial gradient in the concentration field. The turbulent
diffusivities employed in the model are dependent on atmospheric stability and
other meteorological variables.
• Chemical reactions. Ozone results from chemical transformations involving
reactive organics and NOX (See Section 3.2). A chemical kinetics mechanism
representing the important reactions that occur in the atmosphere is employed to
estimate the net rate of change of each pollutant simulated by the model.
Description of chemical reactions requires actinic flux, cloud cover, temperature,
and relative humidity.
• Removal processes. Pollutants are removed from the atmosphere via
interactions with surfaces at the ground, so-called "dry deposition", and by
precipitation, called "wet deposition".
Guidelines issued by EPA (U.S. Environmental Protection Agency, 1986b) identify
two kinds of photochemical model: (1) the grid-based UAM is the recommended model for
modeling O3 over urban areas, and (2) the trajectory model EKMA is identified as an
acceptable approach. The 1990 CAAA (U.S. Congress, 1990) mandate that three-
dimensional, or grid-based, air quality models, such as UAM, be used in SIPs for
O3 nonattainment areas designated as extreme, severe, serious, or multistate moderate (U.S.
Environmental Protection Agency, 1991b).
3.6.1.1 Grid-Based Models
The basis for grid-based air quality models is the atmospheric diffusion equation
that expresses the conservation of mass of each pollutant in a turbulent fluid in which
chemical reactions occur (Seinfeld, 1986). The region to be modeled is bounded on the
bottom by the ground, on the top by some height that characterizes the maximum extent of
vertical mixing, and on the sides by east-west and north-south boundaries. The choice of the
size of the modeling domain will depend on the spatial extent of the O3 problem, including
the distribution of emissions in the region, the meteorological conditions, and, to some extent,
the computational resources available. This space then is subdivided into a three-dimensional
array of grid cells. The horizontal dimensions of each cell are usually a few kilometers for
urban applications up to tens of kilometers for regional applications. Some older grid-based
models assumed only a single, well-mixed vertical cell extending from the ground to the
inversion base; current models subdivide the region into layers. Vertical dimensions can vary,
depending on the number of vertical layers and the vertical extent of the region being
modeled. Increasing the vertical resolution in the computation should be accompanied by
increased vertical resolution of the physical parameters used. A compromise generally must
3-129
-------
be reached between the better vertical resolution afforded by the use of more vertical layers
and the associated increase in computing time. Although aerometric data, such as the vertical
temperature profile, which are needed to define the vertical structure of the atmosphere, are
generally lacking, it is still important to use enough vertical layers so that vertical transport
processes are represented accurately.
There are practical and theoretical limits to the minimum horizontal grid cell size.
Increasing the number of cells increases computing and data acquisition effort and costs.
In addition, the choice of the dimension of a grid cell implies that the input data information
about winds, turbulence, and emissions, for example, are resolved to that scale. The spatial
resolution of the concentrations predicted by a grid-based model corresponds to the size of the
grid cell. Thus, effects that have spatial scales smaller than those of the grid cell cannot be
resolved. Such effects include the depletion of O3 by reaction with NO near strong sources of
NOX like roadways and power plants. Ozone predictions are sensitive to the choice of grid
cell size. The use of a larger grid tends to smooth out VOC and NOX precursor
concentrations, affecting the computed chemical production of O3. Multigrid models, in
which a region with a finer grid resolution is embedded within a larger grid, are an approach
to obtain a better resolution of O3 formation processes in regions of intense source emissions
(Odman and Russell, 1991).
Jang (1992) has examined the sensitivity of O3 predictions to model grid resolution
in regional air quality models. A high-resolution version of the Regional Acid Deposition
Model (RADM) was used to simulate O3 formation over the northeastern United States at
different grid resolutions. The high-resolution version of RADM, with horizontal grid cell
sizes of 20, 40, and 80 km, was operated within the 80-km RADM domain. Coarser grid
sizes were found to result in lower resolved emission intensities of NOX and VOCs. Because
of the smearing effect of the large grid sizes, the coarser grid model tended to underpredict
the O3 highs in the areas downwind of cities and overpredict the O3 lows in the intense NOX
emissions areas. It was found that the impact of model grid resolution on the chemistry of
NOX is more important than that on the chemistry of VOCs, and that model grid resolution
has no significant impact on the total amount of odd oxygen (Ox = O3 + NO2) produced in the
models but has great impact on the interactions of chemistry and transport processes that
control the balance of Ox. As a result, the coarser grid model tends to predict higher O3 and
lower NO2 than does the finer grid model, and the coarser grid tends to transport Ox more in
the form of O3, whereas the finer grid model tends to transport the Ox more in the form of
NO2.
Uncertainties arise in photochemical modeling from the basic model components
(chemical mechanism and numerical techniques in solving the governing equations) and from
inputs to the simulations that reflect the particular episode (boundary and initial conditions,
emission inventory, wind field, and mixing depth). Sensitivity studies aim to determine the
range of uncertainty in model predictions corresponding to ranges of uncertainty in the basic
model components and input quantities. Such studies are valuable in pinpointing those
quantities to which model predictions are most sensitive and, therefore, in directing future
efforts in reducing the uncertainty in key parameters. These studies are also valuable in
assessing the sensitivity of future air quality changes to uncertainties in the base case episode.
It is not possible to state general, widely applicable levels of uncertainty for photochemical
model inputs and parameters. These will depend on the particular region being modeled, and,
in the case of meteorological and emissions inputs, may even depend on the time of day
3-130
-------
during the simulation. All model application exercises should include, to the extent possible,
an analysis of the uncertainties in model inputs and parameters.
Several grid-based photochemical air quality models have been developed to
simulate O3 production in urban areas or in larger regions. They differ primarily in their
treatment of specific atmospheric processes, such as chemistry, and in the numerical
procedures used to solve the governing system of equations. These models will be reviewed
in Section 3.6.3.
3.6.1.2 Trajectory Models
In the trajectory model approach, a hypothetical air parcel moves through the area
of interest along a path calculated from wind trajectories. Emissions are injected into the air
parcel and undergo vertical mixing and chemical transformations. The data requirements for
trajectory models include: (1) initial concentrations of all relevant pollutants and species;
(2) rates of emissions of VOC and NOX precursors into the parcel along its trajectory;
(3) meteorological characteristics, such as wind speed and direction, needed to define the path
of the air parcel through the region; (4) mixing depth; and (5) solar ultraviolet radiation.
The key assumption inherent in the trajectory model is that a hypothetical air
parcel maintains its integrity along the trajectory. Almost certainly, the parcel assumption
fails at night, when flows drift and the atmosphere stratifies; for hilly or mountainous terrain;
and under convergence conditions. Thus, the trajectory model concept does not apply in
many areas and under a variety of conditions (Liu and Seinfeld, 1975).
Trajectory models provide a dynamic description of atmospheric source-receptor
relationships that is simpler and less expensive to derive than that obtained from grid models.
Trajectory models are designed to study the photochemical production of O3 in the presence
of sources and vertical diffusion of pollutants; otherwise the meteorological processes are
highly simplified.
A simple trajectory model is used in EKMA (Dodge, 1977a). This modeling
approach relates the maximum level of O3 observed downwind of an urban area to the levels
of VOCs and NOX observed in the urban area. It is based on the use of a simple, one-cell
moving box model. As the box moves downwind, it encounters emissions of organics and
NOX that are assumed to be uniformly mixed within the box. The height of the box is
allowed to expand to account for the breakup of the nocturnal inversion layer. As the height
of the box increases, pollutants above the inversion layer are transported into the box. The
model is first used to generate a series of constant O3 lines (or isopleths) as depicted in
Figure 3-25. The isopleths show the downwind, peak 1-h O3 levels as a function of the
concentrations of VOCs and NOX for a hypothetical urban area. These isopleths were
generated by carrying out a large number of model simulations in which the initial
concentrations and anthropogenic emissions of VOCs and NOX were varied systematically,
whereas all other model inputs were held constant. When it was first conceived, EKMA
employed a very simple, highly empirical chemical mechanism and the isopleths generated
were for a hypothetical situation in Los Angeles. As understanding of the chemical processes
responsible for O3 formation increased, the EKMA model was updated to include more
complete representations of atmospheric chemistry. Although EKMA has employed the
CBM-IV mechanism, the same mechanism that is currently being used in several grid-based
models, the most recent version allows the input of any mechanism. The EKMA method is
3-131
-------
0.28
0.28
0.24
0.20
- 0.16
- 0.12
- 0.08
0.04
0.2
0.4 0.6 0.8 1.0 1.2 1.4 1.6
Nonmethane Hydrocarbon Concentration (ppm)
1.8 2.0
Figure 3-25. Example of Empirical Kinetic Modeling Approach diagram for high-oxidant
urban area.
Source: Derived from U.S. Environmental Protection Agency (1986a).
now used to generate city-specific isopleth diagrams using information on emissions,
transport, and dilution that are appropriate to the particular city being modeled.
City-specific O3 isopleths can be used to estimate the reduction in NMHC or NOX
levels needed to achieve the NAAQS for O3 in a specific urban area. The first step is to
determine the early-morning NMHC/NOX ratio for the urban area in question and the
maximum 1-h downwind O3 concentration. Both the NMHC/NOX ratio and the peak
O3 concentration are obtained from air monitoring data. These two values define a point on
the isopleth surface and, from this point, the percentage reductions in NMHC or NOX, or both,
needed to achieve the O3 NAAQS can be determined.
As examination of Figure 3-25 reveals, for an NMHC concentration of 0.6 ppmC,
for example, increasing NOX leads to increased O3 until NMHC/NOX ratios of about 5:1 to
6:1 are reached; further NOX increases, leading to lower NMHC/NOX ratios, inhibit
O3 formation. Thus, in this example, there is a "critical" ratio (in the range of 5:1 to 6:1) at
which the NOX effect on O3 changes direction. Besides this "critical" ratio, an "equal control"
NMHC/NOX ratio also exists, above which the reduction of NOX is more beneficial in terms of
O3 reduction than an equal percentage reduction in NMHC. This ratio, for the isopleths
3-132
-------
shown in Figure 3-25, is roughly 8:1 to 9:1 for low levels of control and as high as 20:1 for
the levels of control needed to reduce O3 to 0.12 ppm. Thus, for this particular case (Figure
3-25), the chemical mechanism modeling evidence suggests that NOX control will increase the
peak downwind O3 concentration at NMHC/NOX ratios of between 5:1 and 6:1 or lower; both
NOX control and NMHC control will be beneficial at somewhat higher ratios, with control of
NMHC being more effective; and, for ratios above 20:1, NOX control is relatively more
effective in reducing O3 to attain the O3 NAAQS.
The EKMA-based method for determining control strategies has some limitations,
the most serious of which is that predicted emissions reductions are critically dependent on
the initial NMHC/NOX ratio used in the calculations. This ratio cannot be determined with
any certainty because it is expected to be quite variable in time and space in an urban area.
Another limitation is that trajectory models have limited spatial and temporal scopes of
application. They are generally 1-day models, simulating only one cell at a time. Another
problem with the use of morning NMHC/NOX ratios is the failure to account for
photochemical evolution as urban emissions are carried downwind. As demonstrated in
simulations by Milford et al. (1989) and in smog chamber studies by Johnson and Quigley
(1989), an urban plume that is in the VOC-controlling regime (low NMHC/NOX ratio) near
city center can move increasingly into the NOx-controlling regime (high NMHC/NOX ratio) as
the air parcels age and move downwind. This progression occurs because NOX is
photochemically removed from an aging plume more rapidly than VOCs, causing the
VOC/NOX ratio to increase. As demonstrated by Milford et al. (1989), the implication of this
evolution is that different locations in a large urban area can show very different
O3 sensitivities to VOC and NOX changes. Because of this and other drawbacks, the 1990
CAAA (U.S. Congress, 1990) require that grid-based models be used in most
O3 nonattainment areas.
3.6.2 Model Components
3.6.2.1 Emissions Inventory
The spatial and temporal characteristics of VOC and NOX emissions must be
supplied as inputs to a photochemical air quality model. Emissions from area and point
sources are injected into ground-level grid cells, and emissions from large point sources are
injected into upper level cells. Total VOC emissions generally are apportioned into groups of
chemically similar species (e.g., alkanes, alkenes, aromatics, etc.) according to the
requirements of the chemical mechanism. This apportionment may be accomplished using
actual emission sampling and analysis or be based on studies of similar emission sources.
Recognition of potential undercounting in existing inventories has spurred efforts to improve
the accuracy of emissions inventories. In fact, at present, the emissions inventory is the most
rapidly changing component of photochemical models. It has been recognized that both
mobile and stationary source components have been highly uncertain and that there is
significant ongoing effort to improve the accuracy of emissions inventories.
Some emissions terminology is as follows (Tesche, 1992):
• Emissions data—the primary information used as input to emissions models.
• Emissions model—the integrated collection of calculational procedures, or
algorithms, properly encoded for computer-based computation.
• Emissions estimates—the output of emissions models; used as input to
photochemical models.
3-133
-------
• Emissions inventory—the aggregated set of emissions estimate files.
• Emissions model evaluation—the testing of a model's ability to produce
accurate emissions estimates over a range of source activity and
physicochemical and meteorological conditions.
Emissions input requirements for the UAM, for example, include:
• Spatial allocation of precursor emissions (VOCs, NOX, CO):
— Actual location of individual point sources;
— Spatial allocation by gridding surrogates;
— Assignment of surrogates to other categories.
• Stack parameters for point sources:
— Temperature, height, diameter, and exit velocity.
• Speciation of VOC emissions for CBM-IV mechanism:
— Region-specific speciation profiles;
— EPA default speciation profiles.
• Temporal allocation of precursor emissions:
— Operating schedules for individual point sources;
— Assignment of diurnal profiles for area and mobile sources.
The emissions inventory component of modeling is moving in the direction of the
use of emissions models rather than inventories. Emissions models are being developed for
the Lake Michigan Oxidant Study (LMOS), the San Joaquin Valley Air Quality Study
(SJVAQS), and the Atmospheric Utility Signatures, Predictions, and Experiments (AUSPEX),
designated as the SJVAQS/AUSPEX Regional Model Adaptation Project (SARMAP). The
consistency of existing inventories was improved in 1990 when EPA released the Emissions
Preprocessor System (EPS) as a component of the UAM (U.S. Environmental Protection
Agency, 1992b). The EPS was updated in 1992 to EPS Version 2 (EPS2). It is an emissions
model that considers spatial and temporal disaggregation factors, speciation data, and
meteorological data to convert daily emissions estimates for each point source and for area
source categories and mobile source emissions factors computed by the EPA MOBILES
model into hourly, gridded speciated estimates that are needed by a photochemical grid
model.
A step beyond the EPS is the Emissions Modeling System (EMS)1 (Tesche, 1992).
The EMS utilizes emissions estimation and information processing methods to provide
gridded, temporally resolved, and chemically speciated base-year emissions estimates for all
relevant source categories; to provide flexibility in forecasts of future-year emissions rates;
and to provide modular code design for use in module updating and replacement. The EMS
provides for easy substitution of alternative assumptions, theories, or input parameters (e.g.,
emissions factors, activity levels, spatial distributions) and facilitates sensitivity and
uncertainty testing.
As a result of a variety of independent studies, it recently has been determined that
urban VOC emissions inventories, particularly motor vehicle emissions, have been
significantly understated. These studies include tunnel studies (Pierson et al., 1990) and
comparisons of ambient and emission inventory VOC/NOX ratios (Fujita et al., 1992).
'The EMS has been renamed the GMEP (Geocoded Model of Emissions and Projections).
3-134
-------
3.6.2.2 Meteorological Input to Air Quality Models
Grid-based air quality models require, as input, the three-dimensional wind field
for the episode being simulated. This input is supplied by a so-called meteorological module.
Meteorological modules for constructing wind fields for air quality models fall into one of
four categories (Tesche, 1987; Kessler, 1988):
(1) Objective analysis procedures that interpolate observed surface and aloft wind
speed and direction data throughout the modeling domain;
(2) Diagnostic methods in which the mass continuity equation is solved to
determine the wind field;
(3) Dynamic, or prognostic, methods based on numerical solution of the governing
equations for mass, momentum, energy, and moisture conservation, along with
the thermodynamic state equations on a three-dimensional, finite-difference
mesh; or
(4) Hybrid methods that embody elements from both diagnostic and prognostic
approaches.
Objective Analysis
Objective wind-field analysis involves the interpolation and extrapolation of wind
speed and direction measurements (collected at a number of unequally spaced monitoring
stations) to grid points throughout the region (Goodin et al., 1980). For flat terrain settings
away from complex mesoscale forcings, this class of techniques may provide an adequate
method for estimating the wind field, provided that appropriate weighting and smoothing
functions are used (Haltiner, 1971). For complex terrain or coastal/lake environments,
however, it is tenuous to interpolate between and extrapolate from surface observational sites
except with an unusually dense monitoring network. In most cases, the routinely available
rawinsonde network sounding data are even more severely limited because of the large
distances (300 to 500 km) between sites and because soundings are made only every 12 h.
The limitations of even the best available data sets are most severe above the surface layer,
where upper level observations are less frequent and more expensive to obtain. It will remain
economically unfeasible to obtain sufficiently dense atmospheric observations to allow any
direct objective analysis scheme to provide the required detail and accuracy necessary for use
in advanced, high-resolution photochemical models.
Diagnostic Modeling
In diagnostic wind modeling, the kinematic details of the flow are estimated by
solving the mass conservation equation. Dynamic interactions such as turbulence production
and dissipation and the effects of pressure gradients are parameterized. Various diagnostic
wind models have been developed, many employing the concepts introduced by Sherman
(1978) and Yocke (1981).
In recent years, attempts have been made to combine the best features of objective
analysis and pure diagnostic wind modeling. The current release of EPA's UAM-IV includes
the Diagnostic Wind Model (DWM) as the suggested wind-field generator for this urban-scale
photochemical model. The DWM (U.S. Environmental Protection Agency, 1990c) is
representative of this class of hybrid objective-diagnostic models. The DWM combines the
features of the Complex Terrain Wind Model (CTWM) (Yocke, 1981) and the objective wind
interpolation code developed at the California Institute of Technology (Goodin et al., 1980).
In the DWM, a two-step procedure normally is followed. First, a "domain-scale" wind is
3-135
-------
estimated from available surface and upper-air synoptic data. This initial field consists of a
single wind vector (e.g., horizontal homogeneity) for each elevation. The domain-scale wind
is adjusted using procedures derived from the CTWM for the kinematic effects of terrain such
as lifting, blocking, and flow acceleration. Thermodynamically generated influences such as
mountain-valley winds are parameterized. This first step produces a horizontally varying field
of wind speed and direction for each vertical layer within the DWM modeling domain.
Typically, 10 to 12 vertical layers are used. In the second step, available hourly surface and
upper air measurements are combined objectively with the step 1 hourly diagnostic flow fields
to produce a resultant wind field that matches the observations at the monitoring points and
obeys the general constraints of topography in regions where data are absent. The DWM
contains a number of user-specified options whereby different final flow fields may be
produced, depending on selection of various smoothing and weighting parameters. The final
output of the DWM is a set of hourly averaged horizontal wind fields for each model layer.
Diagnostic models may invoke scaling algorithms that propagate the influence of
the surface-flow field into upper levels according to the local height of the inversion and the
Pasquill-Gifford-Turner stability category for the hour. Once the winds are created by DWM,
they must be "mapped" onto the photochemical model's vertical grid structure. This function
is normally accomplished in a two-step process. First, the DWM winds are interpolated onto
the photochemical model grid using simple linear interpolation. Second, the three-
dimensional divergence is computed in each grid cell and an iterative scheme is used to
minimize this divergence to a user-specified level. Typically, the output consists of
"nondivergent" x- and y-direction wind components for direct input to the photochemical
model.
Among the advantages of the diagnostic modeling approach are its intuitive appeal
and modest computing requirements. The method generally reproduces the observed wind
values at the monitoring locations and provides some information on terrain-induced airflows
in regions where local observations are absent. In addition, diagnostic model parameters for a
particular locale based on site-specific field measurements may be calibrated. However, there
are several disadvantages. Diagnostic models cannot represent complex mesoscale
circulations, unless these features are well represented by surface and aloft observations.
Often the vertical velocities produced by a diagnostic model are unrealistic and, in regions of
complex terrain, local horizontal flow velocities often may be an order of magnitude too high
(Tesche et al., 1987). Because the diagnostic model is not time-dependent, there is no
inherent dynamic consistency in the winds from one hour to the next. That is, calculation of
the flow field at 1200 hours, for example, is not influenced by the results of the 1100-hour
winds. This is a particular problem in applications involving important flow regimes, such as
land-sea breezes, mountain-valley winds, eddy circulations, and nocturnal valley jets, that take
several hours to develop and whose three-dimensional character is poorly characterized by
even the most intensive sampling networks. Finally, the inadequacy of the upper-air synoptic
data causes significant difficulty in the validation of the model wind fields.
Prognostic Modeling
In prognostic meteorological modeling, atmospheric fields are computed based on
numerical solutions of the coupled, nonlinear conservation equations of mass, momentum,
energy, and moisture. Derivations of these equations are presented extensively in the
literature (Haltiner, 1971; Pielke, 1984; Seinfeld, 1986; Cotton and Anthes, 1989). Many
prognostic models have been developed for computing mesoscale wind fields, as shown in the
3-136
-------
recent survey by Pielke (1989), and they have been applied to a variety of problems,
including the study of land-sea and land-lake circulations. Available prognostic models range
from relatively simple one-dimensional representations to complex three-dimensional codes.
Prognostic wind models are attractive because they explicitly address the various
physical processes governing atmospheric flows. Consequently, they have the potential for
describing a number of wind regimes that are particularly relevant to air pollution modeling,
such as flow reversal, daytime upslope flows, wind shear, and other mesoscale thermally
induced circulations. Drawbacks of prognostic models include the need to gather detailed
data for model performance testing and the large computational costs. Indeed, prognostic
models may require as much or more computer time than regional-scale photochemical
models. More intensive data sets are needed to evaluate prognostic models than for
diagnostic models, but this is not necessarily a disadvantage. Rather, it provides the modeler
and decision-maker with a far better basis for judging the adequacy of the model than can be
achieved with objective or diagnostic models.
Summaries of prognostic models available for use in air quality modeling are
presented extensively in the literature (e.g., Pielke, 1989; Benjamin and Seaman, 1985;
McNally, 1990; Stauffer et al., 1985; Stauffer and Seaman, 1990; Ulrickson, 1988; Wang and
Warner, 1988; and Yamada et al., 1989). From these reviews, two models stand out as
representing the present state-of-science in applications-oriented prognostic modeling. These
are the Mesoscale Model Versions 4 and 5 (MM4/MM5) developed by Pennsylvania State
University and the National Center for Atmospheric Research (NCAR) (Anthes and Warner,
1978; Anthes et al., 1987; Zhang et al., 1986; Seaman, 1990; Stauffer and Seaman, 1990), and
the Coast and Lake Regional Atmospheric Modeling System (CAL-RAMS) (Tripoli and
Cotton, 1982; Pielke, 1974, 1984, 1989; Lyons et al., 1991).
Three ongoing regional O3 modeling programs in the United States are using
prognostic models to drive regional O3 models. These include LMOS; SARMAP; and a
regional O3 modeling program in Southeast Michigan, Northern Ohio, and Southwest Ontario.
Part of EPA's long-range plan (in the Office of Research and Development [ORD]) for model
development is to construct a "third" generation modeling framework referred to as MODELS
3 (Dennis and Novak, 1992). This modeling system will consolidate all of the agency's
three-dimensional models. The current plan calls for meteorological inputs to the MODELS 3
system to be supplied by prognostic models. The MM4 model (the hydrostatic version of
MM5) is presently being examined by EPA for this purpose.
Activities are currently underway in LMOS to supply prognostic model fields to
EPA's ROM for use in simulating regional O3 distributions in four multiple-day O3 episodes
extensively monitored during the 1991 field program in the midwest. The EPA will be
utilizing ROM2.2 (version 2.2) with fields obtained from CAL-RAMS (Lyons et al., 1991) to
examine whether prognostic model output gives improved regional O3 estimates (Guinnup and
Possiel, 1991).
The SARMAP program is the modeling and data analysis component of a
multiyear collaboration between two projects, SJVAQS and AUSPEX. The major near-term
objective of SARMAP is to understand the processes that lead to high O3 concentrations in
the San Joaquin Valley of California. An overview of the regional meteorological and air
quality modeling approach of SARMAP is described by Tesche (1993). For SARMAP, the
MM5 model was chosen as the "platform" prognostic meteorological model because of its
broad application history; its demonstrated reliability on large domains, requiring spatially and
temporally varying boundary conditions; and its capability for four-dimensional data
3-137
-------
assimilation (FDDA)—needed for longer-range simulations (see Section 3.6.2.2). All of these
attributes are crucial to the success of mesoscale meteorological modeling.
Prognostic models are believed to provide a dynamically consistent, physically
realistic, three-dimensional representation of the wind and other meteorological variables at
scales of motion not resolvable by available observations. However, the meteorological
fields generated by a prognostic model do not always agree with observational data.
Numerical approximations, physical parameterizations, and initialization problems are among
the potential sources of error growth in model forecasts that can cause model solutions to
deviate from actual atmospheric behavior. Methods that have been devised over the past
20 years to mitigate these problems are described below.
"Post-processing" refers to methods whereby output fields from prognostic models
are selectively adjusted through a series of objective techniques with the aim of improving the
realism of the resultant fields. Examples of this procedure (sometimes referred to as objective
combination) are given by Cassmassi et al. (1991) in the Los Angeles Basin, Kessler and
Douglas (1989) in the South Central Coast Air Basin, and Moore et al. (1987) in the San
Joaquin Valley.
Ideally, a prognostic model should be initialized with spatially varying, three-
dimensional fields (i.e., wind, temperature, moisture) that represent the state of the
atmosphere at the initial simulation time. A prognostic model that is initialized with such
fields, however, can generate large nonmeteorological "waves" when the initial conditions do
not contain a dynamic balance consistent with the model formulation (Hoke and Anthes,
1976; Errico and Bates, 1988). The objective of an initialization procedure is to bring the
initial conditions into dynamic balance so that the model can integrate forward with a
minimum of noise and a maximum of accuracy (Haltiner and Williams, 1980). Dynamic
initialization makes use of a model's inherent adjustment mechanism to bring the wind and
temperature into balance prior to the initial simulation time. In this technique, a
"presimulation" integration of the model equations produces a set of dynamically balanced
initial conditions. By allowing the simulation to begin with a balanced initial state, this
technique reduces the generation of meteorological noise and thus improves the quality of the
simulation.
Four-Dimensional Data-Assimilation Techniques
Four-dimensional data assimilation refers to a class of procedures in which
observational data are used to enhance the quality of meteorological model predictions
(Harms et al., 1992). The most common use of FDDA today in applications-oriented models
is known as Newtonian relaxation, or simply as "nudging". With this method, model
estimates at a particular time interval are adjusted toward the observations by adding artificial
tendency terms to the governing prognostic equations. The objective of this method is to
improve prognostic model estimates through the use of valid, representative observational
data. As an example of this procedure, a linear term is added to the momentum equations to
"nudge" the dynamic calculation towards the observed state at each time step in regions
where data are available. The FDDA procedures may be thought of as the joint use of a
dynamic meteorological model in conjunction with observed data (or analysis fields based on
these data) in such a manner that the prognostic equations provide temporal continuity and
dynamic coupling of the hourly fields of monitored data (Seaman, 1990).
A recent example of the use of FDDA in regional-scale applications with the
MM4/RADM model is given by Stauffer and Seaman (1990). Attempts to apply FDDA in
3-138
-------
support of urban-scale photochemical grid modeling are described by Tesche et al. (1990b)
and McNally (1990) for the San Diego Air Basin and by Stauffer et al. (1993) for the Grand
Canyon region of Arizona. Currently, FDDA is being used in the CAL-RAMS simulations in
the LMOS program (Lyons et al., 1991) and in the MM5 simulations for SARMAP (Seaman,
1992).
3.6.2.3 Chemical Mechanisms
A chemical kinetic mechanism (a set of chemical reactions), representing the
important reactions that occur in the atmosphere, is used in an air quality model to estimate
the net rate of formation of each pollutant simulated as a function of time.
Various grid models employ different chemical mechanisms. Because so many
VOCs participate in atmospheric chemical reactions, chemical mechanisms that explicitly treat
each individual VOC component are too lengthy to be incorporated into three-dimensional
atmospheric models. "Lumped" mechanisms are therefore used (e.g., Lurmann et al., 1986;
Gery et al., 1989; Carter, 1990; Stockwell et al., 1990). These lumped mechanisms are highly
condensed and do not have the ability to follow explicit chemistry because of this lumping.
Lumped-molecule mechanisms group VOCs by chemical classes (alkanes, alkenes, aromatics,
etc.). Lumped-structure mechanisms group VOCs according to carbon structures within
molecules. In both cases, either a generalized (hypothetical) or surrogate (actual) species
represents all species within a class. Organic product and radical chemistry is limited to a
few generic compounds to represent all products; thus, chemistry after the first oxidation step
is overly uniform. Some mechanisms do not conserve carbon and nitrogen mass. Some
molecules do not easily "fit" the classes used in the reduced mechanisms. Because different
chemical mechanisms follow different approaches to lumping, and because the developers of
the mechanisms made different assumptions about how to represent chemical processes that
are not well understood, models can produce somewhat different results under similar
conditions (Dodge, 1989).
No single chemical mechanism is currently considered "best". Both UAM-IV and
ROM utilize the carbon-bond mechanism (CBM-IV), which, along with the SAPRC
(Statewide Air Pollution Research Center, University of California, Riverside) and RADM
mechanisms, is considered to represent the state of the science (Tesche et al., 1993; National
Research Council, 1991). Agreement among mechanisms is better for O3 than for other
secondary pollutants (Dodge, 1989, 1990; National Research Council, 1991), raising concern
that the mechanisms may suffer from compensating errors. These mechanisms are at least
5 years old and often are tested on much older smog chamber data.
The chemical mechanisms used in existing photochemical O3 models contain
uncertainties that may limit the accuracy of the model predictions. The reactions that are
included in these mechanisms generally fall into one of three categories.
(1) Reactions for which the magnitude of their rate constants and their product
distribution is well known. These include mostly the inorganic reactions and
those for the simple carbonyls.
(2) Reactions with known rate constants and known products but with uncertain
product yields. These are mostly organic reactions, and the actual product
yields assumed may vary among mechanisms.
(3) Reactions with known rate constants but unknown products. Each mechanism
assumes its own set of products for reactions in this class. This class includes
aromatic oxidation reactions.
3-139
-------
Most inorganic gas-phase processes are understood. Regarding classes of VOCs
the following general comments can be made:
• The w-alkanes comprise approximately one-half of the major carbon emissions in
urban areas. Reaction rates are relatively slow. The only important reaction is
with the hydroxyl radical. For alkanes C4 or below, the chemistry is well
understood and the reaction rates are slow. For C5 and higher alkanes the
situation is more complex because few reaction products have been found.
• Branched-chain alkanes have rates of reaction that are highly dependent on
structure. Rate constants have been measured for only a few of the branched
alkanes, and reaction products for this class of organics are not well
characterized.
• Alkenes are reactive with OH, O3, and the NO3 radical. Most rate constants of
these reactions are known. Alkenes make up <15% of the emitted carbon and
constitute about 25% of the hydrocarbon reactions in urban areas. Ozone
reaction products are not well characterized, and the mechanisms are poorly
understood. Mechanisms for the NO3 radical are also uncertain.
• Aromatics constitute about 15 to 20% of the carbon compounds emitted and
25% of the hydrocarbons reacting in urban areas. Aromatics have been studied
frequently, but only a few reaction products have been well characterized.
Aromatics act as strong NOX sinks under low NOX conditions.
Mechanisms used in photochemical air quality models thus have uncertainties,
largely attributable to a lack of fundamental data on products and product yields. The
missing information necessitates that assumptions be made. Current mechanisms provide
acceptable overall simulation of O3 generation in smog chamber experiments. Specific VOCs
may, however, be simulated poorly, and products other than O3 may not be simulated
accurately. Existing mechanisms are mostly applicable to single-day, high NOX conditions
because those are the conditions of almost all smog chamber experiments. Low NOX
condition simulations are verified less thoroughly. Fundamental kinetic data are needed on
the photooxidation of aromatics, higher alkanes, and higher alkenes to fill in areas of
uncertainty in current mechanisms. Whereas these uncertainties are important and require
continued research to remove, the uncertainties are likely not such that general conclusions
about the relative roles of hydrocarbons and NOX in O3 formation will be changed by new
data.
3.6.2.4 Deposition Processes
Species are removed from the atmosphere by interaction with ground-level
surfaces, so-called dry deposition, and by absorption into airborne water droplets followed by
transport of the water droplets, wet deposition. Dry deposition is an important removal
process for ozone and other species on both the urban and regional scales and is included in
all urban and regional scale models as a contribution to the ground-level flux of pollutants.
Wet deposition is a key removal process for gaseous species on the regional scale and is
included in regional scale acid deposition models. Urban-scale photochemical models
generally have not included a treatment of wet deposition as O3 episodes do not occur during
periods of significant clouds or rain.
Dry Deposition
3-140
-------
It is generally impractical to simulate, in explicit detail, the complex of multiple
physical and chemical pathways that result in dry deposition to individual surface elements.
Because of this, the usual practice has been to adopt simple parameterizations that consolidate
the multitude of complex processes. For example, it generally is assumed that the dry
deposition flux is proportional to the local pollutant concentration [at a known reference
height (zr), typically 10 m], resulting in the expression F = ~vdC, where F represents the dry
deposition flux (the amount of pollutant depositing to a unit surface area per unit time), and
C is the local pollutant concentration at the reference height. The proportionality constant, vd,
has units of length per unit time and is known as the deposition velocity.
It is customary to interpret the dry deposition process in terms of the electrical
resistance analogy, where transport of material to the surface is assumed to be governed by
three resistances in series: (1) the aerodynamic resistance (ra), (2) the quasi-laminar layer
resistance (rb), and (3) the surface or canopy resistance (rs) (Wu et al., 1992). The
aerodynamic resistance characterizes the turbulent transport through the atmosphere from
reference height zr down to a thin layer of stagnant air very near the surface. The molecular-
scale diffusive transport across the thin quasi-laminar sublayer near the surface is
characterized by rb. The chemical interaction between the surface and the pollutant of interest
once the gas molecules have reached the surface is characterized by rc. The total resistance
(rt) is the sum of the three individual resistances, and is, by definition, the inverse of the
deposition velocity, l/vd = rt = ra + rb + rs. Note that the deposition velocity is small when
any one of the resistances is large. Hence, either meteorological factors or the chemical
interactions on the surface can govern the rate of dry deposition.
Dry deposition velocities of HNO3 and SO2 are typically ~2 cm s"1, and those of
O3 and PAN are generally ~0.5 and ~1 cm s"1, respectively (Dolske and Gatz, 1985; Colbeck
and Harrison, 1985; Huebert and Robert, 1985; Shepson et al., 1992). With a 1 km-deep
inversion or boundary layer, the time scale for dry deposition is on the order of 1 day for a
deposition velocity of 1 cm s"1. Dry deposition is important for those chemicals with high or
fairly high deposition velocities and long or fairly long lifetimes (>10 days) due to photolysis
and chemical reaction (for example, HNO3, SO2, and H2O2, as well as O3 and PAN).
A number of researchers have reviewed the deposition literature and provided
summaries of deposition velocity data. The rank ordering of deposition velocity values
among pollutant species based on several such studies is summarized as follows:
McRae and Russell (1984):
HNO3 > SO2 > NO2 - O3 > PAN > NO;
Derwent and Hov (1988):
HNO3 > SO2 = O3 > NO2 > PAN;
McRae et al. (1982b):
O3 > NO2 > PAN > NO > CO; and
Chang et al. (1987):
HNO3 > H2O2 > NH3 > HCHO > O3 = SO2 = NO2 = NO > RCHO.
There is general agreement that HNO3 is removed at the highest observed rates,
which is consistent with the relative deposition rates observed by Huebert and Robert (1985).
Most of the surveys are roughly consistent with the relative deposition velocity ordering seen
in the experiments of Hill and Chamberlain (1976): diffusion-limited acids > SO2 > NO2 ~
O3 > PAN > NO > CO. This suggests surface resistance values should be ordered
approximately as CO > NO > PAN > O3 « NO2 > SO2 > HNO3 = 0. However, there is still a
substantial range of variability in reported deposition velocities. For example, McKeen et al.
3-141
-------
(1991) calculated dry deposition velocities for HNO3, O3, and PAN of 10, 0.5, and 0.3 cm s"1,
respectively. Note that a large deposition velocity for HNO3 will limit the lifetime of HNO3
relative to O3 in photochemically aged air.
There are a significant number of other gases for which there are no surface
resistance data and for which values must be estimated using expert judgment. The values
should be consistent with the existing experimental values for vegetative surfaces and should
preserve the apparent rank ordering among the pollutant species (discussed above). For O3,
surface resistance values by land-use type and season recommended by Sheih et al. (1986)
and Wesely (1988) are appropriate. For NO, NO2, NH3, H2O2, HCHO, and CH3CHO, the
surface resistance values for each land use can be estimated from that for SO2 (Wesely,
1988), except different proportionality factors should be used for NO and NO2.
The treatment of dry deposition is perhaps the most primitive of the scientific
modules in photochemical air quality models. Knowledge of deposition rates is limited, and
uncertainties in deposition velocities are high. For travel times of one to several days, the
quantities of pollutants that are predicted to be removed by dry deposition can be substantial
for those species with appreciable deposition velocities. Setting all species deposition
velocities to zero in a model provides an indication of the importance of dry deposition
relative to other processes influencing pollutant dynamics. Further effort to describe the
dynamics of deposition are needed, together with evaluation against available data that can be
used to test deposition modules.
Wet Deposition
Wet deposition refers to the removal of gases and particles from the atmosphere by
precipitation events, through incorporation of gases and particles into rain, cloud, and fog
water followed by precipitation at the earth's surface. Removal of gases and particles during
snow falls is also wet deposition. Wet removal of gases arises from equilibrium partitioning
of the chemical between the gas and aqueous phases (Bidleman, 1988; Mackay, 1991). This
partitioning can be defined by means of a washout ratio, Wg, with Wg = [C]rain/[C]air,
where [C]rain and [C]air are the concentrations of the chemical in the aqueous and gas
phases, respectively. Because Wg is the inverse of the air/water partition coefficient, Kaw,
then Wg = RT/H, where R is the gas constant, T is the temperature, and H is the Henry's
Law constant (Mackay, 1991).
Particles and particle-associated chemicals are efficiently removed from the
atmosphere by precipitation events, and the washout ratios for particles, Wp, are typically in
the range 104 to 106 (Eisenreich et al., 1981; Bidleman, 1988). Wet deposition is important
for particles (and particle-associated chemicals) and for those gas-phase compounds with
washout ratios of Wg > 104. Examples of such gaseous chemicals are HNO3, H2O2, phenol,
and cresols, all of which are highly soluble in water. Formaldehyde is present in the aqueous
phase as the glycol, H2C(OH)2, and has an effective washout ratio of 7 x 103 at 298 K
(Betterton and Hoffmann, 1988; Zhou and Mopper, 1990). Note that the importance of wet
deposition may depend on whether the chemical is present in the gas phase or is particle-
associated. For example, the gas-phase alkanes have low values of Wg and are inefficiently
removed by wet deposition, whereas the particle-associated alkanes are efficiently removed by
wet deposition (Bidleman, 1988), through removal of the host particles.
3.6.2.5 Boundary and Initial Conditions
3-142
-------
When a grid-based photochemical model is applied to simulate a past pollution
episode, it is necessary to specify the concentration fields of all the species computed by the
model at the beginning of the simulation. These concentration fields are called the initial
conditions. Throughout the simulation, it is necessary to specify the species concentrations,
called the boundary conditions, in the air entering the three-dimensional geographic domain.
Three general approaches for specifying boundary conditions for urban-scale
applications can be identified: (1) use the output from a regional-scale photochemical model,
(2) use objective or interpolative techniques with ambient observational data, or (3) use
default regional background values and expand the area that is modeled for urban areas
sufficiently isolated from significant upwind sources.
In the ideal case, observed data would provide information about the
concentrations for all the predicted species at the model's boundaries. An alternative
approach is to use regional models to set boundary and initial conditions. This is, in fact,
preferred when changes in these conditions are to be forecast. In any event, simulation
studies should use boundaries that are far enough from the major source areas of the region
that concentrations approaching regional values can be used for the upwind boundary
conditions. Boundary conditions at the top of the area that is being modeled should use
measurements taken from aloft whenever they are available. Regional background values
often are used in lieu of measurements. An emerging technique for specifying boundary
conditions is the use of a nested grid, in which concentrations from a larger, coarse grid are
used as boundary conditions for a smaller, nested grid with finer resolution. This technique
reduces computational requirements compared to those of a single-size, fine-resolution grid.
Initial conditions are determined mainly with ambient measurements, either from
routinely collected data or from special studies. Where spatial coverage with data is sparse,
interpolation can be used to distribute the surface ambient measurements. Because few
measurements of air quality data are made aloft, it generally is assumed that species
concentrations are initially uniform in the mixed layer and above it. To ensure that the initial
conditions do not dominate the performance statistics, model performance should not be
assessed until the effects of the initial conditions have been swept out of the grid.
3.6.2.6 Numerical Methods
The core of a grid-based O3 air quality model is the numerical solution of the
three-dimensional atmospheric diffusion equation (McRae et al., 1982c). The central
numerical schemes involve horizontal advection and simultaneous vertical mixing, advection,
and chemistry. A possible source of model inaccuracy is the numerical method used to solve
the governing equations. The solution of chemical kinetics is generally the most
computationally intensive step in O3 air quality models. To compute the rate of chemical
reaction one essentially must solve a system of stiff nonlinear ordinary differential equations.
The desirable characteristics of the integration routine are speed and stability, at a certain
prescribed level of accuracy. The chemistry integration routines used in several ozone air
quality models are based on the implicit, hybrid, exponential scheme developed by Young and
Boris (1977). The integration is stable, efficient, and sufficiently accurate. It has been
concluded from several studies that the numerical solution of the vertical/chemical portion of
the model is less likely to be a source of O3 prediction inaccuracy (Odman et al., 1992) than
the horizontal advection, numerical method (McRae et al., 1982; Chock, 1985, 1991; Dabdub
and Seinfeld, 1994). Although the horizontal transport computations typically consume only a
small fraction of the total computer time, it is well known that numerical diffusion and
3-143
-------
dispersion degrade the computed solution and that available methods differ greatly in their
numerical performance in this regard (Chock, 1985, 1991; Dabdub and Seinfeld, 1994).
Continued work on optimizing the numerical methods used in O3 air quality models is
necessary.
3.6.3 Urban and Regional Ozone Air Quality Models
Several grid-based models have been widely used to evaluate O3 and acid
deposition control strategies.
• The Urban Airshed Model, developed by Systems Applications, Inc., has been,
and continues to be, applied to urban areas throughout the country. It is
described in Section 3.6.3.1. The current EPA-approved version is UAM-IV.
The UAM-V, which has been developed for LMOS, is a nested regional-scale
model.
• The California Institute of Technology (CIT) model has been applied to
California's South Coast Air Basin (McRae et al., 1982a,b; McRae and Seinfeld,
1983; Milford et al., 1989; Harley et al., 1993).
• The ROM, developed by EPA, has been applied to the northeastern and
southeastern United States (Schere and Wayland, 1989a,b). It is described in
Section 3.6.3.2.
• The Acid Deposition and Oxidant Model (ADOM) was developed by ENSR
Consulting and Engineering for the Ontario Ministry of the Environment and
Environment Canada (Venkatram et al., 1988) and the German
Umweltbundesamt. Its primary application has been to acidic deposition.
• The RADM was developed by NCAR and the State University of New York for
NAPAP. The primary objective of RADM applications is the calculation of
changes in sulfur and nitrogen deposition over the eastern United States and
southeastern Canada, resulting from changes in emissions (National Acid
Precipitation Assessment Program, 1989). See Section 3.6.3.3 for a description
of RADM.
A summary of the major applications of the above air quality models, including the
Sulfur Transport Eulerian Model (STEM-II), is presented in Table 3-21. All of the models
are based nominally on a 1-h time resolution. The horizontal spatial resolutions vary from
5 to 120 km. Typical spatial resolutions used in past model applications are summarized in
Table 3-22. It is important to note that the spatial scale at which a model is applied is
governed by the manner in which physical processes are treated and the spatial scale of the
inputs. The regional models can have a vertical resolution on the order of 10 to 15 layers
extending up to 6 to 10 km in order to treat vertical redistribution of species above the
planetary boundary layer. This increased vertical resolution often comes at the expense of
decreased horizontal resolution. Urban models typically have two to five layers extending up
to 1,000 to 2,000 m. The treatment of meteorological fields by the six models is summarized
in Table 3-23. Generally, the treatment of meteorology is separate from the air quality model
itself, and models can employ wind fields prepared by different approaches as long as
consistent assumptions, such as nondivergent wind field, are employed in each model. The
regional models, ROM, RADM, ADOM, and STEM-II, address the vertical redistribution of
pollutants resulting from the presence of cumulus clouds.
3-144
-------
Table 3-21. Grid-Based Urban and Regional Air Pollution Models:
Overview of Three-Dimensional Air Quality Models3
Model
Major Applications
Major References for
Model Formulation
Selected References for Model Performance
Evaluation and Application
UAM
CIT
Urban and nonurban areas in the
United States and Europe
Los Angeles Basin
Reynolds et al. (1973, 1974, 1979)
Tesche et al. (1992)
U.S. Environmental Protection Agency
(1990a,b,c; 1992b)
Scheffe and Morris (1993)
McRae et al. (1982a)
Tesche et al. (1993)
McRae and Seinfeld (1983)
Russell et al. (1988a,b)
Harley et al. (1993)
ROM
Eastern United States
(E of 99° W longitude)
Lamb (1983)
Schere and Wayland (1989a,b)
Meyer et al. (1991b)
CO
en
RADM
Eastern North America
Chang et al. (1987)
Middleton et al. (1988, 1993)
Middleton and Chang (1990)
Dennis et al. (1993a)
Cohn and Dennis (1994)
ADOM
STEM-II
Eastern North America and
Northern Europe
Venkatram et al. (1988)
Philadelphia area, Kentucky, and Carmichael et al. (1986)
northeastern United States, central
Japan
Venkatram et al. (1988)
Macdonald et al. (1993)
Karamchandani and Venkatram (1992)
Carmichael et al. (1991)
Saylor et al. (1991)
"See Appendix A for abbreviations and acronyms.
-------
Table 3-22. Grid-based Urban and Regional Air Pollution Models: Treatment of Emissions and Spatial
Resolution3
Model
Emitted Species
Point-Source Emissions
Area-Source Emissions
Vertical Resolution
UAM SO2, sulfate, NO, NO2, CO, Released into grid cell in layer
NH3, and 8 classes of ROG and corresponding to plume rise in
PM (4 size classes) UAM; treated with a reactive
plume model in PARIS
CIT SO2, sulfate, NO, NO2, CO, Treated with a plume model
NH3, and 6 classes of ROG and with simple NOX and O3
PM (4 size classes) chemistry
Grid-average with resolution
ranging from 4 km x 4 km to
10 km x 10 km in past
applications
Typically, 5-6 layers up to about 1.5 km
Grid-average with 5 km x 5 km Five layers up to about 1.5 km
resolution in past applications
ROM CO, NO, NO2, and 8 classes
of ROG
Released into grid cell in layer
corresponding to plume rise
Grid-average with 18.5 km x
18.5 km resolution in present
applications
Three layers up to about 4 km
CO
O)
RADM SO2, sulfate, NO, NO2, CO,
NH3, and 12 classes of ROG
Released into grid cell in layer
corresponding to plume rise
Grid-average with 80 km
80 km resolution in past
applications
Fifteen layers up to about 16 km
ADOM SO2, sulfate, NO, NO2, NH3, Released into grid cell in layer Grid-average with resolution Twelve layers up to about 10 km
and 8 classes of ROG and PM corresponding to plume rise ranging from 60 km x 60 km to
about 120 km x 120 km in past
applications
STEM-II SO2, sulfate, NO, NO2, NH3,
and 8 classes of ROG
Released into grid cell in layer Grid-average with resolution Ten to 14 layers up to about 6 km
corresponding to plume rise ranging from 10 km x 10 km to
56 km x 56 km in past
applications
aSee Appendix A for abbreviations and acronyms.
-------
Table 3-23. Grid-based Urban and Regional Air Pollution Models:
Treatment of Meteorological Fields, Transport, and Dispersion3
Model
Meteorology
Transport
Turbulent Diffusion
UAM
CIT
ROM
RADM
ADOM
STEM-II
Constructed through data interpolation
or calculated with land-sea breeze or
complex terrain wind model.
Constructed through data interpolation
with diagnostic wind model.
Constructed through data interpolation.
Calculated with Community Climate
Model (CCM) and MM4.
Constructed through data interpolation
in combination with prognostic planetary
boundary-layer model.
Calculated with dynamic wind model
(MASS) or constructed through data
interpolation.
3-D wind field. Finite difference
numerical technique.
3-D wind field. Finite element
numerical technique.
3-D wind field with vertical transport
through cumulus clouds. Finite
difference numerical technique.
3-D wind field with vertical transport
through cumulus clouds. Finite
difference numerical technique.
3-D wind field with vertical transport
through cumulus clouds. Cubic spline
numerical technique.
3-D wind field with vertical transport
through clouds. Finite element
numerical technique.
Vertical turbulent diffusion function of atmospheric
stability and friction velocity. Constant horizontal
turbulent diffusion coefficient.
Vertical turbulent diffusion function of atmospheric
stability and friction velocity. Horizontal turbulent
diffusion function of mixing height and convective
velocity scale.
Vertical turbulent diffusion function of atmospheric
stability. Horizontal turbulent diffusion function of
atmospheric stability, convective cloud cover and
velocity scale, and the depths of the boundary layer
and clouds.
Vertical turbulent diffusion function of atmospheric
stability and wind shear. No horizontal turbulent
diffusion.
Vertical turbulent diffusion calculated from planetary
boundary layer model. No horizontal turbulent
diffusion.
Vertical turbulent diffusion function of atmospheric
stability and surface roughness. Horizontal turbulent
diffusion proportional to vertical turbulent diffusion.
"See Appendix A for abbreviations and acronyms.
-------
Table 3-24 summarizes the gas-phase chemical mechanisms incorporated into the six models.
Generally three chemical mechanisms are used in the models: (1) CBM-IV used in ROM and
UAM; (2) versions of the SAPRC mechanism used in ADOM, STEM-II, and CIT; and (3) the
RADM mechanism. Of the three chemical mechanisms, RADM is the largest and CBM-IV is
the smallest. Aqueous-phase chemistry is currently treated only in the regional models.
Cloud processes are treated in the three regional models, RADM, ADOM, and STEM-II
(Table 3-25). Cumulus venting and solar attenuation are treated in ROM. Layer 3 depths
also are influenced by cloud thickness. At present, only RADM, ADOM and STEM-II treat
wet deposition. The treatment of dry deposition in the models also is summarized in Table 3-
25.
Regional-scale modeling is an important contributor to the development of
boundary conditions for urban-scale models. In recent years, regional-scale modeling has
been receiving increased attention as the need for addressing interlinked air quality problems
at broader scales is increasing. At the expanded spatial and temporal scales of regional-scale
models, the simulation of certain dynamic processes becomes more critical. For example, in
regional-scale models the treatment of biogenic VOC emissions and removal by dry and wet
deposition generally require greater attention and accuracy than at the urban scale. On the
other hand, the exact mechanistic details of the oxidation of some highly reactive VOCs may
be somewhat less important.
More detailed descriptions now will be presented for UAM, ROM, and RADM.
The UAM is described, as it is specified officially by EPA, as a grid-based model for urban-
scale O3 control strategy determination. The regional-scale O3 model, ROM, is being used by
EPA to evaluate O3 control measures for the eastern United States and to provide boundary
conditions for urban area simulations using UAM. Representative of a comprehensive state-
of-the-science O3/acid deposition model, RADM has been used to evaluate combined O3 and
acid deposition abatement strategies for the northeastern United States and Canada.
The EPA is embarking on a project to produce the next generation of
photochemical models, termed MODELS 3 (Dennis et al., 1993b). This group of models will
be flexible (scalable grid and domain), will be modular (modules with interchangeable data
structure), will have uniform input/output across subsystems, and will contain advanced
analysis and visualization features. The models will be designed to take advantage of the
latest advances in computer architecture and software.
3.6.3.1 The Urban Airshed Model
The UAM is the most widely applied and broadly tested grid-based photochemical
air quality model. The model is described in a number of sources, including a multi-volume
series of documents issued by the U.S. Environmental Protection Agency (1990a,b,c; 1992b)
and a comprehensive evaluation by Tesche et al. (1993). Current versions include provisions
enabling the user to model transport and dispersion within both the mixed and inversion
layers. The computer codes have been structured to allow inclusion of up to 10 vertical
layers of cells and any number of cells horizontally.
The original UAM developed by Reynolds et al. (1973) simulated the dynamic
behavior of six pollutants: (1) reactive and (2) unreactive hydrocarbons, (3) NO, (4) NO2, (5)
O3, and (6) CO. Since 1977, the UAM has employed various versions of the CBM.
Currently, the model utilizes the CBM-IV Mechanism (Gery et al., 1988, 1989), which treats
3-148
-------
Table 3-24. Grid-based Urban and Regional Air Pollution Models: Treatment of Chemical Processes3
Model
Gas-Phase Chemistry
Aqueous- Phase Chemistry
CO
UAM
CIT
ROM
RADM
ADOM
STEM-II
Eighty-seven reactions among 36 species including NOX, O3,
ROG, and SO2 (CBM-IV) (Gery et al., 1988, 1989)
One hundred and twelve reactions among 53 species
including NOX, O3, ROG, and SO2 (Lurmann et al., 1986)
Eighty-seven reactions among 36 species including NOX, O3,
ROG, and SO2 (CBM-IV)
One hundred and fifty-seven reactions among
59 species including NOX , O3, ROG, and SO2
(Stockwell et al., 1990)
One hundred and twelve reactions among 53 species
including NOX, O3, ROG, and SO2 (Lurmann et al., 1986)
One hundred and twelve reactions among 53 species
including NOX, O3, ROG, and SO2 (Lurmann et al., 1986)
No treatment of aqueous-phase chemistry
No treatment of aqueous-phase chemistry
No treatment of aqueous-phase chemistry
Forty-two equilibria and five reactions for SO2 oxidation
Fourteen equilibria and five reactions for SO2 oxidation
Twenty-six equilibria and about 30 reactions for SO2 and NOX
oxidation, radical chemistry, and transition metal chemistry
CD
aSee Appendix A for abbreviations and acronyms.
-------
Table 3-25. Grid-based Urban and Regional Air Pollution Models: Treatment of Cloud and Deposition Processes3
Model
Cloud Processes
Wet Deposition
Dry Deposition
CO
A
Oi
O
UAM
CIT
ROM
RADM
ADOM
STEM-II
No treatment of cloud processes.
No treatment of cloud processes.
No treatment of cloud processes, except
vertical transport treatment.
No treatment of wet deposition.
No treatment of wet deposition.
No treatment of wet deposition.
Treatment of precipitating cumulus clouds, Calculated from precipitation rate and
precipitating stratus clouds, and fair- cloud average chemical composition;
weather cumulus clouds, based on no below-cloud scavenging.
precipitation amount, temperature, and
relative humidity vertical profiles. Use
of cloud-averaged properties for aqueous
chemistry.
Dry deposition velocity approach; function of wind speed,
friction velocity, land type, and species.
Dry deposition velocity approach; function of atmospheric
stability, wind speed, land type, and species.
Resistance transfer approach; function of land type, wind
speed, atmospheric stability, and species.
Resistance transfer approach; function of atmospheric
stability, wind speed, season, land type, insolation,
surface wetness, and species.
Treatment of cumulus clouds and stratus
clouds, based on precipitation amount
(for stratus clouds), temperature, and
relative humidity vertical profiles.
Vertical resolution for cloud chemistry.
Treatment of clouds with the Advanced
Scavenging Module based on cloud-base
height, precipitation rate, and surface
temperature.
Calculated from precipitation rate and Resistance transfer approach; function of atmospheric
vertically weighted cloud average chemical stability, wind speed, land type, season, insolation, and
composition, below-cloud scavenging species.
included.
Calculated with the Advanced
Scavenging Module. Treats cloud
water, rain water, and snow;
below-cloud scavenging included.
Resistance transfer approach; function of atmospheric
stability, land type, wind speed, and species.
aSee Appendix A for abbreviations and acronyms.
-------
36 reacting species. Reactive organic compounds include alkanes, alkenes, aromatics, and
aldehydes, and nitrogen-bearing species include HNO2, HNO3, and PAN.
Under development at the time of writing of the present document is version V of
UAM (Morris et al., 1991, 1992). This version (UAM-V) contains the following features:
the ability to treat two-way interactive nested grids; the use of state-of-the-science treatment
of atmospheric, meteorological, and chemical processes; the treatment of subgrid-scale plume
processes using a plume-in-grid algorithm; and the use of structured programming techniques
to take advantage of computational speed enhancement opportunities offered by the current
and next generation of computers.
3.6.3.2 The Regional Oxidant Model
The ROM was designed to simulate most of the important chemical and physical
processes that are responsible for the photochemical production of O3 over regional domains
and for episodes of up to 15 days in duration. These processes include horizontal transport;
atmospheric chemistry and subgrid-scale chemical processes; nighttime wind shear and
turbulence associated with the low-level nocturnal jet; the effects of cumulus clouds on
vertical mass transport and photochemical reaction rates; mesoscale vertical motions induced
by terrain and the large-scale flow; terrain effects on advection, diffusion, and deposition;
emissions of natural and anthropogenic O3 precursors; and dry deposition. The processes are
simulated mathematically in a three-dimensional Eulerian model with three vertical layers,
including the boundary layer and the capping inversion or cloud layer. The ROM
geographical domains are summarized in Table 3-26 and illustrated in Figure 3-26.
Meteorological data are used to model objectively both regional winds and
diffusion. The three model layers of ROM are prognostic (predictive) and are free to expand
and contract locally in response to changes in the physical processes occurring within the
layers. During an entire simulation period, horizontal advection and diffusion and gas-phase
chemistry are modeled in the three layers. Predictions from Layer 1 are used as surrogates
for surface concentrations. Layers 1 and 2 model the depth of the well-mixed layer during
the day. Some special features of Layer 1 include the modeling of the substantial wind shear
that can exist in the lowest few hundred meters above ground in local areas where strong
winds exist and the surface heat flux is weak, the thermal internal boundary layer that often
exists over large lakes or near sea coasts, and deposition onto terrain features that protrude
above the layer. At night, Layer 2 represents what remains of the daytime mixed layer.
As stable layers form near the ground and suppress turbulent vertical mixing, a nocturnal jet
forms above the stable layer and can transport aged pollutant products and reactants
considerable distances. At night, emissions from tall stacks and warm cities are injected
directly into Layers 1 and 2. Surface emissions are specified as a mass flux through the
bottom of Layer 1. During the day, the top model layer, Layer 3, represents the synoptic-
scale subsidence inversion characteristic of high O3-concentration periods; the base of Layer 3
is typically 1 to 2 km above the ground. Relatively clean tropospheric air is assumed to exist
above Layer 3 at all times, and stratospheric intrusion of O3 is assumed to be negligible. If
cumulus clouds are present, an upward flux of O3 and precursor species is injected into the
layer by penetrative convection. At night, O3 and the remnants of other photochemical
reaction products may remain in this layer and be transported long distances downwind.
These processes are modeled in Layer 3.
When cumulus clouds are present in a Layer 3 cell, the upward vertical mass flux
from the surface is partially diverted from injection into Layer 1 to injection directly into the
3-151
-------
Table 3-26. Regional Oxidant Model Geographical Domains
GENERAL INFORMATION
ROM grid cells are 1/4° longitude and 1/6° latitude in size or approximately 18.5 km.
Actual domain names are included in parenthesis after the general geographical description.
In addition, all domains can be run independently or windowed from the "super" domain.
SUPER DOMAIN (SUPROXA)
99.00 W to 67.00 W Longitude
26.00 N to 47.00 N Latitude
128 x 126 Grid Cells (columns x rows)
NORTHEAST DOMAIN (NEROXA)
89.00 W to 67.00 W Longitude
35.00 N to 47.00 N Latitude
88 x 72 Grid Cells (columns x rows)
MIDWEST DOMAIN (MIDROXA)
97.00 W to 78.00 W Longitude
35.00 N to 47.00 N Latitude
76 x 72 Grid Cells (columns x rows)
SOUTHEAST DOMAIN (SEROXA)
98.75 W to 76.25 W Longitude
27.67 N to 37.67 N Latitude
90 x 60 Grid Cells (columns x rows)
SOUTHERN DOMAIN (TEXROXA)
99.00 W to 81.00 W Longitude
26.00 N to 37.67 N Latitude
72 x 70 Grid Cells (columns x rows)
NORTHEAST DOMAIN (ROMNET)
85.00 W to 69.00 W Longitude
36.33 N to 45.00 N Latitude
64 x 52 Grid Cells (columns x rows)
NORTHEAST DOMAIN (NEROS1)
84.00 W to 69.00 W Longitude
38.00 N to 45.00 N Latitude
60 x 42 Grid Cells (columns x rows)
SOUTHEAST DOMAIN (SEROS1)
97.00 W to 82.00 W Longitude
28.00 N to 35.00 N Latitude
60 x 42 Grid Cells (columns x rows)
cumulus cloud of Layer 3. In the atmosphere, strong thermal vertical updrafts, primarily
originating near the surface in the lowest portion of the mixed layer, feed growing "fair-
weather cumulus" clouds with vertical air currents that extend in one steady upward motion
from the ground to well above the top of the mixed layer. These types of clouds are termed
fair-weather cumulus because atmospheric conditions are such that the clouds do not grow to
the extent that precipitation forms. The dynamic effects of this transport process and daytime
cloud evolution can have significant effects on the chemical fate of pollutants. Within the
ROM system, a submodel parameterizes the above-cloud flux process and the subsequent
impact on mass fluxes among all layers of the model. In the current implementation of the
chemical kinetics, liquid-phase chemistry is not included, and, thus, part of the effects from
the cloud flux processes are not accounted for in the simulations. The magnitude of the mass
flux proceeding directly from the surface layer to the cloud layer is modeled as being
proportional to the observed amount of cumulus cloud coverage and inversely proportional to
the observed depth of the clouds.
Horizontal transport within the ROM system is governed by hourly wind fields that
are interpolated from periodic wind observations made from upper-air soundings and
3-152
-------
98 97 96 95 94 93 92 91 90 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67
47
99 98 97 96 95 94 93 92 91 90 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67
Longitude
Figure 3-26. Regional oxidant model superdomain with modeling domains.
surface measurements. During the nighttime simulation period, the lowest few hundred
meters of the atmosphere above the ground may become stable as a radiation inversion forms.
Wind speeds increase just above the top of this layer, forming the nocturnal jet. This jet is
capable of carrying O3, other reaction products, and emissions injected aloft considerable
distances downwind. This phenomenon is potentially significant in modeling regional-scale
air quality and is implicitly treated by the model, where the definition of Layer 1 attempts to
account for it.
The ROM system requires five types of "raw" data inputs: (1) air quality,
(2) meteorology, (3) emissions, (4) land use, and (5) topography.
Air quality data required by the ROM include initial conditions and boundary
conditions. The model usually is initialized 2 to 4 days before the start of the period of
interest with clean tropospheric conditions for all species. This period of interest is called an
"episode" and usually lasts around 15 days. Ideally, the initial condition field will have been
transported out of the model domain in advance of the portion of the episode of greatest
interest. Upwind lateral boundary conditions for O3 are updated every 12 h based on
measurements, except for the large superdomain, where tropospheric background values are
3-153
-------
used. Other species concentrations at the boundaries, as well as all species at the top of the
modeling domain, are set to tropospheric clean-air concentrations.
Meteorological data are assimilated by the first stage of preprocessors. These data
contain regular hourly observations from U.S. National Weather Service surface stations (and
from similar stations in Canada as necessary), including wind speed and direction, air
temperature, dew point, atmospheric pressure, and cloud amounts and heights. Twice-daily
sounding data, from the upper-air observation network also are included in the meteorological
database. Upper-air meteorological parameters include atmospheric pressure, wind speed and
direction, and air temperature, dew point. Finally, both buoy and Coastal Marine Automated
Station data are used. The parameters that typically are reported include wind speed and
direction and air and sea temperatures.
Emissions data for the primary species are input to the ROM system as well.
Originally these data were provided from the 1985 emissions inventory of NAPAP, with 18.5-
km spatial resolution. Most recently, the interim regional inventory has been used widely to
support current applications of the ROM. It represents an update and improvement of the
NAPAP inventory and is being used to support SIP modeling until state inventories are
approved (U.S. Environmental Protection Agency, 1993a,b). Species included are CO, NO,
NO2, and 10 hydrocarbon reactivity categories. Natural hydrocarbons also are input, including
isoprene explicitly, monoterpenes divided among the existing reactivity classes, and
unidentified hydrocarbons. The chemical mechanism in ROM is the CBM-IV.
Land-use input data consist of 11 land-use categories in 1/4D longitude by 1/6D
latitude grid cells. The data are more than 20 years old and represent a weakness. New land-
use data slowly are being collected and released. Changes in land use over the last 20 years
may change significantly the estimates of biogenic hydrocarbon emissions for large regions of
the United States. Data are provided for the United States and Canada as far as 55D N. The
land-use categories are (1) urban land, (2) agricultural land, (3) range land, (4) deciduous
forests, (5) coniferous forests, (6) mixed-forest wetlands, (7) water, (8) barren land, (9)
nonforested wetland, (10) mixed agricultural land and range land, and (11) rocky, open places
occupied by low shrubs and lichens. Land-use data are used to obtain biogenic emissions
estimates, as a function of the area of vegetative land cover, and for the determination of
surface heat fluxes.
Topography input data consist of altitude matrices of elevations in a 7.5D X 7.5D
grid. The data are obtained from the GRIDS database operated by EPA's Office of
Information Resources Management. Topography data are used in the calculation of layer
heights.
The ROM does have its limitations, including the large grid size, relatively crude
wind fields, and highly empirical vertical mixing assumptions (Wolff, 1993).
3.6.3.3 The Regional Acid Deposition Model
The RADM initially was developed at the NCAR for EPA and subsequently was
refined and improved at the State University of New York at Albany. The model is an
Eulerian transport, transformation, and removal model that includes a treatment of the relevant
physical and chemical processes leading to acid deposition and the formation of photochemical
oxidants. As summarized in Tables 3-21 through 3-25, these processes include atmospheric
transport and mixing, gas-phase and aqueous-phase chemical transformations, dry deposition,
and cloud mixing and scavenging.
3-154
-------
Chemical trace species are transported and diffused through the three-dimensional
RADM grid using externally specified meteorological data. The RADM uses hourly three-
dimensional fields of horizontal winds, temperature, and water vapor mixing ratio calculated
by the meteorological model MM4 with FDDA. In addition, RADM requires
two-dimensional, hourly fields of surface temperature, surface pressure, and precipitation rates
over the model domain. Kuo et al. (1985) found that in order to calculate accurate mesoscale
trajectories, at least 3-h temporal resolution is desirable, and the 12-h resolution of upper air
observations is inadequate. Recent verification studies with 30 meteorological episodes by
Stauffer and Seaman (1990) further support the use of MM5 data with FDDA. Using
meteorology generated from a dynamically consistent meteorological model can introduce
errors caused by simulation errors associated with the meteorological model. These
uncertainties can be quantified through objective verification studies with observed data
(Anthes et al., 1985; Stauffer and Seaman, 1990).
The RADM2 chemical mechanism has been described by Stockwell et al. (1990),
Chang et al. (1991b), Carter and Lurmann (1990), and Stockwell and Lurmann (1989). For
RADM2, the VOCs are aggregated into 12 classes of reactive organic species. Each category
of VOC is represented by several model species that span the required range for reaction with
the OH radical. Most emitted organic compounds are lumped into surrogate species of similar
reactivity and molecular weight, although organic chemicals with large emissions are treated as
separate model species even though their reactivities may be similar. Categories of VOCs with
large reactivity differences and complicated secondary chemistries are represented by larger
numbers of intermediate and stable species. During the aggregation of organic species, the
principle of reactivity weighting is followed to attempt to account for differences in reactivity.
A major part of the SARMAP program described earlier is the extension of the
RADM. The SARMAP is the modeling and data analysis component of a multi-year
collaboration between two projects— SJVAQS and AUSPEX. In the near term, the objective
of SARMAP is to produce a model that can be used to examine scenarios for control of
O3 precursor emissions as required under the CAAA for the 1994 planning cycle. The goals of
the SARMAP modeling program can be summarized as follows:
• Development of a comprehensive state-of-the-science three-dimensional
modeling system (consisting of emissions, meteorological, and air quality
models) suitable for the simulation of O3 concentrations, PM10 concentrations,
visibility degradation, and acid deposition;
• Evaluation of the modeling system and its individual components against
experimental data collected during the SJVAQS/AUSPEX field program; and
• Application of the model to estimate the effect of changes in emission levels on
O3 concentrations, PM10 concentrations, visibility degradation, and acid
deposition.
The general attributes of the SARMAP modeling system are listed below.
• Integrated system of individual modules, including air quality, meteorological,
emissions, and emissions projection; full compatibility of gridding system
among all models.
• Ozone estimation capability; capability for efficiently incorporating modules for
simulating aerosols, visibility, and acid deposition.
• Applicability at urban, subregional, and regional scales, embodying a full range
of anticipated physical, chemical, and terrain characteristics.
3-155
-------
• Capability of being driven by larger meteorological models, if desired (for
generating initial and boundary conditions).
• Capability of generating as output a full complement of chemical species
concentrations and meteorological parameters.
• Variable horizontal grid size.
• Variable number of vertical layers.
• Variable depth of vertical layers.
• Capability of nested grid application.
• Capability of varying the number of vertical layers with time of day. Selection
of number of layers and timing of changes to be model-driven.
• Improved treatment of emissions injection aloft, including placement of plumes
in the vertical, treatment of inversion penetration, proper vertical dilution of
plumes, and proper treatment of chemistry.
• Inclusion of plume-in-grid capability.
• Capability for use of "computational tracers" for a variety of tests.
• Capability of simulating the O3-VOC-NOX system alone or in tandem with the
aerosol system.
• Capability of simulating aerosols for the O3-VOC-NOX system.
The following modifications to the RADM2 gas-phase chemical mechanism have
been made:
• Updating the rate constants, product parameters, and absorption cross sections
and quantum yields for consistency with current recommendations;
• Improving the treatment of isoprene chemistry;
• Adapting the SAPRC emissions processing scheme to the RADM2 mechanism;
and
• Adding extra species (acetaldehyde, PAN, and an additional aromatic) and their
associated reactions and products.
The Smolarkiewicz scheme currently used in RADM will be replaced with the Bott scheme.
This scheme is more accurate than the Smolarkiewicz scheme for continuous plumes and at low
grid resolutions. The RADM cloud module will be replaced with the ADOM module. The
RADM dry deposition module currently underestimates dry deposition velocities under stable
conditions. This can result in unrealistically high O3 concentrations at night.
3.6.4 Evaluation of Model Performance
Air quality models are evaluated by comparing their predictions with ambient
observations. Because a model's demonstration of attainment of the O3 NAAQS is based on
hypothetical reductions of emissions from a base-year-episode simulation, the accuracy of the
base-year simulation is necessary, but not sufficient. An adequate model should give accurate
predictions of current peak O3 concentrations and temporal and spatial O3 patterns. It should
also respond accurately to changes in VOC and NOX emissions, to differences in VOC
reactivity, and to spatial and temporal changes in emissions patterns for future years.
Model performance can be evaluated at several levels. The important sub-models,
the emissions model, the meteorological model, and the chemical mechanism can be evaluated
independently, and the model as a whole can be evaluated. Evaluation of emissions models
can be carried out with special measurements designed to isolate the effects of emissions from
a particular source category, such as tunnel studies (Pierson et al., 1990) or on-road
3-156
-------
surveillance of motor vehicles (Lawson et al., 1990) to evaluate the accuracy of motor vehicle
emissions models. Meteorological sub-models can be evaluated from the results of tracer
experiments. Chemical mechanisms have traditionally been developed and evaluated on the
basis of smog chamber experiments. A question that merits continued attention is how well
chemical mechanisms developed with reference to smog chamber data perform when
simulating the ambient atmosphere. As noted in this section, comparisons of observed and
predicted concentrations for all important precursors, intermediates, and products are
important in assessing the accuracy of a chemical mechanism.
Compilations of the performance of photochemical models in the South Coast Air
Basin of California and in other urban areas indicate a general tendency toward the
underprediction of O3 concentrations and particularly O3 maxima. It should be noted that
different areas of the country are characterized by different controlling factors in
O3 generation, so the reasons for O3 underprediction in one area may not be the same as in
another. A case in point is the possibility of anthropogenic ROG emissions underestimation in
urban areas versus biogenic ROG emissions underestimation in rural and regional areas. It is
well-recognized that urban and regional photochemical models have a number of uncertain
input quantities, so it is possible, by adjusting these quantities within their ranges of
uncertainty, to improve O3 predictions. This process, which is inherent in any modeling
exercise because of the uncertainty associated with many of the input quantities, can lead to
getting the right answer for the wrong reason. Because the modeling of an O3 episode usually
is carried out to establish a "base case" against which to evaluate the effects of VOC and NOX
emissions changes, the accuracy of the base case is vital for obtaining a valid assessment of the
effects of emissions perturbations. Due to the nonlinear response of the O3/VOC/NOX system,
conclusions drawn about the effect of VOC and NOX emissions changes may not reflect actual
atmospheric response if the base case simulation is inaccurate. For this reason, it is important
to understand the reasons why the base case simulation may not agree with observations.
Several more or less equivalent alternate base cases may exist due to the fact that it often is
possible to vary inputs within their ranges of uncertainties to achieve comparable model
performance. Unfortunately, the O3 responses to identical VOC/NOX controls may be rather
different depending on which base case is used.
3.6.4.1 Model Performance Evaluation Procedures
Specific numerical and graphic procedures have been recommended for evaluation
of the accuracy of grid-based photochemical models (Tesche et al., 1990b). The recommended
methods include the calculation of peak prediction accuracy; various statistics based on
concentration residuals; and time series of predicted and observed hourly concentrations. Four
numerical measures appear to be most helpful in making an initial assessment of the adequacy
of a photochemical simulation (Tesche et al., 1990b): (1) the paired peak prediction accuracy,
(2) the unpaired peak prediction accuracy, (3) the mean normalized bias, and (4) the mean
absolute normalized gross error.
Accurate matching of O3 alone may not be sufficient to ensure that a model is
performing accurately. The possibility of compensatory errors must be recognized (in which
two or more sources of error interact in such a way that O3 is predicted accurately, but for the
wrong reasons). The inaccuracies offset each other in part. The modeling effort should be
designed to minimize the likelihood of the presence of compensatory errors.
3-157
-------
Evaluation of model performance for precursor and intermediate species, as well as
for product species other than O3, when ambient concentration data for these species are
available, significantly improves the chances that a flawed model will be identified.
Comparisons of observed and predicted concentrations for all important precursors,
intermediates, and products involved in photochemical air pollution, such as individual VOCs,
NO, NO2, PAN, O2, H2O2, HNO2, and HNO3, are useful in model evaluation, especially with
respect to the chemistry component of the model (Jeffries et al., 1992). Comparisons of
predictions and observations for total organic nitrates (mainly PAN) and inorganic nitrates
(HNO3 and nitrate aerosol) can be used to test qualitatively whether the emissions inventory
has the correct relative amounts of VOCs and NOX. However, in order to include HNO3 and
nitrate aerosol in the data set for model comparisons, the model should include an adequate
description of the HNO3 depletion process associated with aerosol formation.
Adequate model performance for several reactive species increases the assurance
that correct O3 predictions are not a result of chance or fortuitous cancellation of errors
introduced by various assumptions. Multispecies comparisons could be the key in
discriminating among alternative modeling approaches that provide similar predictions of
O3 concentrations.
As noted above, photochemical models have the potential to produce nearly the
right O3 concentrations when performance is evaluated, but do so because two or more flaws
were compensating each other. The existence of compensating errors in many modeling
applications is suspected because most applications have used emission inventories whose
validity is now in question (National Research Council, 1991). Underestimation of VOC
emissions from motor vehicles may be responsible for the lack of agreement between
inventories and ambient concentration data (Baugues, 1986; Lawson et al., 1990; Pier son
et al., 1990; Fujita et al., 1992). Underestimation of emissions from other sources is also a
possibility. One potentially underestimated VOC source is vegetation, which naturally emits
VOCs. An underestimation of VOC emissions could be compensated for by underestimation
of mixing height or wind speed, by overestimation of boundary concentrations of O3 or
precursors, or by inaccurate chemistry modules. Boundary concentrations (which can be
obtained from measurements or regional models or by assuming background concentrations,
often are poorly defined.
If only a routine database is available for modeling O3 in an urban area, then there
are four areas of concern that require attention (Roth, 1992).
(1) Air Quality Aloft. These data most likely will not be available. These
measurements are important and are instrumental for diagnostic analysis of
model simulations.
(2) Boundary Conditions. If the possibility of significant transport into the region
exists, but the data are not available, the boundary conditions become a
variable that allows the introduction of compensatory errors if the emissions
estimates are inaccurate. An approach to circumventing this problem is to
define the region in such a way that the boundaries become a much less
significant issue.
(3) Ambient VOC Data. These generally are not routinely available. In their
absence, evaluation of model performance is hampered.
(4) Meteorological Data Aloft. Very often, there are only surface measurements
and a few soundings from which to extrapolate the needed data.
3-158
-------
If any of these four areas is missing from the database, the performance evaluation
and subsequent model application must be planned to minimize the possibility of compensatory
errors.
3.6.4.2 Performance Evaluation of Ozone Air Quality Models
Urban Airshed Model
The UAM has been applied to many urban areas in the United States and Europe,
and most of these studies have included some form of performance evaluation (see summary in
Tesche et al., 1993, Table 6-2). Thus, there is a growing body of information concerning the
accuracy of the model's predictions; UAM itself is continuing to undergo revision.
Evaluations of UAM's performance have been carried out for a number of geographic areas.
Evaluations conducted since 1985 have indicated mean discrepancies between predicted and
measured O3 values of 20 to 40% of the observations, when paired in space and time (Roth
et al., 1990). The prediction of peaks exhibits relative errors that are smaller than the average
error, with a tendency toward underprediction (Roth et al., 1990). The discrepancies between
predicted and measured NO2 in UAM applications are on the order of 30 to 50%, with no
improvement over the history of modeling applications (Roth et al., 1990). Underprediction of
NO2 by UAM has been typical, generally on the order of 20 to 40% (Roth et al., 1990).
As a result of the discovery of significantly underestimated mobile source VOC
emissions (in the late 1980s), this emissions underestimation is the leading cause of
O3 underprediction in urban areas.
Regional Oxidant Model
A primary role of the ROM is to estimate boundary conditions for use by UAM in
evaluating hydrocarbon and NOX reduction strategies for urban areas in the eastern United
States. This is especially the case in areas where transport is a significant element (U.S.
Environmental Protection Agency, 1990d). Analysis of regional O3 abatement strategies also
is a major role of the ROM (Possiel et al., 1990).
The ROM has been used in the EPA program, the Regional Ozone Modeling for
Northeast Transport (ROMNET) program, to assess the effectiveness of various regional
emission control strategies in lowering O3 concentrations to nationally mandated levels for the
protection of human health, forests, and crops (Meyer et al., 1991b). As part of the ROMNET
program, the ROM also is being used to provide regionally consistent initial and upwind
boundary conditions to smaller-scale urban models for simulations of future-year scenarios.
The most complete testing of ROM2.0 was accomplished in an evaluation with the
50-day (July 12 to August 31, 1980) Northeastern Regional Oxidant Study database (Schere
and Wayland, 1989a,b). The model underestimated the highest values and overestimated the
lowest. It produced an overall 2% overprediction in predicting maximum daily
O3 concentrations averaged over aggregate groups of monitoring stations. A key indicator of
model performance on the regional scale is the accuracy of simulating the spatial extent and
location, as well as the magnitude, of the pollutant concentrations within plumes from
significant source areas. In ROM2.0 performance analyses, plumes from the major
metropolitan areas of the Northeast Corridor, including Washington, DC; Baltimore, MD;
New York; and Boston, could be clearly discerned in the model predictions under episodic
conditions. Generally, the plumes were well characterized by the model, although there was
evidence of a westerly transport bias and underprediction of O3 concentrations near the center
3-159
-------
of the plume. Using aircraft data, ROM2.0 was found to underpredict the regional
tropospheric burden of O3.
The evaluation of ROM2.1 (Pierce et al., 1990), unlike that of ROM2.0, was based
on routinely archived data from state and local agency monitoring sites rather than on an
intensive field-study period. The evaluation consisted of the comparison of observed and
predicted O3 concentrations during selected episodes (totaling 26 days) of high O3 observed
during the summer of 1985. Evaluation showed that ROM2.1 underestimated the highest
values and slightly overestimated the lowest; underestimates of the upper percentiles tended to
be more prevalent in the southern and western areas of the ROMNET domain (Table 3-26).
The model exhibited an overall 1.4% overprediction in predicting maximum daily O3
concentrations averaged over aggregate groups of monitoring stations, and it appears to correct
for the westerly transport bias of high-O3 plumes in the Northeast Corridor seen in ROM2.0.
As with ROM2.0, model performance degraded as a function of increasingly complex
mesoscale wind fields.
In a recent evaluation of ROM (Systems Applications International, 1993),
ROM2.2 overestimated observed O3 maxima by 20 to 30 ppb over the period of July 4 through
6, 1988, and predicted an episodic peak of 242 ppb on July 9, 1988, when the observed peak
was 138 ppb. The ROM2.2 performance for hourly O3 concentrations in the New York region
exceeded the range of EPA acceptable performance by a factor of two 90% of the time during
the July 1988 episode. The Systems Applications International (1993) report concluded that
"the patchiness of the ROM2.2 predictions compared to the observations raises serious
questions as to whether the model will respond correctly to emission control strategies." The
major conclusions of that report were:
• Model performance downwind of New York City is "unacceptable". The model
significantly overpredicts peak O3 levels, and the predicted diurnal variation of
O3 occurs too late in the afternoon.
• Model performance for the Philadelphia and Baltimore/Washington urban
plumes is "poor" with "unpaired peak estimation accuracy at the outer edge of
the acceptable range."
• Elsewhere, the model seems to give good results, although it produces O3 spatial
distributions that are too "patchy" when compared to observations.
• There is a systematic westerly bias in the ROM2.2 wind fields.
• The model performance for NOX is "extremely poor" indicating that ROM2.2
may be overestimating the VOC/NOX ratios across the region.
3.6.4.3 Database Limitations
As previously mentioned, the use of routine air quality and meteorological data
requires that a number of assumptions be made about key model inputs. Although intensive
field studies are desirable during O3 episodes to acquire the full set of data required, three key
problems arise: (1) such studies are expensive and, therefore, are limited in number; (2) the
time required to carry out field studies usually exceeds the time available; and (3) most field
studies have not captured the worst O3 episodes. Because EPA guidance emphasizes planning
to meet worst-case conditions, field data often must be manipulated to approximate highest
O3 concentrations. Such adjustments invariably increase uncertainty in model projections.
Studies that have, or will, provide data for model evaluation include the St. Louis,
MO, RAPS, conducted in 1975 and 1976; the Northeast Corridor Regional Modeling Project,
3-160
-------
conducted in 1979 and 1980; the South Central Coast Cooperative Aerometric Monitoring
Program, conducted in 1985; SCAQS conducted in 1987; studies in Sacramento and
San Diego, CA, in 1990; SJVAQS/AUSPEX conducted in 1990; LMOS conducted in 1990
and 1991; SOS conducted in 1991 and 1992; and a Gulf Coast study for 1993.
In most cases, field studies have not coincided with periods in which ozone
concentrations have attained values as high as that on which the SIP must be based. Given the
low probabilities of occurrence of the most adverse meteorological conditions and the fact that
field studies typically acquire data for two or three ozone episodes, obtaining a design value
concentration during the course of a field study is unlikely.
The EPA recommends that the five highest daily maximum O3 concentrations at a
design-value site, selected from the three most recent years, be modeled if EKMA is used for a
SIP (U.S. Environmental Protection Agency, 1989b). Because EKMA's data requirements are
minimal, it can be applied to the worst cases. In contrast, the number of episodes available for
grid-based modeling is less than desirable in all areas. In addition, any available intensive
databases often do not include the worst-case meteorology; intensive databases typically restrict
modeling to two or three O3 episodes having a duration of 2 to 3 days each. Moreover, the
intensive databases never encompass the full range of meteorological conditions of interest (if
O3 exceedances occur in an area under different meteorological conditions, the relative
effectiveness of different control strategies might vary with the different meteorological
conditions). The EPA specifies procedures for episode selection for use with grid-based
models (U.S. Environmental Protection Agency, 1991b).
Because the number of intensive databases is limited both in terms of episodes and
regions, EPA has investigated the feasibility of applying UAM without conducting intensive
field studies (Scheffe and Morris, 1990, 1991). These studies, known as the Practice for Low-
cost Application in Nonattainment Regions (PLANR), were conducted for New York;
Philadelphia; Atlanta; Dallas-Fort Worth, TX; and St. Louis. Of the five cities studied,
St. Louis, New York, and Philadelphia had intensive databases available. Simulations were
carried out using both routine and intensive databases for St. Louis and Philadelphia. Model
performance using routine data was much better for St. Louis than for Philadelphia (Scheffe
and Morris, 1990, 1991). Scheffe and Morris (1990, 1991) cautioned that the differing results
may be complicated by the quality of the databases, but they speculate that model performance
using routine databases for Philadelphia might have been poorer because of regional transport.
Performance statistics for all four applications using routine data were consistent with other
UAM applications (Scheffe and Morris, 1990, 1991); however, the paucity of data in the
routine databases precluded any investigation of the possibility that compensating errors
occurred.
Scheffe and Morris (1990, 1991) note that the PLANR lack of air quality data was
addressed by extending the length of the simulations and expanding the upwind boundary,
which, in effect, increased the need for accurate emissions inventories (boundary conditions
could also be obtained through use of ROM). For PLANR applications, gridded emissions
were created from routine county-level emission inventories by utilizing an emissions program
that made use of surrogate information, such as population distribution. The PLANR study
represents an interesting start on the problem of model application to areas without intensive
databases; however the results were not sufficiently definitive for drawing conclusions of a
broad, general nature.
3-161
-------
3.6.5 Use of Ozone Air Quality Models for Evaluating Control Strategies
Photochemical air quality models are used for control strategy evaluation by first
demonstrating that a past episode, or episodes, can be adequately simulated and then reducing
hydrocarbon or NOX emissions in the model inputs and in assessing the effects of these
reductions on O3 in the region. Ozone concentrations can be decreased by reducing either
VOC or NOX concentrations to sufficiently low levels. The effects of NOX emissions
reductions on O3 concentrations vary because NOX is an atypical precursor (i.e., although it is
necessary for O3 formation, fresh NO emissions remove O3, and high concentrations of NOX
retard the rate of O3 formation by removing radicals). Control of NOX tends to accelerate the
rate of O3 formation; however, its effects on peak O3 concentration depend on the location and
timing of the control and on ambient concentrations of VOCs and NOX, which vary widely in
time and space, even within a single urban area during 1 day.
At a given VOC level, as the initial NOX is increased, O3 first increases, then peaks,
and then decreases. The reduction in peak O3 with increasing NOX is a well-established
chemical phenomenon. The peak in O3 formation occurs at an initial VOC/NOX ratio of about
10/1 (i.e., 10 ppbC/1 ppb). At fixed NOX level, as VOC is increased, O3 formation increases
but then levels off. As a result of this behavior, at VOC/NOX ratios below about 10/1, VOC
reduction has been the preferred strategy for O3 reduction. In this region NOX reductions
speed up O3 formation and lead to higher peak O3 values. At VOC/NOX ratios exceeding about
10/1, both VOCs and NOX will reduce O3, but less than proportionally. The reason the
reduction in O3 is less than proportional is because equal reductions of VOCs and NOX at
intermediate ratios tend to keep O3 production at its maximum. The nonlinear chemical
behavior of the VOC/NOX system, discussed earlier in this chapter, is at the heart of the
controversy over the role of NOX in O3 control (Heuss and Wolff, 1993).
As noted in Section 3.6.1.2, the concept that a region is characterized by a single
VOC/NOX ratio is oversimplified and may actually lead to incorrect conclusions concerning the
optimal approach to O3 reduction (Milford et al., 1989). The VOC/NOX ratio in a region is a
function of location and time of day; the source-rich center city area may be characterized by a
lower ratio than that in downwind, suburban areas at any given time of day. Because of the
complex spatial and temporal dependence of O3 formation, grid-based photochemical air
quality models are necessary to evaluate the effect of emission reduction strategies for a
region.
Moreover, location-specific studies need to be performed to ascertain whether a
given area is in the VOC- or NOX-controlled regime. Research is being conducted into the
relationship between O3 and NOy to determine whether NOy is a better indicator of the
O3-forming potential than the VOC/NOX ratio (Shepson et al., 1992b; Trainer et al., 1993;
Kleinman et al., 1994; Milford et al., 1994).
In most modeling applications, inputs are adjusted within their range of uncertainty
to improve performance. A key test of quality of performance is to evaluate the model
predictions for other episodes without adjustments, using the same procedures for establishing
inputs as for the original episode.
Grid modeling applications are currently underway by or for state agencies for
approximately 20 areas within the United States to support regional O3 SIP revisions.
An immediate problem faced for almost all urban areas is that even if an adequate
number of episodes exist, the episodes may not include the most adverse O3 levels.
An inherent question in using a less adverse episode to develop control strategies is how do
3-162
-------
these strategies extrapolate to a more severe set of conditions? There is no clear answer to this
question. At present, control strategies, evaluated by using grid-based models, are determined
based on available episodes that have the largest amount of data, whether or not these episodes
contain the highest O3 concentration achieved. Another issue is that the form of the NAAQS
for O3 does not correspond with the output from a grid-based model. The model output does
not provide a direct answer to whether an area will meet the standard in its current statistically
based form.
Table 3-27 summarizes a number of recent O3 control strategy evaluations for
different areas of the United States. Some general observations can be made concerning issues
that have arisen in control strategy exercises, particularly as they relate to problems associated
with different areas of the country (Roth, 1992). In California, model results indicate that O3
has been underestimated, most likely because VOC emissions from motor vehicles have been
seriously underestimated. The underestimation was hidden by adjusting other model inputs
within their range of uncertainty. In Atlanta, it has been estimated that approximately 60% of
the VOC inventory is of biogenic origin, and the variation of anthropogenic emissions
reductions required to achieve O3 attainment within the uncertainty range of the biogenic
emissions is on the order of 20%. The uncertainty range of the biogenic VOC emissions needs
to be reduced to obtain tighter control strategy estimates.
The eastern United States poses special problems in regional-scale photochemical
modeling. Boundary conditions typically contribute 40 to 70% of pollutant loading in many
urban areas east of the Mississippi River. Regional-scale models are often either not available
or not sufficiently reliable to use in estimating upwind boundary conditions. Furthermore, data
are rarely available. If data are available, their use is limited to estimation of present
conditions. If models are used in control strategy assessment and 40 to 70% of pollutant
loading originates outside of the modeling region, major questions arise as to just how control
strategies are to be determined. If uncertainties at the regional scale are significant and if
regional-scale modeling is inaccurate, the limits of accuracy for urban-scale control strategy
determination need to be carefully assessed.
An essential question is, given the inevitable uncertainties associated with O3 air
quality model predictions, can the effect of VOC and NOX emissions changes on O3 levels be
unambiguously determined? The best approach to answering this question is a combination of
sensitivity/uncertainty studies. Given the estimated uncertainties in model inputs and
parameters for a particular application, the proposed VOC and NOX emissions change scenarios
should be examined for the full range of model inputs and parameters to determine how
sensitive conclusions about the effect on O3 levels are to the inherent uncertainties.
3.6.6 Conclusions
The 1990 CAAA (U.S. Congress, 1990) have mandated the use of photochemical
grid models for demonstrating how most O3 nonattainment areas can attain the NAAQS.
Predicting O3 is a complex problem. There are still many uncertainties in the models;
nonetheless, models are useful for regulatory analysis and constitute one of the major tools for
attacking the O3 problem. These models have developed considerably in the past 10 years.
However, their usefulness is constrained by having limited databases for use in model
evaluation and from having to rely on hydrocarbon emissions data that may be inaccurate.
3-163
-------
Table 3-27. Applications of Photochemical Air Quality Models to Evaluating Ozonea
Investigators Region/Episode Model Used Strategies Evaluated
Chu et al. (1993)
Chu and Cox (1993)
Roselle et al. (1992)
Mathur and Schere (1993)
Possieletal. (1993)
Possiel and Cox (1993)
Milford et al. (1992)
Rao (1987)
Rao et al. (1989)
Rao and Sistla (1993)
Scheffe and Morris
(1990, 1991)
Possieletal. (1990)
Roselle and Schere (1990)
Roselle et al. (1991)
Bunker et al. (1992a,b)
Eastern United States;
July 2-10, 1988
Northeastern United
States;
July 1-12, 1988
Northeastern United
States;
July 2-17, 1988
New York metropolitan
area,
5 days in 1980
New York
St. Louis
Atlanta
Dallas-Ft. Worth
Philadelphia
Northeastern United
States;
July 2-17, 1988
Northeastern United
States;
July 12-18, 1980
Los Angeles
New York
Dallas-Ft. Worth
ROM2.2
ROM2.2
ROM
UAM/ROM2.1
UAM
ROM
ROM2.1
UAM
Across-the-board NOX/VOC
reductions
Estimate Q reductions per 1990
CAAA
Analysis of effect of NOX
reductions
Evaluation of 1988 SIPs and
VOC/NO, strategies
Use of UAM for demonstrating
attainment with routinely
available data
Ozone control strategies in
Northeast
Sensitivity of Oj in Northeast to
biogenic emissions
Effects of alternate fuels and
reformulated gasolines on O3
levels
Milford et al. (1989)
South Coast Air Basin
CIT
Effects of systematic VOC and
NO, reductions
Middleton et al. (1993)
Eastern United States and
southeastern Canada
RADM
2010 emissions projections
"See Appendix A for abbreviations and acronyms.
Primary issues and limitations associated with the use of photochemical air quality
models are described below.
• High noise-to-signal ratios. Model imprecision for ozone predictions typically
ranges from 25 to 40%, and inaccuracy (bias) ranges from 5 to 20%. These
uncertainties are often of the same order as the percentage of reduction in the
peak O3 concentration for an area (from 160 to 120 ppb). Reasons for these
inaccuracies include uncertainties in emissions inventories.
• Inadequacies of supporting databases in most geographical areas. Most areas
are lacking or are deficient in data needed to estimate boundary conditions and
3-164
-------
meteorological and air quality conditions aloft. There are few areas where
speciated VOC concentrations are measured; surface NOX data may be
inaccurate. Where important data gaps exist, modeling accuracy suffers, and
the prospects for reducing or eliminating the presence of compensating errors
are diminished.
• Continuing need for improvements. Examples include the introduction of
prognostic meteorological modeling in the mid-1980s, the discovery of
underestimation of VOC emissions in the late 1980s, the inclusion of NOX
emissions from soils in 1993, and major adjustment of the emissions rates of
isoprene in 1994.
• Presence of compensating errors. It appears that compensating errors have been
present in many past applications, introducing the potential for bias into the
estimation of the impacts of emissions control strategies.
Comparison of model predictions against ozone measurements, although necessary,
is not a robust test of a model's accuracy. Ideally, one should evaluate performance against
more extensive sets of species such as individual VOCs, NOX, and NOy. Compensating errors
in input information to a model and within the model formulation can cause an O3 model to
generate correct O3 predictions for the wrong reasons. Therefore, model evaluation indicators
are needed to demonstrate the reliability of a prediction before the model can be used
effectively in making control strategy decisions.
It is important to stress that, in O3 modeling, a modeling system also is at issue, not
just the air quality model itself. The modeling system includes a meteorological model, an
emissions representation (where an emissions model is preferred to the traditional "inventory"
approach), the air quality model, and a comprehensive supporting database. Where a problem
exists, the entire modeling system must be evaluated.
Models can be used effectively in a relative sense to rank different control
alternatives in terms of their effectiveness in reducing O3 and to indicate the approximate
magnitude of improvement in peak O3 levels expected under various control strategies. To do
so, there must be a sound emissions model and data and an adequate database on which to
construct the modeling. Grid-based O3 air quality modeling is superior to the available
alternatives for O3 control planning, but results can be misleading if the model is not evaluated
sufficiently. The goal is to minimize the chances of incorrect use of the model.
3.7 Summary and Conclusions
3.7.1 Tropospheric Ozone Chemistry
3.7.1.1 Ozone in the Unpolluted Atmosphere
Ozone is found in the stratosphere, the "free" troposphere, and the PEL of the
earth's atmosphere. In the stratosphere, O3 is produced through cyclic reactions that are
initiated by the photolysis of molecular oxygen by short-wavelength radiation from the sun and
are terminated by the recombination of molecular oxygen and ground-state oxygen atoms.
In the "free" troposphere, O3 occurs as the result of incursions from the
stratosphere; upward venting from the PEL (which is the layer next to the earth, extending to
altitudes of Dl to 2 km) through certain cloud processes; and photochemical formation from
precursors, notably CH4, CO, and NOX. These processes contribute to the background O3 in
the troposphere.
3-165
-------
Ozone is present in the PEL as the result of downward mixing from the stratosphere
and free troposphere and as the result of photochemical processes occurring within the PEL.
The photochemical production of O3 and other oxidants found at the earth's surface is the
result of atmospheric physical and chemical processes involving two classes of precursor
pollutants, reactive VOCs and NOX. The formation of O3 and other oxidants from its
precursors is a complex, nonlinear function of many factors, including the intensity and
spectral distribution of sunlight; atmospheric mixing and related meteorological conditions; the
reactivity of the mixture of organic compounds in ambient air; the concentrations of precursor
compounds in ambient air; and, within reasonable concentrations ranges, the ratio between the
concentrations of reactive VOCs and NOX.
In the free troposphere and in many relatively "clean" areas of the PEL, CH4 is the
chief organic precursor to in situ photochemical production of O3 and related oxidants.
Exceptions can include clean forested or vegetated areas emitting biogenic organics. The
major tropospheric removal process for CH4 is by reaction with OH radicals. In the complex
cyclic reactions that result in oxidation of CH4, there can be a net increase in O3 or a net loss
of O3, depending mainly on the NO concentration.
3.7.1.2 Ozone Formation in the Polluted Troposphere
The same basic processes by which CH4 is oxidized occur in the atmospheric
oxidative degradation of other, even more reactive and more complex VOCs. The only
significant initiator of the photochemical formation of O3 in the troposphere is the photolysis of
NO2, yielding NO and a ground-state oxygen atom that reacts with molecular oxygen to form
O3. The O3 thus formed reacts with NO, yielding O2 and NO2. These cyclic reactions attain
equilibrium in the absence of VOCs. In the presence of VOCs, however, the equilibrium is
upset, resulting, from a complex series of chain reactions, in a net increase in O3.
The key reactive species in the troposphere is the OH radical, which is responsible
for initiating the oxidative degradation reactions of almost all VOCs. As in the CH4 oxidation
cycle, the conversion of NO to NO2 during the oxidation of VOCs is accompanied by the
production of O3 and the efficient regeneration of the OH radical. The O3 and PANs formed in
polluted atmospheres increase with the NO2/NO concentration ratio.
At night, in the absence of photolysis of reactants, the simultaneous presence of O3
and NO2 results in the formation of the NO3 radical. The reaction with NO3 radicals appears to
constitute a major sink for alkenes, cresols, and some other compounds, although alkyl NO3
chemistry is not well characterized.
Most inorganic gas-phase processes, that is, the nitrogen cycle and its
interrelationships with O3 production, are well understood; the chemistry of the VOCs in
ambient air, however, is not. The chemical loss processes of gas-phase VOCs, with
concomitant production of O3, include reaction with OH, NO3, O3, and photolysis.
The major classes of VOCs in ambient air are alkanes, alkenes (including alkenes
from biogenic sources), aromatic hydrocarbons, carbonyl compounds, alcohols, and ethers. A
wide range of lifetimes in the atmosphere, from minutes to years, characterize the VOCs.
The only important reaction of alkanes is with OH radicals. For alkanes having
carbon-chain lengths of four or less (DC4), the chemistry is well understood and the reaction
rates are slow. For DC5 alkanes, the situation is more complex because few reaction products
have been found. Branched alkanes (e.g., isobutane) have rates of reaction that are highly
dependent on structure. It is difficult to represent reactions of these VOCs satisfactorily in the
3-166
-------
chemical mechanisms of air quality models. Stable products of alkane photooxidation are
known to include carbonyl compounds, alkyl nitrates, and D-hydroxycarbonyls. Major
uncertainties in the atmospheric chemistry of the alkanes concern the chemistry of alkyl nitrate
formation; these uncertainties affect the amount of NO-to-NO2 conversion occurring and,
hence, the amounts of O3 formed during photochemical degradation of the alkanes.
Alkenes react in ambient air with OH and NO3 radicals and with O3. All three
processes are important atmospheric transformation processes, and all proceed by initial
addition to the > C=C < bonds. Products of alkene photooxidation include carbonyl
compounds, hydroxynitrates and nitratocarbonyls, and decomposition products from the
energy-rich biradicals formed in alkene-O3 reactions. Major uncertainties in the atmospheric
chemistry of the alkenes concern the products and mechanisms of their reactions with O3,
especially the radical yields (which affect the O3 formation yields).
The only tropospherically important loss process for aromatics (benzene and the
alkyl-substituted benzenes) is by reaction with the OH radical, followed by H-atom abstraction
or OH radical addition. Products of aromatic hydrocarbon photooxidation include phenolic
compounds, aromatic aldehydes, D-dicarbonyls (e.g., glyoxal), and unsaturated carbonyl or
hydroxycarbonyl compounds. Aromatics appear to act as strong NOX sinks under low NOX
conditions. Major uncertainties in the atmospheric chemistry of aromatic hydrocarbons are
mainly with regard to reaction mechanisms and products under ambient conditions (i.e., for
NOX concentration conditions that occur in urban and rural areas). These uncertainties impact
on the representation of mechanisms in models.
Tropospherically important loss processes for carbonyl compounds not containing
> C = C < bonds are photolysis and reaction with the OH radical; those that contain such
bonds can undergo the same reactions as alkenes. Photolysis is the major loss process for
HCHO (the simplest aldehyde) and acetone (the simplest ketone), as well as for the
dicarbonyls. Reactions with OH radicals are calculated to be the dominant gas-phase loss
process for the higher aldehydes and ketones. Products formed and the importance of
photolysis are major uncertainties in the chemistry of carbonyl compounds.
Alcohols and ethers in ambient air react only with the OH radical, with the reaction
proceeding primarily via H-atom abstraction from the C-H bonds in these compounds.
It should be noted that the photooxidation reactions of certain higher molecular
weight VOCs can lead to the formation of significant yields of organic particulates in ambient
air. The chemical processes involved in the formation of O3 and other photochemical
pollutants lead to the formation of OH radicals and oxidized VOC reaction products that are of
low enough volatility to be present as organic particulate matter. Hydroxyl radicals that
oxidize VOCs also react with NO2 and SO2 to form HNO3 and H2SO4, respectively, which can
become incorporated into aerosols as particulate nitrate and sulfate. Controls aimed at
reducing O3 will also impact acid and secondary aerosol formation in the atmosphere.
3.7.2 Meteorological Processess Influencing Ozone Formation and Transport
3.7.2.1 Meteorological Processes
The surface energy (radiation) budget of the earth strongly influences the dynamics
of the PEL and, in combination with synoptic winds, provides the forces for the vertical fluxes
of heat, mass, and momentum. The redistribution of energy through the PEL creates
3-167
-------
thermodynamic conditions that influence vertical mixing. Energy balances require study so
that more realistic simulations can be made of the structure of the PEL.
Day-to-day variability in O3 concentrations depends heavily on day-to-day variations
in meteorological conditions. For example, the concentration of an air pollutant depends
significantly on the degree of mixing that occurs between the time a pollutant, or its
precursors, is emitted and the arrival of the pollutant at the receptor. Inversion layers (layers
in which temperature increases with height above ground level) are prominent determinants of
the degree of atmospheric vertical mixing and, thus, the degree to which O3 and other
pollutants will be dispersed or accumulate. Ozone left in a layer aloft, as the result of reduced
turbulence and mixing at the end of daylight hours, can be transported through the night, often
to areas far removed from pollution sources. Downward mixing on the subsequent day can
result in increases in local concentrations from the transported O3.
Growing evidence indicates that the conventional use of mixing heights in modeling
is an oversimplification of the complex processes by which pollutants are redistributed within
urban areas. In addition, it is necessary to treat the turbulent structure of the atmosphere
directly and to acknowledge the vertical variations in mixing.
Geography can significantly affect the dispersion of pollutants along the coast or
shore of oceans and lakes. Temperature gradients between bodies of water and land masses
influence the incidence of surface conditions. The thermodynamics of water bodies may play a
significant role in some regional-scale episodes of high O3 concentrations.
An "air mass" is a region of air, usually of multistate dimension, that exhibits
similar temperature, humidity, and stability characteristics. Episodes of high O3 concentrations
in urban areas often are associated with high concentrations of O3 in the surroundings.
The transport of O3 and its precursors beyond the urban scale (D50 km) to
neighboring rural and urban areas has been well documented and was described in the 1986
EPA criteria document for O3. Areas of O3 accumulation are characterized by synoptic-scale
subsidence of air in the free troposphere, resulting in development of an elevated inversion
layer; relatively low wind speeds associated with a weak horizontal pressure gradient around a
surface high pressure system; a lack of cloudiness; and high temperatures.
3.7.2.2 Meteorological Parameters
Ultraviolet radiation from the sun plays a key role in initiating the photochemical
processes leading to O3 formation and affects individual photolytic reaction steps. There is
little empirical evidence in the literature, however, linking day-to-day variations in observed
UV radiation levels with variations in O3 levels.
An association between tropospheric O3 concentrations and tropospheric
temperature has been demonstrated. Plots of daily maximum O3 concentrations versus
maximum daily temperature for the summer months of 1988 to 1990 for four urban areas, for
example, show an apparent upper bound on O3 concentrations that increases with temperature.
A similar qualitative relationship exists at a number of rural locations.
The relationship between wind speed and O3 buildup varies from one part of the
country to another. Research done during the SOS (in the "Atlanta intensive" field study)
indicates that measurements of variations in wind speed among methods at a particular level
above ground must be larger than about 3 m/s to be considered statistically significant.
3.7.2.3 Normalization of Trends
3-168
-------
Statistical techniques (e.g., regression techniques) can be used to help identify real
trends in O3 concentrations, both intra-annual and inter-annual, by normalizing meteorological
variability. In the SOS, for example, regression techniques were used successfully to forecast
O3 levels to ensure that specialized measurements were made on appropriate days.
3.7.3 Precursors
3.7.3.1 Volatile Organic Compound Emissions
Hundreds of VOCs, commonly containing from 2 to about 12 carbon atoms, are
emitted by evaporative and combustion processes from a large number of source types. Total
U.S. VOC emissions in 1991 were estimated at 21.0 Tg. The two largest source categories
were industrial processes (10.0 Tg) and transportation (7.9 Tg). Emissions of VOCs from
highway vehicles accounted for almost 75% of the transportation-related emissions; studies
have shown that the majority of these VOC emissions come from about 20% of the
automobiles in service, many, of which are older cars that are poorly maintained.
The accuracy of VOC emission estimates is difficult to determine, both for
stationary and mobile sources. Within major area sources, deviations of emission rates from
individual sources from assigned average factors can result in error for the entire area source.
Evaporative emissions, which depend on temperature and other environmental factors,
compound the difficulties of assigning accurate emission factors. In assigning VOC emission
estimates to the mobile source category, models are used that incorporate numerous input
parameters (e.g., type of fuel used, type of emission controls, age of vehicle), each of which
has some degree of uncertainty.
According to recent studies, vegetation emits significant quantities of VOCs into the
atmosphere, chiefly monoterpenes and isoprene, but also oxygenated VOCs. The most recent
biogenic VOC emissions estimate for the United States showed annual emissions of 29.1
Tg/year. Coniferous forests are the largest vegetative contributor on a national basis, because
of their extensive land coverage. Summertime biogenic emissions comprise more than half of
the annual totals in all regions because of their dependence on temperature and vegetational
growth. Biogenic emissions are, for those reasons, expected to be higher in the southern states
than in the northern.
Uncertainties in both biogenic and anthropogenic VOC emission inventories prevent
establishing the relative contributions of these two categories.
3.7.3.2 Nitrogen Oxides Emissions
Anthropogenic NOX is associated with combustion processes. The primary pollutant
emitted is NO, formed at high combustion temperatures from the nitrogen and oxygen in air
and from nitrogen in combustion fuel. Emissions of NOX in 1991 in the United States totaled
21.39 Tg. The two largest NOX emission sources are electric power generation plants and
highway vehicles. Emissions of NOX therefore are highest in areas having a high density of
electric-power-generating stations and in urban regions having high traffic densities. Between
1987 and 1991, transportation-related emissions remained essentially constant, whereas
stationary source NOX emissions increased about 10%.
Natural NOX sources include stratospheric intrusion, oceans, lightning, soils, and
wildfires. Lightning and soil emission are the only two significant natural sources of NOX in
the United States. The estimated annual lightning-produced NOX for the continental United
States is Dl.O Tg, about 60% of which is generated over the southern states. Both nitrifying
3-169
-------
and denitrifying organisms in the soil can produce NOX, principally NO. Emission rates depend
mainly on fertilization levels and soil temperature. Inventorying soil NOX emissions is difficult
because of large temporal and spatial variability, but the nationwide total has been estimated at
1.2 Tg/year, of which about 85% is emitted in spring and summer. About 60% of the total
soil NOX is emitted in the area of the country containing the central corn belt.
Combined natural sources contribute about 2.2 Tg of NOX to the troposphere over
the continental United States. Uncertainties in natural NOX inventories are much larger than
that for anthropogenic NOX emissions. Because a large proportion of anthropogenic NOX
emissions come from distinct point sources, published annual estimates are thought to be very
reliable.
3.7.3.3 Concentrations of Volatile Organic Compounds in Ambient Air
The VOCs most frequently analyzed in ambient air are NMHCs. Morning
concentrations (6:00 a.m. to 9:00 a.m.) have been measured most often because of the use of
morning data in EKMA and in air quality simulation models. Major field studies in 22 cities
in 1984 and in 19 cities in 1985 produced NMHC measurements that showed median values
ranging from 0.39 to 1.27 ppmC for 1984 and 0.38 to 1.63 ppmC in 1985. Overall median
values from all urban sites were about 0.72 ppmC in 1984 and 0.60 ppmC in 1985.
Comparative data over two decades (the 1960s through the 1980s) in the Los
Angeles and New York City areas showed decreases in NMHC concentrations in those areas.
Concomitant compositional changes were observed over the two decades, with increases
observed in the percentage of alkanes and decreases in the percentage of aromatic
hydrocarbons and acetylene.
Concurrent measurements of anthropogenic and biogenic NMHCs have shown that
biogenic NMHCs usually constituted much less than 10% of the total NMHCs. For example,
average isoprene concentrations ranged from 0.001 to 0.020 ppmC and terpenes from 0.001 to
0.030 ppmC.
3.7.3.4 Concentrations of Nitrogen Oxides in Ambient Air
Measurements of NOX at sites in 22 and 19 U.S. cities in 1984 and 1985,
respectively, showed that median NOX concentrations ranged from 0.02 to 0.08 ppm in most of
these cities. The 6 a.m. to 9 a.m. median concentrations in many of these cities exceeded the
annual average NOX values of 0.02 to 0.03 ppm found in U.S. metropolitan areas between
1980 and 1989. Nonurban NOX concentrations, reported as average seasonal or annual NOX,
range from < 0.005 to 0.015 ppm.
Ratios of 6 a.m. to 9 a.m. NMOC to NOX are higher in southeastern and
southwestern U.S. cities than in northeastern and midwestern U.S. cities, according to data
from EPA's multi-city studies conducted in 1984 and 1985. Median ratios ranged from 9.1 to
37.7 in 1984; in 1985, median ratios ranged from 6.5 to 53.2 in the cities studied. Rural
NMOC/NOX ratios tend to be higher than urban ratios. Morning (6 a.m. to 9 a.m.)
NMOC/NOX ratios are used in the EKMA-type of trajectory model. Trends from 1976 to 1990
show decreases in these ratios in the South Coast Air Basin of California. The correlation of
NMOC/NOX ratios with maximum 1-h O3 concentrations, however, was weak in a recent
analysis.
3-170
-------
3.7.3.5 Ratios of Concentrations of Nonmethane Organic Compounds to Nitrogen
Oxides
The ratios of NMOC/NOX vary substantially between cities and within a given city.
With certain exceptions, urban NMOC/NOX ratios have been in the range of 10 and below.
In contrast, ratios of NMOC/NOX in rural areas tend to equal or exceed 20. Discrepancies
have been found between ambient NMOC/NOX ratios and emission inventory NMOC/NOX
ratios, with ambient ratios of NMOC/NOX significantly exceeding emission ratios of
NMOC/NOX.
Trends in ratios of NMOC/NOX have shown downward trends to well below
10 during the 1980s, both for the South Coast Air Basin and for cities in the eastern United
States. Based on these low ratios, hydrocarbon control should be more effective than NOX
control within a number of cities.
3.7.3.6 Source Apportionment and Reconciliation
Source apportionment (now regarded as synonymous with receptor modeling) refers
to determining the quantitative contributions of various sources of VOCs to ambient air
pollutant concentrations. Source reconciliation refers to the comparison of measured ambient
VOC concentrations with emissions inventory estimates of VOC source emission rates for the
purpose of validating the inventories.
Early studies in Los Angeles employing a "mass balance" approach to receptor
modeling showed the following estimated contributions of respective sources to ambient air
concentrations of NMOCs through C10: automotive exhaust, 53%; whole gasoline
evaporation, 12%; gasoline headspace vapor, 10%; commercial natural gas, 5%; geogenic
natural gas, 19%; and liquefied natural gas, 1%. Recent studies in eight U.S. cities showed
that vehicle exhaust was the dominant contributor to ambient VOCs (except in Beaumont,
where 14% was reported). Estimates of the contributions of gasoline evaporation differ in
methodology; the more appropriate methods used result in estimates of large whole gasoline
contributions (i.e., equal to vehicle exhaust in one study and 20% of vehicle exhaust in a
second study).
The chemical mass balance approach used for estimating anthropogenic VOC
contributions to ambient air cannot be used for receptor modeling of biogenic sources.
A modified approach, applied to 1990 data from a downtown site in Atlanta, indicated a lower
limit of 2% (24-h average) for the biogenic percentage of total ambient VOCs at that location
(isoprene was used as the biogenic indicator species). The percentage varies during the 24-h
period because of the diurnal (e.g., temperature, light intensity) dependence of isoprene
concentrations.
Source reconciliation data have shown disparities between emission inventory
estimates and receptor-estimated contributions. For biogenics, emission estimates are greater
than receptor-estimated contributions. The reverse has been true for natural gas contributions
estimated for Los Angeles, Columbus, and Atlanta and for refinery emissions in Chicago.
3.7.4 Analytical Methods for Oxidants and Their Precursors
3.7.4.1 Oxidants
Current methods used to measure O3 are CL, UV absorption spectrometry, and
newly developed spectroscopic and chemical approaches, including chemical approaches
applied to passive sampling devices for O3.
3-171
-------
The CL method, designated as the reference method by EPA, involves the direct
gas-phase reaction of O3 with an alkene (C2H4) to produce electronically excited products,
which decay with the emission of light. Detection limits of 0.005 ppm and a response time of
less than 30 s are typical of currently available commercial instruments. A positive
interference from atmospheric water vapor was reported in the 1970s and has recently been
confirmed. Proper calibration can minimize this source of error.
Commercial UV photometers for measuring O3 have detection limits of about
0.005 ppm, long-term precision within about + 5%, and a response time of < 1 min. Ozone
has a fairly strong absorption band with a maximum near 254 nm; its molar absorption
coefficient at that wavelength is well known. Because the measurement is absolute, UV
photometry also is used to calibrate other O3 methods.
A potential disadvantage of UV photometry is that atmospheric constituents that
absorb 254-nm radiation (and that are removed fully or partially by the MnO2 scrubber used in
UV O3 photometers) will be positive interferences in O3 measurements. Interferences have
been reported in two recent studies but assessment of the potential importance of such
interferences (e.g., toluene, styrene, cresols, nitrocresols) is hindered by lack of absorption
spectra data in the 250-nm range and by lack of ambient measurements of most of the aromatic
photochemical reaction products. An interference from water also appears to occur from
condensation of moisture in sampling level. Results from collocated UV and CL instruments
indicated positive biases in the UV data of 20 to 40 ppb on hot, humid days.
Differential optical absorption spectrometry has been used to measure ambient O3,
but further intercomparisons with other methods and interference tests are recommended.
Passive sampling devices permit acquisition of personal human exposure data and of
O3 monitoring data in areas where the use of instrumental methods is not feasible. Three PSDs
are commercially available; all employ solid absorbents that react with O3.
Calibration of O3 measurement methods (other than PSDs) is done by UV
spectrometry or by GPT of O3 with NO. Ultraviolet photometry is the reference calibration
method approved by EPA. Ozone is unstable and must be generated in situ at time of use to
produce calibration mixtures.
Two methods generally have been employed to measure atmospheric PAN and its
higher homologues: IR and GC using an BCD. A third method, less often used, couples GC
with a molybdenum converter that reduces PAN to NO in the gas phase and subsequently
measures the NO with a CL analyzer. Peroxyacetyl nitrate and the higher PANs are normally
measured by GC-ECD. Detection limits have been extended to 1 to 5 ppt, using cryogenic
enrichment of samples and specified desorption procedures that limit losses associated with
cryosampling. Because PAN is unstable (explosive, and subject to surface-related
decomposition), the preparation of reliable calibration standards is difficult. Methods devised
to generate calibration standards include photolysis of static concentrations of gases, nitration
of peracetic acid in single hydrocarbons, and analysis of PAN as NO under specified
conditions of the dissociation of PAN into its precursors.
Early measurements of 10 to 80 ppb H2O2 reported in the 1970s have been found to
be in error because of artifact formation of H2O2 from reactions of absorbed gaseous O3.
Modeling results also indicate that lower levels of H2O2, on the order of 1 ppb, occur in the
atmosphere.
In situ measurement methods for H2O2 include FTIR and TOLAS. The FTIR
method is specific for H2O2 but has a high detection level of D50 ppb (using a 1-km path
3-172
-------
length). The TDLAS method also is specific and has a detection level of 0.1 ppb over
averaging times of several minutes. Four frequently used wet chemical methods for
measurement of H2O2 are available. All involve the oxidation of a substrate followed by
instrumental detection and quantification of the resulting CL or fluorescence. Detection limits
are comparable to those of FTIR and TDLAS, but interferences are common and must be
obviated or minimized with specified procedures.
Calibration of methods for gaseous H2O2 measurement requires the immediate use
of standard mixtures prepared by one of several wet chemical methods.
3.7.4.2 Volatile Organic Compounds
Increased monitoring of VOCs is required under Title I, Section 182, of the CAAA
of 1990 because of the role of VOCs as precursors to the formation of O3 and other
photochemical oxidants. Volatile organic compounds are those gaseous organic compounds
that have a vapor pressure greater than 0.15 mm and, generally, have a carbon content ranging
from C, through C12.
Traditionally, NMHCs have been measured by methods that employ a FID as the
sensing element that measures a change in ion intensity resulting from the combustion of air
containing organic compounds. The method recommended by EPA for total NMOC
measurement involves the cryogenic preconcentration of nonmethane organic compounds and
the measurement of the revolatilized NMOCs using FID. The main technique for speciated
NMOC/NMHC measurements is cryogenic preconcentration followed by GC-FID. Systems
for sampling and analysis of VOCs have been developed that require no liquid cryogen for
operation, yet provide sufficient resolution of species.
Stainless steel canisters have become the containers of choice for collection of
whole-air samples for NMHC/NMOC data. Calibration procedures for NMOC
instrumentation require the generation, by static or dynamic systems, of dilute mixtures at
concentrations expected to occur in ambient air.
Preferred methods for measuring carbonyl species (aldehydes and ketones) in
ambient air are spectroscopic methods, on-line colorimetric methods, and HPLC method
employing DNPH derivatization in a silica gel cartridge. The most common method in current
use for measuring aldehydes in ambient air is the HPLC-DNPH method. Use of an O3 scrubber
has been recommended to prevent interference in this method by O3 in ambient air. Carbonyl
species are reactive, making preparation of stable calibration mixtures difficult; but several
methods are available.
Impetus for the development of methods for measuring the more reactive oxygen-
and nitrogen-containing organic compounds has come from their roles as precursors or
products of photochemical oxidation and also from the inclusion of many of these compounds
on the list of hazardous air pollutants in the 1990 CAAA. Measurement of these PVOCs is
difficult because of their reactivity and water solubility. Methods are still in development.
3.7.4.3 Oxides of Nitrogen
Nitric oxide and NO2 comprise the NOX involved as precursors to O3 and other
photochemical oxidants.
The most common method of NO measurement is the gas-phase CL reaction with
O3. The CL method is essentially specific for NO. Commercial NO monitors have detection
limits of a few parts per billion by volume in ambient air. Commercial NO analyzers may not
3-173
-------
have sensitivity sufficient for surface measurements in rural or remote areas or for airborne
measurements. Direct spectroscopic methods for NO exist that have very high sensitivity and
selectivity for NO. Major drawbacks of these methods are their complexity, size, and cost,
which restrict these methods to research applications. No PSDs exist for measurement of NO.
Chemiluminescence analyzers are the method of choice for NO2 measurement, even
though they do not measure NO2 directly. Minimum detection levels for NO2 have been
reported to be 5 to 13 ppb, but more recent evaluations have indicated detection limits of 0.5 to
1 ppbv. Reduction of NO2 to NO is required for measurement. In practice, selective
measurement of NOX by this approach has proved difficult. Commercial instruments that use
heated catalytic converters to reduce NO2 to NO measure not NO and NOX, but more nearly
NO and total NO . Thus, the NO2 value inferred from such measurements may be
significantly in error, which may in turn affect the results of modeling of ambient O3.
Several spectroscopic approaches to NO2 detection have been developed. As noted
above for NO, however, these methods have major drawbacks that include their complexity,
size, and cost, which, at present, outweigh the advantages of their sensitivity and selectivity.
Passive samplers for NO2 exist but are still in the developmental stage for ambient air
monitoring.
Calibration of methods for NO measurement is done using standard cylinders of NO
in nitrogen. Calibration of methods for NO2 measurement include the use of cylinders of NO2
in nitrogen or air, the use of permeation tubes, and gas-phase titration.
3.7.5 Ozone Air Quality Models
3.7.5.1 Definitions, Descriptions, and Uses
Photochemical air quality models are used to predict how O3 concentrations change
in response to prescribed changes in source emissions of NOX and VOCs. They are
mathematical descriptions of the atmospheric transport, diffusion, removal, and chemical
reactions of pollutants. They operate on sets of input data that characterize the emissions,
topography, and meteorology of a region and produce outputs that describe air quality in that
region.
Two kinds of photochemical models are recommended in guidelines issued by EPA:
(1) the grid-based UAM is recommended for modeling O3 over urban areas, and (2) EKMA is
identified as an acceptable approach under certain circumstances. The 1990 CAAA mandate
the use of three-dimensional (grid-based) air quality models such as UAM in developing SIPs
for areas designated as extreme, severe, serious, or multistate moderate.
In grid-based air quality models, the region to be modeled (the modeling domain) is
subdivided into a three-dimensional array of grid cells. Pertinent atmospheric processes and
chemical reactions are represented for each cell.
In trajectory models, such as EKMA, a hypothetical air parcel moves through the
area of interest along a path calculated from wind trajectories. Emissions are injected into the
air parcel and undergo vertical mixing and chemical transformations. Trajectory models
provide a dynamic description of atmospheric source-receptor relationships that is simpler and
less expensive to derive than that obtained from grid models, but meterological processes are
highly simplified in trajectory models
The EKMA-based method for determining O3 control strategies has some
limitations, the most serious of which is that predicted emissions reductions are critically
dependent on the initial NMHC/NOX ratio used in the calculations. This ratio cannot be
3-174
-------
determined with any certainty because it is expected to be quite variable in time and space in
an urban area. Grid-based models have their limitations as well. These are pointed out
subsequently.
3.7.5.2 Model Components
Spatial and temporal characteristics of VOC and NOX emissions are major inputs to
a photochemical air quality model. Greater accuracy in emissions inventories is needed, for
biogenics and for both mobile and stationary source components. Grid-based air quality
models also require as input the three-dimensional wind field for the photochemical episode
being simulated. This input is supplied by "meteorological modules" which fall into one of
four categories: (1) objective analysis procedures; (2) diagnostic methods; (3) dynamic, or
prognostic, methods; and (4) hybrid methods that embody elements from both diagnostic and
prognostic approaches. Prognostic models are believed to provide a dynamically consistent,
physically realistic, three-dimensional representation of the wind and other meteorological
variables at scales of motion not resolvable by available observations. Outputs of prognostic
models do not always agree with observational data, but methods have been devised to mitigate
these problems.
A chemical kinetic mechanism (a set of chemical reactions), representing the
important reactions that occur in the atmosphere, is used in an air quality model to estimate the
net rate of formation of each pollutant simulated as a function of time. Chemical mechanisms
that explicitly treat each individual VOC component of ambient air are too lengthy to be
incorporated into three-dimensional atmospheric models. " Lumped" mechanisms are therefore
used. The chemical mechanisms used in existing photochemical O3 models contain
uncertainties that may limit the accuracy of their predictions. Because of different approaches
to "lumping" of reactions, models can produce somewhat different results under similar
conditions. Both the UAM (UAM-IV) and EPA's ROM use the CMB-IV. The CBM-IV and
the SAPRC and RADM mechanisms are considered to represent the state of the science.
Dry deposition, the removal of chemical species from the atmosphere by interaction
with ground-level surfaces, is an important removal process for O3 on both urban and regional
scales; and is included in all urban- and regional-scale models. Wet deposition (the removal of
gases and particles from the atmosphere by precipitation events) generally is not included in
urban-scale photochemical models, because O3 episodes do not occur during periods of
significant clouds or rain.
Concentration fields of all species computed by the model must be specified at the
beginning of the simulation; these concentration fields are called the initial conditions. These
initial conditions are determined mainly with ambient measurements, either from routinely
collected data or from special studies, but interpolation can be used to distribute the surface
ambient measurements.
3.7.5.3 Evaluation of Model Performance
Air quality models are evaluated by comparing their predictions with ambient
observations. An adequate model should give accurate predictions of current peak
O3 concentrations and temporal and spatial O3 patterns. It also should respond accurately to
changes in VOC and NOX emissions, to differences in VOC reactivity, and to spatial and
temporal changes in emissions patterns for future years. Likewise, multispecies comparisons
could be the key in discriminating among alternative modeling approaches that provide similar
3-175
-------
predictions of O3 concentrations. Adequate model performance for several reactive species
increases the assurance that correct O3 predictions are not a result of chance or fortuitous
cancellation of errors introduced by various assumptions.
If only a routine database is available for modeling O3 in an urban area, then several
concerns require attention relative to model performance evaluation: air quality aloft,
boundary conditions, ambient VOC data, and meteorological data aloft. If any of these four
areas is missing from the database, the performance evaluation and subsequent model
application must be adequately planned to minimize the possibility of compensatory errors.
3.7.5.4 Use of Ozone Air Quality Model for Evaluating Control Strategies
Photochemical air quality models are used for control strategy evaluation by first
demonstrating that a past episode, or episodes, can be simulated adequately and then reducing
hydrocarbon or NOX emissions, or both, in the model inputs and assessing the effects of these
reductions on O3 in the region. The adequacy of control strategies based on grid-based models
depends, in part, on the nature of input data for simulations and model validation, on input
emissions inventory data, and on the relationship between model output and the current form
of the NAAQS for O3.
Grid-based models that have been widely used to evaluate control strategies for
O3 or acid deposition, or both, are the UAM, the CIT model, the ROM, the ADOM, and the
RADM.
3.7.5.5 Conclusions
Urban air quality models are becoming readily available for application and have
been applied in recent years in several urban areas. Significant progress also has been made in
the development of regional models and in the integration of state-of-the-art prognostic
meteorological models as drivers.
There are still many uncertainties in photochemical air quality modeling. Prime
among these are emission inventories. However, models are essential for regulatory analysis
and constitute one of the major tools for attacking the O3 problem. Grid-based O3 air quality
modeling is superior to the available alternatives for O3 control planning, but the chances of its
incorrect use must be minimized.
3-176
-------
References
Allen, A. O.; Hochanadel, C. J.; Ghormley, J. A.; Davis, T. W. (1952) Decomposition of water and aqueous
solutions under mixed fast neutron and gamma radiation. J. Phys. Chem. 56: 575-586.
Altshuller, A. P. (1983a) Measurements of the products of atmospheric photochemical reactions in laboratory
studies and in ambient air-relationships between ozone and other products. Atmos. Environ.
17: 2383-2427.
Altshuller, A. P. (1983b) Review: natural volatile organic substances and their effect on air quality in the United
States. Atmos. Environ. 17: 2131-2165.
Altshuller, A. P. (1986) The role of nitrogen oxides in nonurban ozone formation in the planetary boundary layer
over N America, W Europe and adjacent areas of ocean. Atmos. Environ. 20: 245-268.
Altshuller, A. P. (1987) Estimation of the natural background of ozone present at surface rural locations. JAPCA
37: 1409-1417.
Altshuller, A. P. (1988) Some characteristics of ozone formation in the urban plume of St. Louis, MO. Atmos.
Environ. 22: 499-510.
Altshuller, A. P. (1989a) Sources and levels of background ozone and its precursors and impact at ground level.
In: Schneider, T.; Lee, S. D.; Wolters, G. J. R.; Grant, L. D., eds. Atmospheric ozone research
and its policy implications: proceedings of the 3rd US-Dutch international symposium; May 1988;
Nijmegen, The Netherlands. Amsterdam, The Netherlands: Elsevier Science Publishers;
pp. 127-157. (Studies in environmental science 35).
Altshuller, A. P. (1989b) Nonmethane organic compound to nitrogen oxide ratios and organic composition in cities
and rural areas. JAPCA 39: 936-943.
Altshuller, A. P. (1993) Production of aldehydes as primary emissions and from secondary atmospheric reactions
of alkenes and alkanes during the night and early morning hours. Atmos. Environ. Part A 27: 21-32.
Altshuller, A. P.; Bufalini, J. J. (1971) Photochemical aspects of air pollution: a review. Environ. Sci. Technol.
5: 39-64.
Altshuller, A. P.; Lefohn, A. S. (1996) Background ozone in the planetary boundary layer over the United States.
J. Air Waste Manage. Assoc. 46: 134-141.
Altshuller, A. P.; Leng, L. J. (1963) Application of the 3-methyl-2-benzothiazolone hydrazone method for
atmospheric analysis of aliphatic aldehydes. Anal. Chem. 35: 1541-1542.
Altshuller, A. P.; McPherson, S. P. (1963) Spectrophotometric analysis of aldehydes in the Los Angeles
atmosphere. J. Air Pollut. Control Assoc. 13: 109-111.
Altshuller, A. P.; Miller, D. L.; Sleva, S. F. (1961) Determination of formaldehyde in gas mixtures by the
chromotropic acid method. Anal. Chem. 33: 621-625.
Anderson, J. G.; Toohey, D. W.; Brune, W. H. (1991) Free radicals within the Antarctic vortex: the role of
CFCs in Antarctic ozone loss. Science (Washington, DC) 251: 39-46.
Andersson-Skold, Y.; Grennfelt, P.; Pleijel, K. (1992) Photochemical ozone creation potentials: a study of
different concepts. J. Air Waste Manage. Assoc. 42: 1152-1158.
3-177
-------
Andreae, W. A. (1955) A sensitive method for the estimation of hydrogen peroxide in biological materials. Nature
(London) 175: 859-860.
Anthes, R. A.; Warner, T. T. (1978) Development of hydrodynamic models suitable for air pollution and other
mesometeorological studies. Mon. Weather Rev. 106: 1045-1078.
Anthes, R. A.; Kuo, Y.-H.; Baumhefner, D. P.; Errico, R. M.; Bettge, T. W. (1985) Predictability of mesoscale
atmospheric motions. Adv. Geophys. 28B: 159-202.
Anthes, R. A.; Hsie, E. Y.; Kao, Y. H. (1987) Description of the Penn State/NCAR mesoscale model
version 4 (MM4). Boulder, CO: National Center for Atmospheric Research; NCAR technical note
282.
Aoyanagi, S.; Mitsumisha, H. (1985) Determination of hydrogen peroxide by chemiluminescence. Jpn. Kokai
Tokkyo Koho 62-123336 (CA 107: 93083V, 1987).
Arey, J.; Atkinson, R.; Aschniann, S. M. (1990) Product study of the gas-phase reactions of nionoterpenes with
the OH radical in the presence of NO,. J. Geophys. Res. [Atmos.] 95: 18,539-18,546.
Arey, J.; Winer, A. M.; Atkinson, R.; Aschmann, S. M.; Long, W. D.; Morrison, C. L. (1991a) The emission
of (Z)-S-hexen-l-ol, (Z)-3-hexenylacetate and other oxygenated hydrocarbons from agricultural plant
species. Atmos. Environ. Part A 25: 1063-1075.
Arey, J.; Winer, A. M.; Atkinson, R.; Aschmann, S. M.; Long, W. D.; Morrison, C. L.; Olszyk, D. M. (1991b)
Terpenes emitted from agricultural species found in California's Central Valley. J. Geophys. Res.
[Atmos.] 96: 9329-9336.
Armstrong, W. A.; Humphreys, W. G. (1965) A L.E.T. independent dosimeter based on the chemiluminescent
determination of H2O2. Can. J. Chem. 43: 2576-2584.
Arnts, R. R.; Tejada, S. B. (1989) 2,4-Dinitrophenylhydrazine-coated silica gel cartridge method for
determination of formaldehyde in air: identification of an ozone interference. Environ. Sci. Technol.
23: 1428-1430.
Aronian, P. P.; Scheff, P. A.; Wadden, R. A. (1989) Wintertime source-reconciliation of ambient organics.
Atmos. Environ. 23: 911-920.
Atkinson, R. (1988) Atmospheric transformations of automotive emissions. In: Watson, A. Y.; Bates, R. R.;
Kennedy, D., eds. Air pollution, the automobile, and public health. Washington, DC: National
Academy Press; pp. 99-132.
Atkinson, R. (1989) Kinetics and mechanisms of the gas-phase reactions of the hydroxyl radical with organic
compounds. Washington, DC: American Chemical Society. (J. Phys. Chem. Ref. Data Monograph
no. 1).
Atkinson, R. (1990) Gas-phase tropospheric chemistry of organic compounds: a review. Atmos. Environ.
Part A 24: 1-41.
Atkinson, R. (1991) Kinetics and mechanisms of the gas-phase reactions of the NO, radical with organic
compounds. J. Phys. Chem. Ref. Data 20: 459-507.
Atkinson, R. (1994) Gas-phase tropospheric chemistry of organic compounds. Washington, DC: American
Chemical Society. (J. Phys. Chem. Ref. Data Monograph no. 2).
3-178
-------
Atkinson, R. ; Aschmann, S. M. (1993) OH radical production from the gas-phase reactions of 03 with a series of
alkenes under atmospheric conditions. Environ. Sci. Technol. 27: 1357-1363.
Atkinson, R. ; Aschmann, S. M. (1994) Products of the gas-phase reactions of aromatic hydrocarbons: effect of
NO2 concentration. Int. J. Chem. Kinet. 26: 929-944.
Atkinson, R. ; Carter, W. P. L. (1984) Kinetics and mechanisms of the gas-phase reactions of ozone with organic
compounds under atmospheric conditions. Chem. Rev. 84: 437-470.
Atkinson, R. ; Carter, W. P. L. (1991) Reactions of alkoxy radicals under atmospheric conditions: the relative
importance of decomposition versus reaction with O2. J. Atmos. Chem. 13: 195-210.
Atkinson, R. ; Winer, A. M.; Pitts, J. N., Jr. (1986) Estimation of night-time tyOj concentrations from ambient
NO2 and NO3 radical concentrations and the role of N2O5 in night-time chemistry. Atmos. Environ.
20: 331-339.
Atkinson, R.; Winer, A. M.; Arey, J.; Bierniann, H. W.; Dinoff, T.; Harger, W. P.; McElroy, P. A.;
Tuazon, E. C.; Zielinska, B. (1988) Measurements of NOj, HONO, NO,, HCHO, PAH, nitroarenes
and particulate mutagenic activities during the carbonaceous species methods comparison study. Final
report. Sacramento, CA: California Air Resources Board; contract no. A5-150-32.
Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. P., Jr.; Kerr, J. A.; Troe, J. (1992a) Evaluated kinetic
and photochemical data for atmospheric chemistry: supplement IV, IUPAC Subcommittee on Gas
Kinetic Data Evaluation for Atmospheric Chemistry. J. Phys. Chem. Ref. Data 21: 1125-1568.
Atkinson, R. ; Aschmann, S. M.; Arey, J.; Shorees, B. (1992b) Formation of OH radicals in the gas phase
reactions of O3 with a series of terpenes. J. Geophys. Res. [Atmos.] 97: 6065-6073.
Atwater, M. A. (1984) Influence of meteorology on high ozone concentrations. In: Evaluation of the scientific
basis for ozone/oxidants standards: proceedings of an international specialty conference; November;
Houston, TX. Pittsburgh, PA: Air Pollution Control Association.
Ayers, G. P.; Penkett, S. A.; Gillett, R. W.; Bandy, B.; Galbally, I. E.; Meyer, C. P.; Elsworth, C. M.;
Bentley, S. T.; Forgan, B. W. (1992) Evidence for photochemical control of ozone concentrations in
unpolluted marine air. Nature (London) 360: 446-449.
Barnes, I.; Bastian, V.; Becker, K. H.; Tong, Z. (1990) Kinetics and products of the reactions of NQ with
monoalkenes, dialkenes, and nionoterpenes. J. Phys. Chem. 94: 2413-2419.
Baugues, K. (1986) A review of NMOC, NOX and NMOC/NOX ratios measured in 1984 and 1985. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards; report no. EPA-450/4-86-015. Available from: NTIS, Springfield, VA;
PB87-166963/HSU.
Becker, K. H.; Brockmann, K. J.; Bechara, J. (1990) Production of hydrogen peroxide in forest air by reaction of
ozone with terpenes. Nature (London) 346: 256-258.
Becker, K. H.; Bechara, J.; Brockmann, K. J. (1993) Studies on the formation of HjOj in the ozonolysis of
alkenes. Atmos. Environ. Part A 27: 57-61.
Benjamin, S. G.; Seaman, N. L. (1985) A simple scheme for objective analysis in curved flow. Mon. Weather
Rev. 113: 1184-1198.
Bersis, D.; Vassiliou, E. (1966) A chemiluminescence method for determining ozone. Analyst (London)
91: 499-505.
3-179
-------
Betterton, E. A.; Hoffmann, M. R. (1988) Henry's law constants of some environmentally important aldehydes.
Environ. Sci. Technol. 22: 1415-1418.
Bidleman, T. F. (1988) Atmospheric processes. Environ. Sci. Technol. 22: 361-367.
Biermann, H. W.; Tuazon, E. C.; Winer, A. M.; Wallington, T. J.; Pitts, J. N., Jr. (1988) Simultaneous absolute
measurements of gaseous nitrogen species in urban ambient air by long pathlength infrared and
ultraviolet-visible spectroscopy. Atmos. Environ. 22: 1545-1554.
Blake, N. J.; Penkett, S. A.; Clemitshaw, K. C.; Anwyl, P.; Lightman, P.; Marsh, A. R. W.; Butcher, G. (1993)
Estimates of atmospheric hydroxyl radical concentrations from the observed decay of many reactive
hydrocarbons in well-defined urban plumes. J. Geophys. Res. [Atmos.] 98: 2851-2864.
Bradshaw, J.; Smyth, S.; Bai, R.; Sandholm, S. (1994) Recent advances in the design of NO, photolytic
converters. In: Measurement of toxic and related air pollutants: proceedings of the U.S.
EPA/A&WMA international symposium; May; Durham, NC. Pittsburgh, PA: Air & Waste
Management Association; p. 228. (A&WMA specialty conference VIP-39).
Bruckmann, P. W.; Willner, H. (1983) Infrared spectroscopic study of peroxyacetyl nitrate (PAN) and its
decomposition products. Environ. Sci. Technol. 17: 352-357.
Buhr, M. P.; Parrish, D. D.; Norton, R. B.; Fehsenfeld, F. C.; Sievers, R. E.; Roberts, J. M. (1990)
Contribution of organic nitrates to the total reactive nitrogen budget at a rural eastern U.S. site.
J. Geophys. Res. [Atmos.] 95: 9809-9816.
Burkhardt, M. R.; Maniga, N. I.; Stedman, D. H.; Paur, R. J. (1988) Gas chroniatographic method for measuring
nitrogen dioxide and peroxyacetyl nitrate in air without compressed gas cylinders. Anal. Chem. 60:
816-819.
Burns, W. P.; Tingey, D. T.; Evans, R. C.; Bates, E. H. (1983) Problems with a Nafion® membrane dryer for
drying chroniatographic samples. J. Chromatogr. 269: 1-9.
Busness, K. (1992) [Unpublished data]. Richland, WA: Pacific Northwest Laboratory.
Cadle, S. H.; Gorse, R. A.; Lawson, D. R. (1993) Real-world vehicle emissions: a summary of the third annual
CRC-APRAC on-road vehicle emissions workshop. Air Waste 43: 1084-1090.
Cadoff, B. C.; Hodgeson, J. (1983) Passive sampler for ambient levels of nitrogen dioxide. Anal. Chem.
55: 2083-2085.
California Air Resources Board. (1976) A study of the effect of atmospheric humidity on analytical oxidant
measurement methods. Presented at: 15th conference [on] methods in air pollution studies; January;
Long Beach, CA. Sacramento, CA: Air and Industrial Hygiene Laboratory.
California Air Resources Board. (1992) Analysis of the ambient VOC data collected in the southern California Air
Quality Study [final report]. Sacramento, CA: California Air Resources Board; contract
no. A8320-130.
Cardelino, C. A.; Chameides, W. L. (1990) Natural hydrocarbons, urbanization, and urban ozone. J. Geophys.
Res. [Atmos.] 95: 13,971-13,979.
3-180
-------
Cardin, D. B.; Lin, C. C. (1991) Analysis of selected polar and non-polar compounds in air using automated
2-dimensional chromatography. In: Measurement of toxic and related air pollutants: proceedings of
the 1991 U.S. EPA/A&WMA international symposium, v. 1; May; Durham, NC. Pittsburgh, PA:
Air & Waste Management Association; pp. 552-557. (A&WMA publication VIP-21).
Carmichael, G. R.; Peters, L. K.; Kitada, T. (1986) A second generation model for regional-scale
transport/chemistry/deposition. Atmos. Environ. 20: 173-188.
Carmichael, G. R.; Peters, L. K.; Saylor, R. D. (1991) The STEM-II regional scale acid deposition and
photochemical oxidant model—I. an overview of model development and applications. Atmos.
Environ. Part A 25: 2077-2090.
Carroll, M. A.; Hastie, D. R.; Ridley, B. A.; Rodgers, M. O.; Torres, A. L.; Davis, D. D.; Bradshaw, J. D.;
Sandholm, S. T.; Schiff, H. I.; Karecki, D. R.; Harris, G. W.; Mackay, G. I.; Gregory, G. L.;
Condon, E. P.; Trainer, M.; Hubler, G.; Montzka, D. D.; Madronich, S.; Albritton, D. L.; Singh,
H. B.; Beck, S. M.; Shiphani, M. C.; Bachmeier, A. S. (1990) Aircraft measurement of NO* over
the eastern Pacific and continental United States and implications for ozone production. J. Geophys.
Res. [Atmos.] 95: 10,205-10,233.
Carter, W. P. L. (1990) A detailed mechanism for the gas-phase atmospheric reactions of organic compounds.
Atmos. Environ. Part A 24: 481-518.
Carter, W. P. L. (1991) Development of ozone reactivity scales for volatile organic compounds. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure
Assessment Laboratory; report no. EPA-600/3-91/050. Available from: NTIS, Springfield, VA;
PB91-243386.
Carter, W. P. L. (1994) Development of ozone reactivity scales for volatile organic compounds. J. Air Waste
Manage. Assoc. 44: 881-899.
Carter, W. P. L.; Atkinson, R. (1987) An experimental study of incremental hydrocarbon reactivity. Environ. Sci.
Technol. 21: 670-679.
Carter, W. P. L.; Atkinson, R. (1989a) Alkyl nitrate formation from the atmospheric photooxidation of alkanes; a
revised estimation method. J. Atmos. Chem. 8: 165-173.
Carter, W. P. L.; Atkinson, R. (1989b) Computer modeling study of incremental hydrocarbon reactivity. Environ.
Sci. Technol. 23: 864-880.
Carter, W. P. L.; Lurmann, F. W. (1990) Evaluation of the RADM gas-phase chemical mechanism. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure
Assessment Laboratory; report no. EPA-600/3-90-001. Available from: NTIS, Springfield, VA;
PB90-164526/HSU.
Cassmassi, J. C.; Mitsutomi, S.; Shepherd, M. (1991) Three-dimensional wind fields for use in the urban airshed
model. In: Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the
environment: papers from an international conference; March 1990; Los Angeles, CA. Pittsburgh,
PA: Air & Waste Management Association; pp. 460-479. (A&WMA transactions series no. TR-19).
Chanieides, W. L.; Tan, A. (1981) The two-dimensional diagnostic model for tropospheric OH: an uncertainty
analysis. J. Geophys. Res. C: Oceans Atmos. 86: 5209-5223.
Chameides, W. L.; Lindsay, R. W.; Richardson, J.; Kiang, C. S. (1988) The role of biogenic hydrocarbons
in urban photochemical smog: Atlanta as a case study. Science (Washington, DC) 241: 1473-1475.
3-181
-------
Chameides, W. L.; Fehsenfeld, P.; Rodgers, M. O.; Cardelino, C.; Martinez, J.; Parrish, D.; Lonneman, W.;
Lawson, D. R.; Rasmussen, R. A.; Zimmerman, P.; Greenberg, J.; Middleton, P.; Wang, T. (1992)
Ozone precursor relationships in the ambient atmosphere. J. Geophys. Res. [Atmos.] 97: 6037-6055.
Chandler, A. S.; Choularton, T. W.; Bollard, G. J.; Eggleton, A. E. J.; Gay, M. J.; Hill, T. A.;
Jones, B. M. R.; Tyler, B. J.; Bandy, B. J.; Penkett, S. A. (1988) Measurements of H2O, and SO,
in clouds and estimates of their reaction rate. Nature (London) 336: 562-565.
Chang, T. Y.; Rudy, S. J. (1990) Ozone-forming potential of organic emissions from alternative-fueled vehicles.
Atmos. Environ. Part A 24: 2421-2430.
Chang, J. S.; Brost, R. A.; Isaken, I. S. A.; Madronich, S.; Middleton, P.; Stockwell, W. R.; Walcek, C. J.
(1987) A three-dimensional Eulerian acid deposition model: physical concepts and formulation.
J. Geophys. Res. [Atmos.] 92: 14,681-14,700.
Chang, T. Y.; Hammerle, R. H.; Japar, S. M.; Salmeen, I. T. (1991a) Alternative transportation fuels and air
quality. Environ. Sci. Technol. 25: 1190-1197.
Chang, J. S.; Middleton, P. B.; Stockwell, W. R.; Walcek, C. J.; Pleim, J. E.; Lansford, H. H.; Madronich, S.;
Binkowski, F. S.; Seaman, N. L.; Stauffer, D. R. (1991b) The regional acid deposition model and
engineering model. In: Irving, P. M., ed. Acidic deposition: state of science and technology,
volume I: emissions, atmospheric processes, and deposition. Washington, DC: The U.S. National
Acid Precipitation Assessment Program. (State of science and technology report no. 4).
Chapman, S. (1930) A theory of upper-atmospheric ozone. Mem. R. Meteorol. Soc. 3: 103-125.
Chock, D. P. (1985) A comparison of numerical methods for solving the advection equation—II. Atmos. Environ.
19: 571-586.
Chock, D. P. (1991) A comparison of numerical methods for solving the advection problem—III. Atmos. Environ.
Part A 25: 853-871.
Chock, D. P.; Kumar, S.; Herrmann, R. W. (1982) An analysis of trends in oxidant air quality in the south coast
air basin of California. Atmos. Environ. 16: 2615-2624.
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards,
L. W. (1993) PM,0 and PM2 5 compositions in California's San Joaquin Valley. Aerosol Sci. Technol.
18: 105-128.
Chu, S.-H.; Cox, W. M. (1993) Differences in regional ozone responses to precursor reductions as demonstrated
in a regional oxidant model. Presented at: AMS conference on the role of meteorology in managing
the environment in the 1990's; January; Scottsdale, AZ. Boston, MA: American Meteorological
Society.
3-182
-------
Chu, S.-H.; Meyer, E. L.; Cox, W. M.; Scheffe, R. D. (1993) The response of regional ozone to VOC and NQ
emissions reductions: an analysis for the eastern United States based on regional oxidant modeling.
In: Vostal, J. J., ed. Tropospheric ozone: nonattainment and design value issues [proceedings of a
U.S. EPA/A&WMA international specialty conference]; October 1992; Boston, MA. Pittsburgh,
PA: Air & Waste Management Association; pp. 305-315. (A&WMA transactions series no. 23).
Ciccioli, P.; Brancaleoni, E.; Frattoni, M.; Cecinato, A.; Brachetti, A. (1993) Ubiquitous occurrence of
semi-volatile carbonyl compounds in tropospheric samples and their possible sources. Atnios.
Environ. Part A 27: 1891-1901.
Cicerone, R. J. (1989) Analysis of sources and sinks of atmospheric nitrous oxide (HO). J. Geophys. Res.
[Atmos.] 94: 18,265-18,271.
Clark, T. L.; Clarke, J. F. (1984) A Lagrangian study of the boundary layer transport of pollutants in the
northeastern United States. Atmos. Environ. 18: 287-297.
Clark, T. L.; Karl, T. R. (1982) Application of prognostic meteorological variables to forecasts of daily maximum
one-hour ozone concentrations in the northeastern United States. J. Appl. Meteorol. 21: 1662-1671.
Code of Federal Regulations. (1994) Subpart B—procedures for testing performance characteristics of automated
methods SO2, CO, O3, and NO2. C. F. R. 40: §53.20.
Code of Federal Regulations. (1987) Ambient air monitoring reference and equivalent methods. C. F. R. 40: §53.
Cohen, I. R.; Purcell, T. C. (1967) Spectrophotometric determination of hydrogen peroxide with 8-quinolinol.
Anal. Chem. 39: 131-132.
Cohen, I. R.; Purcell, T. C.; Altshuller, A. P. (1967) Analysis of the oxidant in photooxidation reactions.
Environ. Sci. Technol. 1: 247-252.
Cohn, R. D.; Dennis, R. L. (1994) The evaluation of acid deposition models using principal component spaces.
Atmos. Environ. 28: 2531-2543.
Colbeck, I.; Harrison, R. M. (1985) Dry deposition of ozone: some measurements of deposition velocity and of
vertical profiles to 100 metres. Atmos. Environ. 19: 1807-1818.
Comes, F. J.; Armerding, W.; Grigonis, R.; Herbert, A.; Spiekermann, M.; Walter, J. (1992) Tropospheric OH:
local measurements and their interpretations. Ber. Bunsen-Ges. Phys. Chem. 96: 284-286.
Cotton, W. R.; Anthes, R. A. (1989) Storm and cloud dynamics. New York, NY: Academic Press, Inc.
(Dniowska, R.; Holton, J. R., eds. International geophysics series: v. 44).
Coutant, R. W. (1993) Theoretical evaluation of stability of volatile organic chemicals and polar volatile organic
chemicals in canisters. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/R-94/030.
Available from: NTIS, Springfield, VA; PB94-159902.
Cox, R. A. (1974) The photolysis of gaseous nitrous acid. J. Photochem. 3: 175-188.
Cox, R. A.; Roffey, M. J. (1977) Thermal decomposition of peroxyacetylnitrate in the presence of nitric oxide.
Environ. Sci. Technol. 11: 900-906.
Cox, R. D.; Balfour, W. D.; Langley, G. J. (1982) Quality control for ambient level hydrocarbon sampling and
analysis. Presented at: 75th annual meeting of the Air Pollution Control Association; June; New
Orleans, LA. Pittsburgh, PA: Air Pollution Control Association; paper no. 82-23.2.
3-183
-------
Crutzen, P. J. (1970) The influence of nitrogen oxides on the atmospheric ozone content. Q. J. R. Meteorol. Soc.
96: 320-325.
Dabdub, D.; Seinfeld, J. H. (1994) Numerical advection schemes used in air quality models—sequential and
parallel implementation. Atmos. Environ. 28: 3369-3385.
Barley, E. P.; Kettner, K. A.; Stephens, E. R. (1963) Analysis of peroxyacyl nitrates by gas chroniatography with
electron capture detection. Anal. Cheni. 35: 589-591.
Darnall, K. R.; Lloyd, A. C.; Winer, A. M.; Pitts, J. N., Jr. (1976) Reactivity scale for atmospheric
hydrocarbons based on reaction with hydroxyl radical. Environ. Sci. Technol. 10: 692-696.
Das, T. N.; Moorthy, P. N.; Rao, K. N. (1982) Chemiluminescent method for the determination of low
concentration of hydrogen peroxide. J. Indian Chem. Soc. 59: 85-88.
Dasgupta, P. K.; Hwang, H. (1985) Application of a nested loop system for the flow injection analysis of trace
aqueous peroxide. Anal. Chem. 57: 1009-1012.
Dasgupta, P. K.; Dong, S.; Hwang, H.; Yang, H.-C.; Genfa, Z. (1988) Continuous liquid-phase fluorometry
coupled to a diffusion scrubber for the real-time determination of atmospheric formaldehyde,
hydrogen peroxide and sulfur dioxide. Atmos. Environ. 22: 949-963.
Dasgupta, P. K.; Dong, S.; Hwang, H. (1990) Diffusion scrubber-based field measurements of atmospheric
formaldehyde and hydrogen peroxide. Aerosol Sci. Technol. 12: 98-104.
Davis, D. D. (1988) Atmospheric nitrogen oxides: their detection and chemistry. Third year report. Atlanta, GA:
Georgia Institute of Technology; pp. 1-13.
Davis, D. D.; Bradshaw, J. D.; Rodgers, M. O.; Sandholni, S. T.; KeSheng, S. (1987) Free tropospheric and
boundary layer measurements of NO over the central and eastern North Pacific Ocean. J. Geophys.
Res. [Atmos.] 92: 2049-2070.
Delany, A. C.; Dickerson, R. R.; Melchior, F. L., Jr.; Wartburg, A. F. (1982) Modification of a commercial
NOX detector for high sensitivity. Rev. Sci. Instrum. 53: 1899-1902.
Demerjian, K. L.; Schere, K. L.; Peterson, J. T. (1980) Theoretical estimates of actinic (spherically integrated)
flux and photolytic rate constants of atmospheric species in the lower troposphere. In: Pitts,
J. N., Jr.; Metcalf, R. L.; Grosjean, D., eds. Advances in environmental science and technology: v.
10. New York, NY: John Wiley & Sons; pp. 369-459.
DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. P.; Kurylo, M. J.; Howard, C. J.; Ravishankara,
A. R.; Kolb, C. E.; Molina, M. J. (1992) Chemical kinetics and photochemical data for use in
stratospheric modeling. Pasadena, CA: NASA Panel for Data Evaluation, Jet Propulsion Laboratory;
publication no. 92-20.
Dennis, R. L.; Novak, J. H. (1992) EPA's third generation modeling system (MODELS-3): an overview.
In: Berglund, R. L., ed. Tropospheric ozone and the environment II: effects, modeling and control
[papers from an international specialty conference]; November; Atlanta, GA. Pittsburgh, PA: Air &
Waste Management Association; pp. 137-147. (A&WM transactions series no. 20).
Dennis, R. L.; McHenry, J. N.; Barchet, W. R.; Binkowski, F. S.; Byun, D. W. (1993a) Correcting RADM's
sulfate underprediction: discovery and correction of model errors and testing the corrections through
comparisons against field data. Atmos. Environ. Part A 27: 975-997.
3-184
-------
Dennis, R. L.; Bynn, D. W.; Novak, J. H.; Coates, C. J.; Galuppi, K. J. (1993b) The next generation of
integrated air quality modeling: EPA's Models-3. Presented at: the international conference on
regional photochemical measurement and modeling studies; November; San Diego, CA. Pittsburgh,
PA: Air & Waste Management Association.
Dentener, F. J.; Crutzen, P. J. (1993) Reaction of N2O5 on tropospheric aerosols: impact on the global
distributions of NO,, O3, and OH. J. Geophys. Res. [Atmos.] 98: 7149-7163.
Derwent, R.; Hov, O. (1988) Application of sensitivity and uncertainty analysis techniques to a photochemical
ozone model. J. Geophys. Res. [Atmos.] 93: 5185-5199.
Derwent, R. G.; Jenkin, M. E. (1991) Hydrocarbons and the long-range transport of ozone and PAN across
Europe. Atmos. Environ. Part A 25: 1661-1678.
Derwent, R. G.; Kay, P. J. A. (1988) Factors influencing the ground level distribution of ozone in Europe.
Environ. Pollut. 55: 191-219.
Dickerson, R. R.; Delany, A. C.; Wartburg, A. F. (1984) Further modification of a commercial NOX detector for
high sensitivity. Rev. Sci. Instrum. 55: 1995-1998.
Dickerson, R. R.; Huffman, G. J.; Luke, W. T.; Nunnermacker, L. J.; Pickering, K. E.; Leslie, A. C. D.;
Lindsey, C. G.; Slinn, W. G. N.; Kelly, T. J.; Daum, P. H.; Delany, A. C.; Greenberg, J. P.;
Zimmerman, P. R.; Boatman, J. P.; Ray, J. D.; Stedman, D. H. (1987) Thunderstorms: an
important mechanism in the transport of air pollutants. Science (Washington, DC) 235: 460-465.
Dietz, W. A. (1967) Response factors for gas chromatographic analyses. J. Gas Chromatogr. 5: 68-71.
Dodge, M. C. (1977) Combined use of modeling techniques and smog chamber data to derive ozone-precursor
relationships. In: Dimitriades, B., ed. International conference on photochemical oxidant pollution
and its control—proceedings: volume II; September 1976; Raleigh, NC. Research Triangle Park,
NC: U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory; pp. 881-
889; report no. EPA-600/3-77-001b. Available from: NTIS, Springfield, VA; PB-264233.
Dodge, M. C. (1989) A comparison of three photochemical oxidant mechanisms. J. Geophys. Res. [Atmos.]
94: 5121-5136.
Dodge, M. C. (1990) Formaldehyde production in photochemical smog as predicted by three state-of-the-science
chemical oxidant mechanisms. J. Geophys. Res. [Atmos.] 95: 3635-3648.
Dolske, D. A.; Gatz, D. F. (1985) A field intercomparison of methods for the measurement of particle and gas
dry deposition. J. Geophys. Res. 90: 2076-2084.
Drummond, J. W.; Volz, A.; Ehhalt, D. H. (1985) An optimized chemiluminescence detector for tropospheric
NO measurements. J. Atmos. Chem. 2: 287-306.
Drummond, J.; Schiff, H.; Karecki, D.; Mackay, G. (1989) Measurements of N02, O3, PAN, HNO3, H2O2, and
H2CO during the southern California air quality study. Presented at: 82nd annual meeting and
exhibition of the Air & Waste Management Association; June; Anaheim, CA. Pittsburgh, PA: Air &
Waste Management Association; report no. 89-139.4.
3-185
-------
Drummond, J. W.; Shepson, P. B.; Mackay, G. I.; Schiff, H. I. (1992) Measurements of NQ, NO*, and NO,
using a new converter-sequencer and sensitive Luminox® detection. In: Measurement of toxic and
related air pollutants: proceedings of the 1992 U.S. EPA/A&WMA international symposium; May;
Durham, NC. Pittsburgh, PA: Air & Waste Management Association; pp. 750-755. (A&WMA
publication VIP-25).
Dunker, A. M.; Schleyer, C. H.; Morris, R. E.; Pollack, A. K. (1992a) Effects of methanol/gasoline blends used
in flexible/variable fuel vehicles on urban air quality in year 2005/2010—auto/oil air quality
improvement research program. Presented at: 85th annual meeting and exhibition of the Air & Waste
Management Association; June; Kansas City, MO. Pittsburgh, PA: Air & Waste Management
Association; paper no. 92-119.06.
Dunker, A. M.; Morris, R. E.; Pollack, A. K.; Cohen, J. P.; Schleyer, C. H.; Chock, D. P. (1992b) Effects
of aromatics, MTBE, olefms, and T90 on urban air quality in year 2005/2010—auto/oil air quality
improvement research program. Presented at: 85th annual meeting and exhibition of the Air & Waste
Management Association; June; Kansas City, MO. Pittsburgh, PA: Air & Waste Management
Association; paper no. 92-119.03.
Ehhalt, D. H.; Dorn, H.-P.; Poppe, D. (1991) The chemistry of the hydroxyl radical in the troposphere. In: Last,
F. T.; Walling, R., eds. Acidic deposition: its nature and impacts. Proceedings of the international
symposium; September 1990; Glasgow, United Kingdom. Proc. R. Soc. Edinburgh, Sect. B: Biol.
Sci. 97: 17-34.
Ehrenfeld, J. R. (1974) Analysis of the composition of the atmosphere in the Los Angeles basin. Washington, DC:
U.S. Environmental Protection Agency, Office of Research and Development; report no.
EPA-650/2-74-105. Available from: NTIS, Springfield, VA; PB-239466.
Eisele, F. L.; Tanner, D. J. (1991) Ion-assisted tropospheric OH measurements. J. Geophys. Res. [Atmos.]
96: 9295-9308.
Eisenreich, S. J.; Looney, B. B.; Thornton, J. D. (1981) Airborne organic contaminants in the Great Lakes
ecosystem. Environ. Sci. Technol. 15: 30-38.
Eldering, A.; Larson, S. M.; Hall, J. R.; Hussey, K. J.; Cass, G. R. (1993) Development of an improved image
processing based visibility model. Environ. Sci. Technol. 27: 626-635.
Errico, R.; Bates, G. T. (1988) Implicit normal-mode initialization of the PSU/NCAR mesoscale model. Boulder,
CO: National Center for Atmospheric Research; NCAR technical note 312.
Fahey, D. W.; Eubank, C. S.; Hubler, G.; Fehsenfeld, F. C. (1985) Evaluation of a catalytic reduction technique
for the measurement of total reactive odd-nitrogen NOy in the atmosphere. J. Atmos. Chem. 3: 435-
468.
Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30) 36: 8186-
8201.
Federal Register. (1975) Ambient air monitoring reference and equivalent methods. F. R. (February 18) 40: 7042-
7070.
Federal Register. (1979) National primary and secondary ambient air quality standards: calibration of ozone
reference methods. F. R. (February 8) 44: 8221-8233.
Federal Register. (1990) Office of Research and Development; ambient air monitoring reference and equivalent
methods; equivalent method designation. F. R. (September 18) 55: 38386-38387.
3-186
-------
Federal Register. (1992) Office of Research and Development ambient air monitoring reference and equivalent
methods; equivalent method and reference method designations. F. R. (September 28)
57: 44565-44566.
Federal Register. (1993) Ambient air monitoring reference and equivalent methods; equivalent method
designations. F. R. (February 3) 58: 6964-6965.
Fehsenfeld, F. C.; Dickerson, R. R.; Hiibler, G.; Luke, W. T.; Nunnermacker, L. J.; Williams, E. J.; Roberts,
J. M.; Calvert, J. G.; Curran, C. M.; Delany, A. C.; Eubank, C. S.; Fahey, D. W.; Fried, A.;
Gandrud, B. W.; Langford, A. O.; Murphy, P. C.; Norton, R. B.; Pickering, K. E.; Ridley, B. A.
(1987) A ground-based intercomparison of NO, NOX, and NO, measurement techniques. J. Geophys.
Res. [Atmos.] 92: 14,710-14,722.
Fehsenfeld, F. C.; Parrish, D. D.; Fahey, D. W. (1988) The measurement of NOX in the non-urban troposphere.
In: Isaksen, I. S. A., ed. Tropospheric ozone: regional and global scale interactions. Dordrecht, The
Netherlands: D. Reidel Publishing; pp. 185-215.
Fehsenfeld, F. C.; Drummond, J. W.; Roychowdhury, U. K.; Galvin, P. J.; Williams, E. J.; Buhr, M. P.;
Parrish, D. D.; Hubler, G.; Langford, A. O.; Calvert, J. G.; Ridley, B. A.; Grahek, P.; Heikes, B.
G.; Kok, G. L.; Shelter, J. D.; Walega, J. G.; Elsworth, C. M.; Norton, R. B.; Fahey, D. W.;
Murphy, P. C.; Hovermale, C.; Mohnen, V. A.; Demerjian, K. L.; Mackay, G. L; Schiff, H. I.
(1990) Intercomparison of NO2 measurement techniques. J. Geophys. Res. [Atmos.] 95: 3579-3597.
Fehsenfeld, P.; Calvert, J.; Fall, R.; Goldan, P.; Guenther, A. B.; Hewitt, C. N.; Lamb, B.; Liu, S.;
Trainer, M.; Westberg, H.; Zimmerman, P. (1992) Emissions of volatile organic compounds from
vegetation and the implications for atmospheric chemistry. Global Biogeocheni. Cycles 6: 389-430.
Felton, C. C.; Sheppard, J. C.; Campbell, M. J. (1990) The radiochemical hydroxyl radical measurement method.
Environ. Sci. Technol. 24: 1841-1847.
Finlayson-Pitts, B. J.; Pitts, J. N., Jr. (1986) Atmospheric chemistry: fundamentals and experimental techniques.
New York, NY: John Wiley & Sons.
Finlay son-Pitts, B. J.; Pitts, J. N., Jr. (1993) Atmospheric chemistry of tropospheric ozone formation: scientific
and regulatory implications. Air Waste 43: 1091-1100.
Fried, A.; Hodgeson, J. (1982) Laser photoacoustic detection of nitrogen dioxide in the gas-phase titration of nitric
oxide with ozone. Anal. Chem. 54: 278-282.
Fujita, E. M. (1993) Trends in emissions and ambient concentrations of CO, NMHC and NOX in the South Coast
Air Basin. In: Southern California Air Quality Study data analysis: proceedings of an international
specialty conference; July 1992; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
Association; pp. 43-48. (A&WMA publication VIP-26).
Fujita, E. M.; Croes, B. E.; Bennett, C. L.; Lawson, D. R.; Lurmann, F. W.; Main, H. H. (1992) Comparison
of emission inventory and ambient concentration ratios of CO, NMOG, and NOX in California's south
coast air basin. J. Air Waste Manage. Assoc. 42: 264-276.
Fung, K. (1989) Carbonyl observations during the SCAQS. Presented at: 82nd annual meeting and exhibition of
the Air & Waste Management Association; June; Anaheim, CA. Pittsburgh, PA: Air & Waste
Management Association; paper no. 89-152.3.
Fung, K.; Grosjean, D. (1981) Determination of nanogram amounts of carbonyls as 2,4-dinitrophenylhydrazones
by high-performance liquid chromatography. Anal. Chem. 53: 168-171.
3-187
-------
Fung, I.; John, J.; Lerner, J.; Matthews, E.; Prather, M.; Steele, L. P.; Fraser, P. J. (1991a) Three-dimensional
model synthesis of the global methane cycle. J. Geophys. Res. [Atmos.] 96: 13,033-13,065.
Fung, C. S.; Misra, P. K.; Bloxam, R.; Wong, S. (1991b) A numerical experiment on the relative importance of
H2O2 and O3 in aqueous conversion of SQ to SO42". Atmos. Environ. Part A 25: 41 1-423.
Gab, S.; Hellpointner, E.; Turner, W. V.; Korte, F. (1985) Hydroxymethyl hydroperoxide and
bis(hydroxymethyl) peroxide from gas-phase ozonolysis of naturally occuring alkenes. Nature
(London) 3 16: 535-536.
Gaffney, J. S.; Fajer, R.; Senum, G. I. (1984) An improved procedure for high purity gaseous peroxyacyl nitrate
production: use of heavy lipid solvents. Atmos. Environ. 18: 215-218.
Gay, B. W., Jr.; Bufalini, J. J. (1972a) Hydrogen peroxide in the urban atmosphere. In: Photochemical smog and
ozone reactions: proceedings of the 161st meeting of the American Chemical Society; March- April
1971; Los Angeles, CA. Washington, DC: American Chemical Society; pp. 255-263. (Gould, R. P.,
ed. Advances in chemistry series: v. 113).
Gay, B. W., Jr.; Bufalini, J. J. (1972b) Hydrogen peroxide in the urban atmosphere. Environ. Lett. 3: 21-24.
Gay, B. W., Jr.; Noonan, R. C.; Bufalini, J. J.; Hanst, P. L. (1976) Photochemical synthesis of peroxyacyl
nitrates in gas phase via chlorine-aldehyde reaction. Environ. Sci. Technol. 10: 82-85.
Gay, B. W.; Meeks, S.; Bufalini, J. J. (1988) Peroxide formation and detection in various systems. Presented
at: 196th meeting of the American Chemical Society; Los Angeles, CA. 28: 93-94.
Gertler, A. W.; Pierson, W. R.; Wittoff, D. N.; Robinson, N. F. (1994) Where do tunnel studies fit in for
emissions inventory development? Presented at: Emissions inventory workshop; November.
Gervat, G. P.; Clark, P. A.; Marsh, A. R. W.; Teasdale, L; Chandler, A. S.; Choularton, T. W.; Gay, M. J.;
Hill, M. K.; Hill, T. A. (1988) Field evidence for the oxidation of SO, by H2Q, in cap clouds.
Nature (London) 333: 241-243.
Gery, M. W.; Whitten, G. Z.; Killus, J. P. (1988) Development and testing of the CMB-IV for urban and
regional modeling. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Atmospheric Sciences Research Laboratory; report no. EPA-600/3-88-012. Available from: NTIS,
Springfield, VA; PB88-180039.
Gery, M. W.; Whitten, G. Z.; Killus, J. P.; Dodge, M. C. (1989) A photochemical kinetics mechanism for urban
and regional scale computer modeling. J. Geophys. Res. [Atmos.] 94: 12,925-12,956.
Gholson, A. R.; Jayanty, R. K. M.; Storm, J. F. (1990) Evaluation of aluminum canisters for the collection and
storage of air toxics. Anal. Chem. 62: 1899-1902.
Global Thinking Project. (1993) The Global Thinking Project. Supported by the Eisenhower Higher Education
Program and Georgia State University.
Goldan, P. D.; Kuster, W. C.; Fehsenfeld, F. C.; Montzka, S. A. (1993) The observation of a Q alcohol
emission in a North American pine forest. Geophys. Res. Lett. 20: 1039-1042.
Goodin, W. R.; McRae, G. J.; Seinfeld, J. H. (1980) An objective analysis technique for constructing
three-dimensional urban-scale wind fields. J. Appl. Meteorol. 19: 98-108.
Gordon, G. E. (1988) Receptor models. Environ. Sci. Technol. 22: 1132-1142.
3-188
-------
Gordon, S. M.; Miller, M. (1989) Analysis of ambient polar volatile organic compounds using chemical
ionization-ion trap detector. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/3-89/070.
Available from: NTIS, Springfield, VA; PB90-106451.
Graedel, T. E.; Weschler, C. J. (1981) Chemistry within aqueous atmospheric aerosols and raindrops.
Rev. Geophys. Space Phys. 19: 505-539.
Graedel, T. E.; Hawkins, D. T.; Claxton, L. D. (1986a) Atmospheric chemical compounds: sources, occurrence,
and bioassay. Orlando, PL: Academic Press, Inc.
Graedel, T. E.; Mandich, M. L.; Weschler, C. J. (1986b) Kinetic model studies of atmospheric droplet chemistry:
2. homogenous transition metal chemistry in raindrops. J. Geophys. Res. [Atmos.] 91: 5205-5221.
Greenhut, G. K. (1986) Transport of ozone between boundary layer and cloud layer by cumulus clouds.
J. Geophys. Res. [Atmos.] 91: 8613-8622.
Gregory, G. L.; Hoell, J. M., Jr.; Torres, A. L.; Carroll, M. A.; Ridley, B. A.; Rodgers, M. O.; Bradshaw, J.;
Sandholm, S.; Davis, D. D. (1990a) An intercomparison of airborne nitric oxide measurements:
a second opportunity. J. Geophys. Res. [Atmos.] 95: 10,129-10,138.
Gregory, G. L.; Hoell, J. M., Jr.; Carroll, M. A.; Ridley, B. A.; Davis, D. D.; Bradshaw, J.; Rodgers, M. O.;
Sandholm, S. T.; Schiff, H. I.; Hastie, D. R.; Karecki, D. R.; Mackay, G. I.; Harris, G. W.;
Torres, A. L.; Fried, A. (1990b) An intercomparison of airborne nitrogen dioxide instruments. J.
Geophys. Res. [Atmos.] 95: 10,103-10,127.
Griffith, D. W. T.; Schuster, G. (1987) Atmospheric trace gas analysis using matrix isolation-Fourier transform
infrared spectroscopy. J. Atmos. Chem. 5: 59-81.
Grosjean, D. (1982) Formaldehyde and other carbonyls in Los Angeles ambient air. Environ. Sci. Technol.
16: 254-262.
Grosjean, D. (1988) Aldehydes, carboxylic acids and inorganic nitrate during NSMCS. Atmos. Environ.
22: 1637-1648.
Grosjean, D. (1991) Ambient levels of formaldehyde, acetaldehyde, and formic acid in southern California: results
of a one-year base-line study. Environ. Sci. Technol. 25: 710-715.
Grosjean, D.; Fung, K. (1984) Hydrocarbons and carbonyls in Los Angeles air. J. Air Pollut. Control Assoc.
34: 537-543.
Grosjean, D.; Harrison, J. (1985a) Peroxyacetyl nitrate: comparison of alkaline hydrolysis and chemiluminescence
methods. Environ. Sci. Technol. 19: 749-752.
Grosjean, D.; Harrison, J. (1985b) Response of chemiluminescence NOX analyzers and ultraviolet ozone analyzers
to organic air pollutants. Environ. Sci. Technol. 19: 862-865.
Grosjean, D.; Hisham, M. W. M. (1992) A passive sampler for atmospheric ozone. J. Air Waste Manage. Assoc.
42: 169-173.
Grosjean, D.; Williams, E. L., II. (1992) Field tests of a passive sampler for atmospheric ozone at California
mountain forest locations. Atmos. Environ. Part A 26: 1407-1411.
Grosjean, D.; Fung, K.; Collins, J.; Harrison, J.; Breitung, E. (1984) Portable generator for on-site calibration of
peroxyacetyl nitrate analyzers. Anal. Chem. 56: 569-573.
3-189
-------
Grosjean, D.; Williams, E. L., II; Grosjean, E. (1993a) A biogenic precursor of peroxypropionyl nitrate:
atmospheric oxidation of ds-3-hexen-l-ol. Environ. Sci. Technol. 27: 979-981.
Grosjean, E.; Williams, E. L., II; Grosjean, D. (1993b) Ambient levels of formaldehyde and acetaldehyde in
Atlanta, Georgia. Air Waste 43: 469-474.
Guilbault, G. G.; Brignac, P. J., Jr.; Juneau, M. (1968) New substrates for the fluoronietric determination of
oxidative enzymes. Anal. Chem. 40: 1256-1263.
Guinnup, D.; Possiel, N. (1991) Regional ozone modeling to support the Lake Michigan ozone study [draft].
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards.
Gunz, D. W.; Hoffmann, M. R. (1990) Atmospheric chemistry of peroxides: a review. Atmos. Environ.
Part A 24: 1601-1633.
Hakola, H.; Shorees, B.; Arey, J.; Atkinson, R. (1993) Product formation from the gas-phase reactions of
OH radicals and O3 with D-phellandrene. Environ. Sci. Technol. 27: 278-283.
Hakola, H.; Arey, J.; Aschniann, S. M.; Atkinson, R. (1994) Product formation from the gas-phase reactions of
OH radicals and O3 with a series of monoterpenes. J. Atmos. Chem. 18: 75-102.
Haltiner, G. J. (1971) Numerical weather prediction. New York, NY: John Wiley & Sons, Inc.
Haltiner, G. J.; Williams, R. T. (1980) Numerical prediction and dynamic meteorology. 2nd ed. New York,
NY: John Wiley & Sons.
Hampson, R. P.; Braun, W.; Brown, R. L.; Garvin, D.; Herron, J. T.; Huie, R. E.; Kurylo, M. J.;
Laufer, A. H.; McKinley, J. D.; Okabe, H.; Scheer, M. D.; Tsang, W. (1973) Survey of
photochemical and rate data for twenty-eight reactions of interest in atmospheric chemistry. J. Phys.
Chem. Ref. Data 2: 267-311.
Hansen, D. A. (1989) Measuring trace gases with FM spectroscopy. EPRI J. 14(4): 42-43.
Hanst, P. L.; Wong, N. W.; Bragin, J. (1982) A long-path infra-red study of Los Angeles smog. Atmos. Environ.
16: 969-981.
Hard, T. M.; Mehrabzadeh, A. A.; Chan, C. Y.; O'Brien, R. J. (1992) PAGE measurements of tropospheric HO
with measurements and model of interferences. J. Geophys. Res. [Atmos.] 97: 9795-9817.
Harley, R. A.; Hannigan, M. P.; Cass, G. R. (1992) Respeciation of organic gas emissions and the detection of
excess unburned gasoline in the atmosphere. Environ. Sci. Technol. 26: 2395-2408.
Harley, R. A.; Russell, A. G.; McRae, G. J.; Cass, G. R.; Seinfeld, J. H. (1993) Photochemical modeling of the
Southern California Air Quality Study. Environ. Sci. Technol. 27: 378-388.
Harms, D. E.; Raman, S.; Madala, R. V. (1992) An examination of four-dimensional data-assimilation techniques
for numerical weather prediction. Bull. Am. Meteorol. Soc. 73: 425-440.
Harris, G. W.; Carter, W. P. L.; Winer, A. M.; Pitts, J. N., Jr.; Plait, U.; Perner, D. (1982) Observations of
nitrous acid in the Los Angeles atmosphere and implications for predictions of ozone-precursor
relationships. Environ. Sci. Technol. 16: 414-419.
Harrison, J. W.; Timmons, M. L.; Denyszyn, R. B.; Decker, C. E. (1977) Evaluation of the EPA reference
method for measurement of non-methane hydrocarbons. Research Triangle Park, NC: U.S.
3-190
-------
Environmental Protection Agency, National Environmental Research Center; report no.
EPA-600/4-77-033. Available from: NTIS, Springfield, VA; PB-278296.
Hatakeyama, S.; Izuini, K.; Fukuyama, T.; Akimoto, H. (1989) Reactions of ozone with D-pinene and D-pinene in
air: yields of gaseous and particulate products. J. Geophys. Res. [Atmos.] 94: 13,013-13,024.
Hatakeyama, S.; Izuini, K.; Fukuyama, T.; Akimoto, H.; Washida, N. (1991) Reactions of OH withD-pinene and
D-pinene in air: estimate of global CO production from the atmospheric oxidation of terpenes.
J. Geophys. Res. [Atmos.] 96: 947-958.
Heikes, B. G. (1984) Aqueous H2O2 production from O3 in glass impingers. Atmos. Environ. 18: 1433-1445.
Heikes, B. G.; Lazrus, A. L.; Kok, G. L.; Kunen, S. M.; Gandrud, B. W.; Gitlin, S. N.; Sperry, P. D. (1982)
Evidence for aqueous phase hydrogen peroxide synthesis in the troposphere. J. Geophys. Res. C:
Oceans Atmos. 87: 3045-3051.
Heikes, B. G.; Kok, G. L.; Walega, J. G.; Lazrus, A. L. (1987) H2O2, O3 and SO2 measurements in the lower
troposphere over the eastern United States during fall. J. Geophys. Res. 92: 915-931.
Helas, G.; Flanz, M.; Warneck, P. (1981) Improved NOX monitor for measurements in tropospheric clean air
regions. Int. J. Environ. Anal. Chem. 10: 155-166.
Hellpointner, E.; Gab, S. (1989) Detection of methyl, hydroxymethyl and hydroxyethyl hydroperoxides in air and
precipitation. Nature (London) 337: 631-634.
Helmig, D.; Mueller, J.; Klein, W. (1989) Improvements in analysis of atmospheric peroxyacetyl nitrate (PAN).
Atmos. Environ. 23: 2187-2192.
Heuss, J. M.; Wolff, G. T. (1993) Measurement needs for developing and assessing ozone-control strategies.
Water Air Soil Pollut. 67: 79-92.
Hewitt, C. N.; Kok, G. L. (1991) Formation and occurence of organic hydroperoxides in the troposphere:
laboratory and field observations. J. Atmos. Chem. 12: 181-194.
Hildemann, L. M.; Markowski, G. R.; Cass, G. R. (1991a) Chemical composition of emissions from urban
sources of fine organic aerosol. Environ. Sci. Technol. 25: 744-759.
Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991b) Quantitative characterization of
urban sources of organic aerosol by high-resolution gas chroniatography. Environ. Sci. Technol.
25: 1311-1325.
3-191
-------
Hill, A. C.; Chamberlain, E. M., Jr. (1976) The removal of water soluble gases from the atmosphere by
vegetation. In: Atmosphere-surface exchange of paniculate and gaseous pollutants (1974):
proceedings of a symposium; September 1974; Richland, WA. Oak Ridge, TN: Energy Research and
Development Administration; pp. 153-170. (ERDA symposium series no. 38). Available from:
NTIS, Springfield, VA; CONF-740921.
Hjorth, J.; Lohse, C.; Nielsen, C. J.; Skov, H.; Restelli, G. (1990) Products and mechanisms of gas-phase
reactions between NO3 and a series of alkenes. J. Phys. Chem. 94: 7494-7500.
Hochanadel, C. J. (1952) Effects of cobalt D-radiation on water and aqueous solutions. J. Phys. Chem.
56: 587-594.
Hodgeson, J. A.; Krost, K. J.; O'Keeffe, A. E.; Stevens, R. K. (1970) Chemiluminescent measurement of
atmospheric ozone: response characteristics and operating variables. Anal. Chem. 42: 1795-1802.
Hoell, J. M., Jr.; Gregory, G. L.; McDougal, D. S.; Torres, A. L.; Davis, D. D.; Bradshaw, J.;
Rodgers, M. O.; Ridley, B. A.; Carroll, M. A. (1987) Airborne intercomparison of nitric oxide
measurement techniques. J. Geophys. Res. [Atmos.] 92: 1995-2008.
Hofzumahaus, A.; Dorn, H.-P.; Callies, J.; Plait, U.; Ehhalt, D. H. (1991) Tropospheric OH concentration
measurements by laser long-path absorption spectroscopy. Atmos. Environ. Part A 25: 2017-2022.
Hoke, J. E.; Anthes, R. A. (1976) The initialization of numerical models by a dynamic initialization technique.
Mon. Weather Rev. 104: 1551-1556.
Holdren, M. W.; Rasmussen, R. A. (1976) Moisture anomaly in analysis of peroxyacetyl nitrate (PAN). Environ.
Sci. Technol. 10: 185-187.
Holdren, M. W.; Smith, D. L. (1987) Stability of volatile organic compounds while stored in SUMMA polished
stainless steel canisters [final report]. Columbus, OH: Battelle Columbus Laboratory; EPA contract
no. 68-02-4127, WA-13.
Holdren, M. W.; Spicer, C. W. (1984) Field compatible calibration procedure for peroxyacetyl nitrate. Environ.
Sci. Technol. 18: 113-116.
Holdren, M. W.; Smith, D. L.; Pollack, A. J.; Pate, A. D. (1993) The 1992 demonstration/field study of new
designs of automated gas chromatographs in Connecticut and other locations. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment
Laboratory; report no. EPA/600/R-93/193. Available from: NTIS, Springfield, VA; PB94-105376.
Holland, D. M.; McElroy, F. F. (1986) Analytical method comparisons by estimates of precision and lower
detection limit. Environ. Sci. Technol. 20: 1157-1161.
Holzworth, G. C. (1964) Estimates of mean maximum mixing depths in the contiguous United States. Mon.
Weather Rev. 92: 235-242.
Holzworth, G. C. (1972) Mixing heights, wind speeds, and potential for urban air pollution throughout the
contiguous United States. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Air Programs; publication no. AP-101. Available from: NTIS, Springfield, VA;
PB-207103.
Horie, O.; Moortgat, G. K. (1991) Decomposition pathways of the excited Criegee intermediates in the ozonolysis
of simple alkenes. Atmos. Environ. Part A 25: 1881-1896.
3-192
-------
Horowitz, A.; Calvert, J. G. (1982) Wavelength dependence of the primary processes in acetaldehyde photolysis.
J. Phys. Chem. 86: 3105-3114.
Hoshino, H.; Hinze, W. L. (1987) Exploitation of reversed micelles as a medium in analytical cheniiluininescence
measurements with application to the determination of hydrogen peroxide using Luminol. Anal.
Chem. 59: 496-504.
Hosier, C. R. (1961) Low-level inversion frequency in the contiguous United States. Mon. Weather Rev.
339.
319-
Hudgens, E. E.; Kleindienst, T. E.; McElroy, F. P.; Ollison, W. M. (1994) A study of interferences in ozone UV
and chemiluminescence monitors. In: Measurement of toxic and related air pollutants: proceedings of
the U.S. EPA/A&WMA international symposium; May; Durham, NC. Pittsburgh, PA: Air & Waste
Management Association; pp. 405-415. (A&WMA specialty conference VIP 39).
Huebert, B. J.; Robert, C. H. (1985) The dry deposition of nitric acid to grass. J. Geophys. Res. [Atmos.]
90: 2085-2090.
Ibusuki, T. (1983) Influence of trace metal ions on the determination of hydrogen peroxide in rainwater by using a
chemiluminescent technique. Atmos. Environ. 17: 393-396.
Isidorov, V. A.; Zenkevich, I. G.; loffe, B. V. (1985) Volatile organic compounds in the atmosphere of forests.
Atmos. Environ. 19: 1-8.
Jacob, D. J.; Gottlieb, E. W.; Prather, M. J. (1989) Chemistry of a polluted cloudy boundary layer. J. Geophys.
Res. [Atmos.] 94: 12,975-13,002.
Jacob, D. J.; Logan, J. A.; Gardner, G. M.; Yevich, R. M.; Spivakovsky, C. M.; Wofsy, S. C.; Sillman, S.;
Prather, M. J. (1993a) Factors regulating ozone over the United States and its export to the global
atmosphere. J. Geophys. Res. [Atmos.] 98: 14,817-14,826.
Jacob, D. J.; Logan, J. A.; Yevich, R. M.; Gardner, G. M.; Spivakovsky, C. M.; Wofsy, S. C.; Munger, J. W.;
Sillman, S.; Prather, M. J.; Rodgers, M. O.; Westberg, H.; Zimmerman, P. R. (1993b) Simulation
of summertime ozone over North America. J. Geophys. Res. [Atmos.] 98: 14,797-14,816.
Jang, J.-C. C. (1992) Sensitivity of ozone to model grid resolution [dissertation]. Chapel Hill, NC: University of
North Carolina-Chapel Hill, Department of Environmental Sciences and Engineering. Available
from: University Microfilms International; publication no. AAD93-09896.
Japar, S. M.; Wallington, T. J.; Richer!, J. F. O.; Ball, J. C. (1990) The atmospheric chemistry of oxygenated
fuel additives: f-butyl alcohol, dimethyl ether, and methyl f-butyl ether. Int. J. Chem. Kinet.
22: 1257-1269.
Jayanty, R. K. M.; Blackard, A.; McElroy, F. P.; McClenny, W. A. (1982) Laboratory evaluation of non
methane organic carbon determination in ambient air by cryogenic preconcentration and flame
ionization detection. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Environmental Monitoring Systems Laboratory; report no. EPA-600/4-82-019. Available from:
NTIS, Springfield, VA; PB82-224965.
Jeffries, H.; Fox, D.; Kaniens, R. (1975) Outdoor smog chamber studies: effect of hydrocarbon reduction on
nitrogen dioxide. Washington, DC: U.S. Environmental Protection Agency, Office of Research and
Development; report no. EPA-650/3-75-011. Available from: NTIS, Springfield, VA; PB-245829.
Jeffries, H.; Fox, D.; Kaniens, R. (1976) Outdoor smog chamber studies: light effects relative to indoor
chambers. Environ. Sci. Technol. 10: 1006-1011.
3-193
-------
Jeffries, H. E.; Gery, M. W.; Carter, W. P. L. (1992) Protocols for evaluating oxidant mechanisms for urban and
regional models. Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric
Research and Exposure Assessment Laboratory; report no. EPA-600/R-92/112. Available from:
NTIS, Springfield, VA; PB92-205848.
Jenkin, M. E.; Cox, R. A.; Williams, D. J. (1988) Laboratory studies of the kinetics of formation of nitrous acid
from the thermal reaction of nitrogen dioxide and water vapour. Atmos. Environ. 22: 487-498.
Johansson, C. (1984) Field measurements of emission of nitric oxide from fertilized and unfertilized forest soils in
Sweden. J. Atmos. Chem. 1: 429-442.
Johansson, C.; Granat, L. (1984) Emission of nitric oxide from arable land. Tellus Ser. B 36: 25-37.
Johnson, G. M.; Quigley, S. M. (1989) A universal monitor for photochemical smog. Presented at: 82nd annual
meeting and exhibition of the Air & Waste Management Association; June; Anaheim, CA.
Pittsburgh, PA: Air & Waste Management Association; paper no. 89-29.8.
Johnson, W. B.; Viezee, W. (1981) Stratospheric ozone in the lower troposphere—I. presentation and
interpretation of aircraft measurements. Atmos. Environ. 15: 1309-1323.
Johnston, H. (1971) Reduction of stratospheric ozone by nitrogen oxide catalysts from supersonic transport
exhaust. Science (Washington, DC) 173: 517-522.
Jones, K.; Militana, L.; Martini, J. (1989) Ozone trend analysis for selected urban areas in the continental U.S.
Presented at: 82nd annual meeting and exhibition of the Air & Waste Management Association; June;
Anaheim, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 89-3.6.
Joos, L. P.; Landolt, W. P.; Leuenberger, H. (1986) Calibration of peroxyacetyl nitrate measurements with
an NOX analyzer. Environ. Sci. Technol. 20: 1269-1273.
Joseph, D. W.; Spicer, C. W. (1978) Chemiluminescence method for atmospheric monitoring of nitric acid and
nitrogen oxides. Anal. Chem. 50: 1400-1403.
Joshi, S. B.; Bufalini, J. J. (1978) Halocarbon interferences in chemiluminescent measurements of NC^. Environ.
Sci. Technol. 12: 597-599.
Junge, C. E. (1963) Air chemistry and radioactivity. New York, NY: Academic Press. (Van Miegheni, J.; Hales,
A. L., eds. International geophysics series: v. 4).
Jiittner, P. (1986) Analysis of organic compounds (VOC) in the forest air of the southern Black Forest.
Chemosphere 15: 985-992.
Kalabokis, P.; Carlier, P.; Fresnet, P.; Mouvier, G.; Toupance, G. (1988) Field studies of aldehyde chemistry
in the Paris area. Atmos. Environ. 22: 147-155.
Kamens, R. M.; Gery, M. W.; Jeffries, H. E.; Jackson, M.; Cole, E. I. (1982) Ozone-isoprene reactions: product
formation and aerosol potential. Int. J. Chem. Kinet. 14: 955-975.
Kaplan, W. A.; Wofsy, S. C.; Keller, M.; Da Costa, J. M. (1988) Emission of NO and deposition of
O3 in a tropical forest system. J. Geophys. Res. [Atmos.] 93: 1389-1395.
Karanichandani, P.; Venkatram, A. (1992) The role of non-precipitating clouds in producing ambient sulfate
during summer: results from simulations with the Acid Deposition and Oxidant Model (ADOM).
Atmos. Environ. Part A 26: 1041-1052.
3-194
-------
Katz, M. (1976) Nitrogen compounds and oxidants. In: Stern, A. C., ed. Air pollution: v. Ill, measuring,
monitoring, and surveillance of air pollution. 3rd ed. New York, NY: Academic Press; pp. 259-305.
Kelly, T. J.; Fortune, C. R. (1994) Continuous monitoring of gaseous formaldehyde using an improved
fluorescence approach. Int. J. Environ. Anal. Chem. 54: 249-263.
Kelly, N. A.; Gunst, R. F. (1990) Response of ozone to changes in hydrocarbon and nitrogen oxide concentrations
in outdoor smog chambers filled with Los Angeles air. Atmos. Environ. Part A 24: 2991-3005.
Kelly, T. J.; Stedman, D. H.; Ritter, J. A.; Harvey, R. B. (1980) Measurements of oxides of nitrogen and nitric
acid in clean air. J. Geophys. Res. C: Oceans Atmos. 85: 7417-7425.
Kelly, N. A.; Ferman, M. A.; Wolff, G. T. (1986) The chemical and meteorological conditions associated with
high and low ozone concentrations in southeastern Michigan and nearby areas of Ontario. J. Air
Pollut. Control Assoc. 36: 150-158.
Kelly, T. J.; Spicer, C. W.; Ward, G. F. (1990) An assessment of the luminol chemiluminescence technique for
measurement of NO2 in ambient air. Atmos. Environ. Part A 24: 2397-2403.
Kelly, T. J.; Callahan, P. J.; Pleil, J.; Evans, G. F. (1993) Method development and field measurements for polar
volatile organic compounds in ambient air. Environ. Sci. Technol. 27: 1146-1153.
Kenski, D. M.; Wadden, R. A.; Scheff, P. A.; Lonneman, W. A. (1993) A receptor modeling approach to VOC
emission inventory validation in five U.S. cities. Presented at: 86th annual meeting and exhibition of
the Air & Waste Management Association; June; Denver, CO. Pittsburgh, PA: Air & Waste
Management Association; paper no. 93-WP-100.04.
Kessler, R. C. (1988) What techniques are available for generating windfields? Presented at: Conference on
photochemical modeling as a tool for decision makers. Pasadena, CA: California Air Resources
Board.
Kessler, R. C.; Douglas, S. G. (1989) Numerical simulation of mesoscale airflow in the South Central Air Coast
Basin. San Rafael, CA: Systems Applications, Inc.; paper no. SYSAPP-89/108.
Khalil, M. A. K.; Rasmussen, R. A. (1992) Forest hydrocarbon emissions: relationships between fluxes and
ambient concentrations. J. Air Waste Manage. Assoc. 42: 810-813.
Kirollos, K. S.; Attar, A. J. (1991) Direct-read passive dosimetry of nitrogen dioxide and ozone. In: Measurement
of toxic and related air pollutants: proceedings of the 1991 U.S. EPA/A&WMA international
symposium, v. 1; May; Durham, NC. Pittsburgh, PA: Air & Waste Management Association;
pp. 297-302. (A&WMA publication VIP-21).
Kleindienst, T. E.; Shepson, P. B.; Hodges, D. N.; Nero, C. M.; Arnts, R. R.; Dasgupta, P. K.; Hwang, H.;
Kok, G. L.; Lind, J. A.; Lazrus, A. L.; Mackay, G. I.; Mayne, L. K.; Schiff, H. I. (1988)
Comparison of techniques for measurement of ambient levels of hydrogen peroxide. Environ. Sci.
Technol. 22: 53-61.
Kleindienst, T. E.; Hudgens, E. E.; Smith, D. P.; McElroy, F. P.; Bufalini, J. J. (1993) Comparison of
chemiluminescence and ultraviolet ozone monitor responses in the presence of humidity and
photochemical pollutants. Air Waste 43: 213-222.
Kleinman, L.; Lee, Y.-N.; Springston, S. R.; Nunnerrnacker, L.; Zhou, X.; Brown, R.; Hallock, K.; Klotz, P.;
Leahy, D.; Lee, J. H.; Newman, L. (1994) Ozone formation at a rural site in the southeastern United
States. J. Geophys. Res. [Atmos.] 99: 3469-3482.
3-195
-------
Kley, D.; McFarland, M. (1980) Chemiluminescence detector for NO and NQ. Atmos. Technol. 12: 63-69.
Klockow, D.; Jacob, P. (1986) The peroxyoxalate Chemiluminescence and its application to the determination of
hydrogen peroxide in precipitation. In: Jaeschke, W., ed. Chemistry of multiphase atmospheric
systems. Berlin, Germany: Springer-Verlag; pp. 117-130. (NATO Advanced Science Institutes
series: v. G6).
Klouda, G. A.; Norris, J. E.; Currie, L. A.; Rhoderick, G. C.; Sams, R. L.; Dorko, W. D.; Lewis, C. W.;
Lonneman, W. A.; Seila, R. L.; Stevens, R. K. (1993) A method for separating volatile organic
carbon from 0.1 m3 of air to identify sources of ozone precursors via isotope (14C) measurements.
In: Measurement of toxic and related air pollutants: proceedings of the 1993 EPA/A&WMA
international symposium; May; Durham, NC. Pittsburgh, PA: Air & Waste Management
Association; pp. 585-603. (A&WMA specialty conference no. VIP-34).
Knispel, R.; Koch, R.; Siese, M.; Zetzsch, C. (1990) Adduct formation of OH radicals with benzene, toluene,
and phenol and consecutive reactions of the adducts with NOt and O2. Ber. Bunsen-Ges. Phys. Cheni.
94: 1375-1379.
Kok, G. L.; Darnall, K. R.; Winer, A. M.; Pitts, J. N., Jr.; Gay, B. W. (1978a) Ambient air measurements of
hydrogen peroxide in the California south coast air basin. Environ. Sci. Technol. 12: 1077-1080.
Kok, G. L.; Holler, T. P.; Lopez, M. B.; Nachtrieb, H. A.; Yuan, M. (1978b) Chemiluminescent method for
determination of hydrogen peroxide in the ambient atmosphere. Environ. Sci. Technol.
12: 1072-1076.
Kok, G. L.; Thompson, K.; Lazrus, A. L.; McLaren, S. E. (1986) Derivatization technique for the determination
of peroxides in precipitation. Anal. Chem. 58: 1192-1194.
Kok, G. L.; Walega, J. G.; Heikes, B. G.; Lind, J. A.; Lazrus, A. L. (1990) Measurements of hydrogen
peroxide and formaldehyde in Glendora, California. Aerosol Sci. Technol. 12: 49-55.
Kondo, Y.; Matthews, W. A.; Iwata, A.; Morita, Y.; Takagi, M. (1987) Aircraft measurements of oxides of
nitrogen along the eastern rim of the Asian continent: winter observations. J. Atmos. Chem.
5: 37-58.
Korsog, P. E.; Wolff, G. T. (1991) An examination of urban ozone trends in the northeastern U.S. (1973-1983)
using a robust statistical method. Atmos. Environ. Part B 25: 41-51.
Kosmus, W. (1985) Summation method for monitoring nitrogen oxides. Int. J. Environ. Anal. Chem.
22: 269-279.
Kotzias, D.; Hjorth, J. L.; Skov, H. (1989) A chemical mechanism for dry deposition—the role of biogenic
hydrocarbon (terpene) emissions in the dry deposition of O3, SO2 and NOX in forest areas. Toxicol.
Environ. Chem. 20-21: 95-99.
Kumar, S.; Chock, D. P. (1984) An update on oxidant trends in the south coast air basin of California. Atmos.
Environ. 18: 2131-2134.
Kuntasai, G.; Chang, T. Y. (1987) Trends and relationships of O3 NOX and HC in the south coast air basin of
California. JAPCA 37: 1158-1163.
Kuntz, R.; Lonneman, W.; Namie, G.; Hull, L. A. (1980) Rapid determination of aldehydes in air analyses.
Anal. Lett. 13: 1409-1415.
3-196
-------
Kuo, Y.-H.; Skumanich, M.; Haagenson, P. L.; Chang, J. S. (1985) The accuracy of trajectory models as
revealed by the observing system simulation experiments. Mon. Weather Rev. 113: 1852-1867.
Lamb, R. G. (1983) Regional scale (1000 km) model of photochemical air pollution. Part 1. Theoretical
formulation. Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental
Sciences Research Laboratory; report no. EPA-600/3-83-035. Available from: NTIS, Springfield,
VA; PB83-207688.
Lamb, B.; Westberg, H.; Allwine, G.; Quarles, T. (1985) Biogenic hydrocarbon emissions from deciduous and
coniferous trees in the United States. J. Geophys. Res. 90: 2380-2390.
Lamb, B.; Guenther, A.; Gay, D.; Westberg, H. (1987) A national inventory of biogenic hydrocarbon emissions.
Atmos. Environ. 21: 1695-1705.
Lamb, B.; Gay, D.; Westberg, H.; Pierce, T. (1993) A biogenic hydrocarbon emission inventory for the U.S.A.
using a simple forest canopy model. Atmos. Environ. Part A 27: 1673-1690.
Lammel, G.; Perner, D. (1988) The atmospheric aerosol as a source of nitrous acid in the polluted atmosphere. J.
Aerosol Sci. 19: 1199-1202.
Langner, J.; Rodhe, H.; Olofsson, M. (1990) Parameterization of subgrid scale vertical tracer transport in a global
two-dimensional model of the troposphere. J. Geophys. Res. [Atmos.] 95: 13,691-13,706.
Larson, S. M.; Cass, G. R.; Gray, H. A. (1989) Atmospheric carbon particles and the Los Angeles visibility
problem. Aerosol Sci. Technol. 10: 118-130.
Lawson, D. R.; Groblicki, P. J.; Stedman, D. H.; Bishop, G. A.; Guenther, P. L. (1990) Emissions froniin-use
motor vehicles in Los Angeles: a pilot study of remote sensing and the inspection and maintainence
program. J. Air Waste Manage. Assoc. 40: 1096-1105.
Lazrus, A. L.; Kok, G. L.; Gitlin, S. N.; Lind, J. A.; McLaren, S. E. (1985) Automated fluorometric method for
hydrogen peroxide in atmospheric precipitation. Anal. Chem. 57: 917-922.
Lazrus, A. L.; Kok, G. L.; Lind, J. A.; Gitlin, S. N.; Heikes, B. G.; Shelter, R. E. (1986) Automated
fluorometric method for hydrogen peroxide in air. Anal. Chem. 58: 594-597.
Lee, J. H.; Tang, I. N.; Weinstein-Lloyd, J. B. (1990) Nonenzymatic method for the determination of hydrogen
peroxide in atmospheric samples. Anal. Chem. 62: 2381-2384.
Lee, J. H.; Chen, Y.; Tang, I. N. (1991) Heterogeneous loss of gaseous H2Q in an atmospheric air sampling
system. Environ. Sci. Technol. 25: 339-342.
Lee, J. H.; Leahy, D. P.; Tang, I. N.; Newman, L. (1993) Measurement and speciation of gas phase peroxides in
the atmosphere. J. Geophys. Res. [Atmos.] 98: 2911-2915.
Lee, J. H.; Tang, I. N.; Weinstein-Lloyd, J. B.; Harper, E. B. (1994) Improved nonenzymatic method for the
determination of gas-phase peroxides. Environ. Sci. Technol. 28: 1180-1185.
Lelieveld, J.; Crutzen, P. J. (1990) Influences of cloud photochemical processes on tropospheric ozone. Nature
(London) 343: 227-233.
Lelieveld, J.; Crutzen, P. J. (1991) The role of clouds in tropospheric photochemistry. J. Atmos. Chem. 12: 229-
267.
3-197
-------
Lesion, A.; Ollison, W. (1993) Estimated accuracy of ozone design values: are they compromised by method
interferences? In: Vostal, J. J., ed. Tropospheric ozone: nonattainment and design value issues
[proceedings of a U.S. EPA/A&WMA international specialty conference]; October 1992; Boston,
MA. Pittsburgh, PA: Air & Waste Management Association; pp. 541-560. (A&WMA transactions
series no. 23).
Levaggi, D. A.; Oyung, W.; Zerrudo, R. V. (1992) Noncryogenic concentration of ambient hydrocarbons for
subsequent nonmethane and volatile organic compound analysis. In: Measurement of toxic and related
air pollutants: proceedings of the 1992 U.S. EPA/A&WMA international symposium; May; Durham,
NC. Pittsburgh, PA: Air & Waste Management Association; pp. 857-863. (A&WMA publication
VIP-25).
Lewis, C. W.; Conner, T. L. (1991) Source reconciliation of ambient volatile organic compounds measured in the
Atlanta 1990 summer study: the mobile source component. Presented at: A&WMA specialty
conference on emission inventory issues in the 1990's; September; Durham, NC. Pittsburgh, PA: Air
& Waste Management Association; pp. 514-523. (A&WMA publication VIP-22).
Lewis, C. W.; Baunigardner, R. E.; Stevens, R. K.; Claxton, L. D.; Lewtas, J. (1988) Contribution of
woodsmoke and motor vehicle emissions to ambient aerosol mutagenicity. Environ. Sci. Technol.
22: 968-971.
Lewis, C. W.; Stevens, R. K.; Zweidinger, R. B.; Claxton, L. D.; Barraclough, D.; Klouda, G. A. (1991)
Source apportionment of mutagenic activity of fine particle organics in Boise, Idaho. Presented at:
84th annual meeting and exhibition of the Air & Waste Management Association; June; Vancouver,
BC, Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-131.3.
Lewis, C. W.; Conner, T. L.; Stevens, R. K.; Collins, J. P.; Henry, R. C. (1993) Receptor modeling of volatile
hydrocarbons measured in the 1990 Atlanta ozone precursor study. Presented at: 86th annual meeting
and exhibition of the Air & Waste Management Association; June; Denver, CO. Pittsburgh, PA: Air
& Waste Management Association; paper no. 93-TP-58.04.
Lipari, P.; Swarin, S. J. (1982) Determination of formaldehyde and other aldehydes in automobile exhaust with an
improved 2,4-dinitrophenylhydrazine method. J. Chromatogr. 247: 297-306.
Liu, M.-K.; Seinfeld, J. H. (1975) On the validity of grid and trajectory models of urban air pollution. Atmos.
Environ. 9: 555-574.
Liu, S. C.; Trainer, M. (1988) Responses of the tropospheric ozone and odd hydrogen radicals to column ozone
change. J. Atmos. Chem. 6: 221-233.
Liu, L.-J. S.; Koutrakis, P.; Suh, H. H.; Mulik, J. D.; Burton, R. M. (1992) Use of personal measurements for
ozone exposure assessment—a pilot study. In: Measurement of toxic and related air pollutants:
proceedings of the 1992 U.S. EPA/A&WMA international symposium; May; Durham, NC.
Pittsburgh, PA: Air & Waste Management Association; pp. 962-967. (A&WMA publication
VIP-25).
Logan, J. A. (1985) Tropospheric ozone: seasonal behavior, trends, and anthropogenic influence. J. Geophys.
Res. [Atmos.] 90: 10,463-10,482.
Logan, J. A.; Prather, M. J.; Wofsy, S. C.; McElroy, M. B. (1981) Tropospheric chemistry: a global
perspective. J. Geophys. Res. C: Oceans Atmos. 86: 7210-7254.
Lonneman, W. A. (1977) PAN measurement in dry and humid atmospheres. Environ. Sci. Technol. 11: 194-195.
3-198
-------
Lonneman, W. A.; Seila, R. L. (1993) Hydrocarbon compositions in Los Angeles and New York 20 years later.
Presented at: International symposium on measurement of toxic and related air pollutants. Pittsburgh,
PA: Air & Waste Management Association.
Lonneman, W. A.; Kopczynski, S. L.; Darley, P. E.; Sutterfield, F. D. (1974) Hydrocarbon composition of
urban air pollution. Environ. Sci. Technol. 8: 229-236.
Lonneman, W. A.; Bufalini, J. J.; Namie, G. R. (1982) Calibration procedure for PAN based on its thermal
decomposition in the presence of nitric oxide. Environ. Sci. Technol. 16: 655-660.
Lonneman, W. A.; Seila, R. L.; Meeks, S. A. (1986) Non-methane organic composition in the Lincoln Tunnel.
Environ. Sci. Technol. 20: 790-796.
Lonneman, W. A.; Seila, R. L.; Ellenson, W. (1989) Speciated hydrocarbon and NOX comparisons at SCAQS
source and receptor sites. Presented at: 82nd annual meeting and exhibition of the Air & Waste
Management Association; June; Anaheim, CA. Pittsburgh, PA: Air & Waste Management
Association; paper no. 89-152.5.
Ludwig, F. L.; Reiter, E.; Shelar, E.; Johnson, W. B. (1977) The relation of oxidant levels to precursor
emissions and meteorological features: v. I, analysis and findings. Research Triangle Park, NC:
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; report
no. EPA-450/3-77-022a. Available from: NTIS, Springfield, VA; PB-275 001.
Lurniann, F. W.; Lloyd, A. C.; Atkinson, R. (1986) A chemical mechanism for use in long-range transport/acid
deposition computer modeling. J. Geophys. Res. [Atmos.] 91: 10,905-10,936.
Lyons, W. A.; Olsson, L. E. (1972) Mesoscale air pollution transport in the Chicago lake breeze. J. Air Pollut.
Control Assoc. 22: 876-881.
Lyons, W. A.; Calby, R. H.; Keen, C. S. (1986) The impact of mesoscale convective systems on regional
visibility and oxidant distributions during persistent elevated pollution episodes. J. Clim. Appl.
Meteorol. 25: 1518-1531.
Lyons, W. A.; Tremback, C. J.; Tesche, T. W. (1991) Lake Michigan ozone study prognostic modeling: model
performance evaluation and sensitivity testing. Crested Butte/ Fort Collins, CO: Alpine Geophysics,
ASTeR, Inc.
Macdonald, A. M.; Banic, C. M.; Leaitch, W. R.; Puckett, K. J. (1993) Evaluation of the Eulerian Acid
Deposition and Oxidant Model (ADOM) with summer 1988 aircraft data. Atmos. Environ.
Part A 27: 1019-1034.
Mackay, D. (1991) Multimedia environmental models: the fugacity approach. Chelsea, MI: Lewis Publishers.
Mackay, G. L; Schiff, H. I. (1987a) Methods comparison measurements during the carbonaceous species methods
comparison study, Glendora, California, August 1986: tunable diode laser absorption spectrometer
measurements of HCHO, H2O2 and HNO3. Sacramento, CA: California State Air Resources Board;
field report no. ARB-R-87/318. Available from: NTIS, Springfield, VA; PB88-133269.
Mackay, G. L; Schiff, H. I. (1987b) Reference measurements of HNO, and NOj by tunable diode laser absorption
spectroscopy. In: Proceedings of the 1987 EPA/APCA symposium on measurement of toxic and
related air pollutants; May; Research Triangle Park, NC. Pittsburgh, PA: Air Pollution Control
Association; pp. 367-372. (APCA publication VIP-8).
3-199
-------
Mackay, G. I.; Mayne, L. K.; Schiff, H. I. (1990) Measurements of H2O, and HCHO by tunable diode laser
absorption spectroscopy during the 1986 Carbonaceous Species Methods Comparison Study in
Glendora, California. Aerosol Sci. Technol. 12: 56-63.
MacKenzie, A. R.; Harrison, R. M.; Colbeck, I.; Hewitt, C. N. (1991) The role of biogenic hydrocarbons in the
production of ozone in urban plumes in southeast England. Atnios. Environ. Part A 25: 351-359.
Maeda, Y.; Aoki, K.; Munemori, M. (1980) Chemiluminescence method for the determination of nitrogen
dioxide. Anal. Chem. 52: 307-311.
Main, H. H.; Lurniann, F. W. (1993) Ambient versus emission inventory NMOC speciation during the SCAQS.
In: Southern California Air Quality Study data analysis: proceedings of an international specialty
conference; July 1992; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management Association;
pp. 49-54. (A&WMA publication VIP-26).
Martin, R. S.; Westberg, H.; Allwine, E.; Ashman, L.; Farmer, J. C.; Lamb, B. (1991) Measurement of
isoprene and its atmospheric oxidation products in a central Pennsylvania deciduous forest. J. Atmos.
Chem. 13: 1-32.
Mason, J. P.; Kirk, L; Windsor, C. G.; Tipler, A.; Spragg, R. A.; Rendle, M. (1992) A novel algorithm for
chromatogram matching in qualitative analysis. J. High Resolut. Chromatogr. 15: 539-547.
Mast, G. M.; Saunders, H. E. (1962) Research and development of the instrumentation of ozone sensing.
ISA Trans. 1: 325-328.
Mathur, R.; Schere, K. L. (1993) A regional modeling analysis of the dependencies of atmospheric oxidants
to perturbations in NOX and hydrocarbon emissions. Presented at: AMS special session on
atmospheric chemistry; January; Anaheim, CA. Boston, MA: American Meteorological Society.
Mayrsohn, H.; Crabtree, J. H. (1976) Source reconciliation of atmospheric hydrocarbons. Atmos. Environ.
10: 137-143.
Mayrsohn, H.; Crabtree, J. H.; Kuranioto, M.; Sothern, R. D.; Mano, S. H. (1977) Source reconciliation of
atmospheric hydrocarbons 1974. Atmos. Environ. 11: 189-192.
McClenny, W. A. (1993) Instrumentation to meet requirements for measurement of ozone precursor hydrocarbons
in the U. S. A. In: Proceedings of the international conference on volatile organic compounds;
October; London, United Kingdom.
McClenny, W. A.; Pleil, J. D.; Holdren, M. W.; Smith, R. N. (1984) Automated cryogenic preconcentration and
gas chromatographic determination of volatile organic compounds in air. Anal. Chem.
56: 2947-2951.
McClenny, W. A.; Pleil, J. D.; Evans, G. P.; Oliver, K. D.; Holdren, M. W.; Winberry, W. T. (1991a)
Canister-based method for monitoring toxic VOCs in ambient air. J. Air Waste Manage. Assoc.
41: 1308-1318.
McClenny, W. A.; Yarns, J. L.; Daughtridge, J. V. (1991b) The emergence of automated gas chromatographs
as air quality network monitors for volatile organic compounds. Presented at: 84th annual meeting
and exhibition of the Air & Waste Management Association; June; Vancouver, BC, Canada.
Pittsburgh, PA: Air & Waste Management Association; paper no. 91-68.5.
McElroy, F. P.; Thompson, V. L. (1975) Hydrocarbon measurement discrepancies among various analyzers using
flame-ionization detectors. Research Triangle Park, NC: U.S. Environmental Protection Agency,
3-200
-------
Environmental Monitoring and Support Laboratory; report no. EPA-600/4-75-010. Available from:
NTIS, Springfield, VA; PB-247821.
McElroy, F. P.; Thompson, V. L.; Holland, D. M.; Lonneman, W. A.; Seila, R. L. (1986) Cryogenic
preconcentration-direct FID method for measurement of ambient NMOC: refinement and comparison
with GC speciation. J. Air Pollut. Control Assoc. 36: 710-714.
Mcllveen, R. (1992) Fundamentals of weather and climate. London, United Kingdom: Chapman & Hall.
McKeen, S. A.; Hsie, E.-Y.; Trainer, M.; Tallamraju, R.; Liu, S. C. (1991) A regional model study of the ozone
budget in the eastern United States. J. Geophys. Res. [Atmos.] 96: 10,809-10,845.
McNair, L.; Russell, A.; Odman, M. T. (1992) Airshed calculation of the sensitivity of pollutant formation to
organic compound classes and oxygenates associated with alternative fuels. J. Air Waste Manage.
Assoc. 42: 174-178.
McNair, L. A.; Russell, A. G.; Odman, M. T.; Croes, B. E.; Kao, L. (1994) Airshed model evaluation of
reactivity adjustment factors calculated with the maximum incremental reactivity scale for
transitional-low emission vehicles. J. Air Waste Manage. Assoc. 44: 900-907.
McNally, D. E. (1990) Incorporation of four-dimensional data assimilation into the Colorado State University
mesoscale model. Davis, CA: University of California, Department of Land, Air and Water; project
no. UCD201.
McRae, G. J.; Russell, A. G. (1984) Dry deposition of nitrogen-containing species. In: Hicks, B. B.,
ed. Deposition both wet and dry. Boston, MA: Butterworth Publishers; pp. 153-193. (Teasley, J. L,
ed. Acid precipitation series: v. 4).
McRae, G. J.; Seinfeld, J. H. (1983) Development of a second-generation mathematical model for urban air
pollution—II. evaluation of model performance. Atmos. Environ. 17: 501-522.
McRae, G. J.; Goodin, W. R.; Seinfeld, J. H. (1982a) Development of a second-generation mathematical model
for urban air pollution—I. model formulation. Atmos. Environ. 16: 679-696.
McRae, G. J.; Goodin, W. R.; Seinfeld, J. H. (1982b) Mathematical modeling of photochemical air pollution.
Pasadena, CA: California Institute of Technology; Environmental Quality Laboratory report no. 18.
McRae, G. J.; Goodin, W. R.; Seinfeld, J. H. (1982c) Numerical solution of the atmospheric diffusion equation
for chemically reacting flows. J. Comp. Phys. 45: 1-42.
Meagher, J. P.; Olszyna, K. J.; Weatherford, F. P.; Mohnen, V. A. (1990) The availability of H,q, and O, for
aqueous phase oxidation of SO2. The question of linearity. Atmos. Environ. Part A 24: 1825-1829.
Meyer, C. P.; Elsworth, C. M.; Galbally, I. E. (1991a) Water vapor interference in the measurement of ozone in
ambient air by ultraviolet absorption. Rev. Sci. Instrum. 62: 223-228.
Meyer, E. L.; Possiel, N. C.; Doll, D. C.; Baugues, K. A.; Baldridge, K. W. (1991b) A summary of ROMNET
results and outputs. In: Proceedings of the seventh joint AMS/A&WMA conference on applications
of air pollution meteorology; January; New Orleans, LA. Pittsburgh, PA: Air & Waste Management
Association; pp. 246-249.
Meyrahn, H.; Moortgat, G. K.; Warneck, P. (1982) The photolysis of acetaldehyde under atmospheric conditions.
In: 15th informal conference on photochemistry; June-July; Stanford, CA.
3-201
-------
Meyrahn, H.; Pauly, J.; Schneider, W.; Warneck, P. (1986) Quantum yields for photodissociation of acetone in
air and an estimate for the life time of acetone in the lower troposhere. J. Atnios. Chem. 4: 277-291.
Meyrahn, H.; Helas, G.; Warneck, P. (1987) Gas chromatographic determination of peroxyacetyl nitrate: two
convenient calibration techniques. J. Atmos. Chem. 5: 405-415.
Michie, R. M., Jr.; Sokash, J. A.; Fritschel, B. P.; McElroy, F. P.; Thompson, V. L. (1983) Performance test
results and comparative data for designated reference methods for nitrogen dioxide. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems
Laboratory; report no. EPA-600/4-83-019. Available from: NTIS, Springfield, VA; PB83-200238.
Middleton, P.; Chang, J. S. (1990) Analysis of RADM gas concentration predictions using OSCAR and NEROS
monitoring data. Atmos. Environ. Part A 24: 2113-2125.
Middleton, P.; Chang, J. S.; del Corral, J. C.; Geiss, H.; Rosinski, J. M. (1988) Comparison of RADM and
OSCAR precipitation chemistry data. Atmos. Environ. 22: 1195-1208.
Middleton, P.; Chang, J. S.; Beauharnois, M.; Hash, L.; Binkowski, F. S. (1993) The role of nitrogen oxides in
oxidant production as predicted by the Regional Acid Deposition Model (RADM). In: Vanarsdale,
A., ed. The role and importance of NOX emission control: refereed papers from the international
symposium; May-June 1990; Cambridge, MA. Water Air Soil Pollut. 67: 133-159.
Milford, J. B.; Russell, A. G.; McRae, G. J. (1989) A new approach to photochemical pollution control:
implications of spatial patterns in pollutant responses to reductions in nitrogen oxides and reactive
organic gas emissions. Environ. Sci. Technol. 23: 1290-1301.
Milford, J. B.; Gao, D.; Odman, M. T.; Russell, A. G.; Possiel, N. C.; Scheffe, R. D.; Pierce, T. E.; Schere,
K. L. (1992) Air quality responses to NOX reductions: analysis of ROMNET results. Presented at:
85th annual meeting and exhibition of the Air & Waste Management Association; June; Kansas City,
MO. Pittsburgh, PA: Air & Waste Management Association; paper no. 92-89.2.
Milford, J. B.; Gao, D.; Sillman, S.; Blossey, P.; Russell, A. G. (1994) Total reactive nitrogen (NCp
as an indicator of the sensitivity of ozone to reductions in hydrocarbon and NOX emissions.
J. Geophys. Res. 99: 3533-3542.
Miller, D. P. (1988) Low-level determination of nitrogen dioxide in ambient air using the Palmes Tube. Atmos.
Environ. 22: 945-947.
Moore, G. E.; Daly, C.; Liu, M.-K.; Huang, S.-J. (1987) Modeling of mountain-valley wind fields in the
southern San Joaquin Valley, California. J. Clim. Appl. Meteorol. 26: 1230-1242.
Morris, R. E.; Kessler, R. C. (1991) Development of a variable grid regional oxidant model and application to the
San Joaquin valley. In: Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone
and the environment: papers from an international conference; March 1990; Los Angeles, CA.
Pittsburgh, PA: Air & Waste Management Association; pp. 588-610. (A&WMA transactions series
no. TR-19).
Morris, R. E.; Myers, T. C.; Douglas, S. G.; Yocke, M. A.; Mirabella, V. (1991) Development of a nested-grid
urban airshed model and application to southern California. Presented at: 84th annual meeting and
exhibition of the Air & Waste Management Association; June; Vancouver, BC, Canada. Pittsburgh,
PA: Air & Waste Management Association; paper no. 91-66.8.
Morris, R. E.; Yocke, M. A.; Myers, T. C.; Mirabella, V. (1992) Overview of the variable-grid Urban Airshed
Model (UAM-V). Presented at: 85th annual meeting and exhibition of the Air & Waste Management
3-202
-------
Association; June; Kansas City, MO. Pittsburgh, PA: Air & Waste Management Association; paper
no. 92-86.10.
Mottola, H. A.; Simpson, B. E.; Gorin, G. (1970) Absorptiometric determination of hydrogen peroxide in
submicrogram amounts with leuco crystal violet and peroxidase as catalyst. Anal. Chem.
42:410-411.
Mount, G. H.; Eisele, F. L. (1992) An interconiparison of tropospheric OH measurements at Fritz Peak
Observatory, Colorado. Science (Washington, DC) 256: 1187-1190.
Mukammal, E. I.; Neumann, H. H.; Gillespie, T. J. (1982) Meteorological conditions associated with ozone in
southwestern Ontario, Canada. Atmos. Environ. 16: 2095-2106.
Mulik, J. D.; Williams, D. (1986) Passive sampling devices for NO^. In: Proceedings of the 1986 EPA/APCA
symposium on measurement of toxic air pollutants; April; Raleigh, NC. Research Triangle Park,
NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; pp. 61-
70; report no. EPA-600/9-86-013. (APCA publication VIP-7). Available from: NTIS, Springfield,
VA; PB87-182713.
Mulik, J. D.; Williams, D. E. (1987) Passive sampling device measurements of NC«, in ambient air.
In: Proceedings of the 1987 EPA/APCA symposium on measurement of toxic and related air
pollutants; May; Research Triangle Park, NC. Pittsburgh, PA: Air Pollution Control Association;
pp. 387-397. (APCA publication VIP-8).
Mulik, J. D.; Lewis, R. G.; McClenny, W. A. (1989) Modification of a high-efficiency passive sampler to
determine nitrogen dioxide or formaldehyde in air. Anal. Chem. 61: 187-189.
Mulik, J. D.; Yarns, J. L.; Koutrakis, P.; Wolfson, M.; Bunyaviroch, A.; Williams, D. D.; Kromniller, K. G.
(1991) Using passive sampling devices to measure selected air volatiles for assessing ecological
change. In: Measurement of toxic and related air pollutants: proceedings of the 1991 U.S.
EPA/A&WMA international symposium, v. 1; May; Durham, NC. Pittsburgh, PA: Air & Waste
Management Association; pp. 285-290. (A&WMA publication VIP-21).
National Acid Precipitation Assessment Program. (1989) NAPAP assessment process compendium. Volume 5:
models planned for use in the NAPAP integrated assessment program. Washington, DC: National
Acid Precipitation Assessment Program. Available from: NTIS, Springfield, VA; PB91-
192955/XAB.
National Aeronautics and Space Administration. (1983) Assessment of techniques for measuring tropospheric
NxOy: proceedings of a workshop; August 1982; Palo Alto, CA. Hampton, VA: Langley Research
Center; NASA conference publication NASA-CP-2292. Available from: NTIS, Springfield, VA;
N84-13706.
National Research Council. (1991) Rethinking the ozone problem in urban and regional air pollution. Washington,
DC: National Academy Press.
Nederbragt, G. W.; van der Horst, A.; van Duijn, J. (1965) Rapid ozone determination near an accelerator.
Nature (London) 206: 87.
Nelson, P. P.; Quigley, S. M.; Smith, M. Y. (1983) Sources of atmospheric hydrocarbons in Sydney:
a quantitative determination using a source reconciliation technique. Atmos. Environ. 17: 439-449.
Nieboer, H.; Van Ham, J. (1976) Peroxyacetyl nitrate (PAN) in relation to ozone and some meteorological
parameters at Delft in The Netherlands. Atmos. Environ. 10: 115-120.
3-203
-------
Nielsen, T.; Hansen, A. M.; Thomsen, E. L. (1982) A convenient method for preparation of pure standards of
peroxyacetyl nitrate for atmospheric analyses. Atmos. Environ. 16: 2447-2450.
Niki, H.; Maker, P. D.; Savage, C. M.; Breitenbach, L. P. (1983) Atmospheric ozone-olefm reactions. Environ.
Sci. Technol. 17: 312A-322A.
Niki, H.; Maker, P. D.; Savage, C. M.; Breitenbach, L. P. (1985) An FTIR spectroscopic study of the reactions
Br + CH3CHO D HBr + CH3CO and CH3C(O)OO + NO2 D CH3C(O)OONO2 (PAN). Int. J. Chem.
Kinet. 17: 525-534.
Niki, H.; Maker, P. D.; Savage, C. M.; Breitenbach, L. P.; Hurley, M. D. (1987) FTIR spectroscopic study of
the mechanism for the gas-phase reaction between ozone and tetraniethylethylene. J. Phys. Chem.
91: 941-946.
Nondek, L.; Rodler, D. R.; Birks, J. W. (1992) Measurement of sub-ppv concentrations of aldehydes in a forest
atmosphere using a new HPLC technique. Environ. Sci. Technol. 26: 1174-1178.
Notholt, J.; Hjorth, J.; Raes, F. (1992a) Formation of HNO2 on aerosol surfaces during foggy periods in the
presence of NO and NO2. Atmos. Environ. Part A 26: 211-217.
Notholt, J.; Hjorth, J.; Raes, P.; Schrems, O. (1992b) Simultaneous long path field measurements of HNQ,
CH2O and aerosol. Ber. Bunsen Ges. Phys. Chem. 96: 290-293.
Noxon, J. F. (1983) NO3 and NO2 in the mid-Pacific troposphere. J. Geophys. Res. C: Oceans Atmos.
88: 11,017-11,021.
O'Shea, W. J.; Scheff, P. A. (1988) A chemical mass balance for volatile organics in Chicago. JAPCA 38: 1020-
1026.
Odman, M. T.; Russell, A. G. ( 1991) Multiscale modeling of pollutant transport and chemistry. J. Geophys. Res.
[Atmos.] 96: 7363-7370.
Odman, M. T.; Kumar, N.; Russell, A. G. (1992) A comparison of fast chemical kinetic solvers for air quality
modeling. Atmos. Environ. Part A 26: 1783-1789.
3-204
-------
Ogle, L. D.; Hall, R. C.; Crow, W. L.; Jones, A. E.; Gise, J. P. (1982) Development of preconcentration and
chromatographic procedures for the continuous and unattended monitoring of hydrocarbons in
ambient air. Presented at: 184th national meeting of the American Chemical Society; September;
Kansas City, MO. Austin, TX: Radian Corporation.
Ogle, L. D.; Brynier, D. A.; Jones, C. J.; Nahas, P. A. (1992) Moisture management techniques applicable to
whole air samples analyzed by method TO-14. In: Measurement of toxic and related air pollutants:
proceedings of the 1992 U.S. EPA/A&WMA international symposium; May; Durham, NC.
Pittsburgh, PA: Air & Waste Management Association; pp. 25-30. (A&WMA publication VIP-25).
Oke, T. R. (1978) Boundary layer climates. London, United Kingdom: Methuen & Co., Ltd.
Oliver, K. D.; Pleil, J. D.; McClenny, W. A. (1986) Sample integrity of trace level volatile organic compounds in
ambient air stored in SUMMA® polished canisters. Atmos. Environ. 20: 1403-1411.
Olszyna, K. J.; Meagher, J. P.; Bailey, E. M. (1988) Gas-phase, cloud and rain-water measurements of hydrogen
peroxide at a high-elevation site. Atmos. Environ. 22: 1699-1706.
Olszyna, K. J.; Bailey, E. M.; Simonaitis, R.; Meagher, J. G. (1994) O, and NOy relationships at a rural site.
J. Geophys. Res. [Atmos.] 99: 14,557-14,563.
Palmes, E. D.; Tomczyk, C. (1979) Personal sampler for NOX. Am. Ind. Hyg. Assoc. J. 40: 588-591.
Pandis, S. N.; Seinfeld, J. H. (1989) Sensitivity analysis of a chemical mechanism for aqueous-phase atmospheric
chemistry. J. Geophys. Res. [Atmos.] 94: 1105-1126.
Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1991) Aerosol formation in the photooxidation of
isoprene and D-pinene. Atmos. Environ. Part A 25: 997-1008.
Pandis, S. N.; Harley, R. A.; Cass, G. R.; Seinfeld, J. H. (1992) Secondary organic aerosol formation and
transport. Atmos. Environ. Part A 26: 2269-2282.
Pankow, J. F.; Bidleman, T. F. (1991) Effects of temperature, TSP and per cent non-exchangeable material in
determining the gas-particle partitioning of organic compounds. Atmos. Environ. Part A
25: 2241-2249.
Pankow, J. P.; Bidleman, T. F. (1992) Interdependence of the slopes and intercepts from log-log correlations of
measured gas-particle partitioning and vapor pressure -1. theory and analysis of available data.
Atmos. Environ. Part A 26: 1071-1080.
Papa, L. J.; Turner, L. P. (1972) Chromatographic determination of carbonyl compounds as their
2,4-dinitrophenylhydrazones: II. high pressure liquid chromatography. J. Chromatogr. Sci. 10: 747-
750.
Parrish, D. D.; Hahn, C. H.; Fahey, D. W.; Williams, E. J.; Bollinger, M. J.; Hiibler, G.; Buhr, M. P.;
Murphy, P. C.; Trainer, M.; Hsie, E. Y.; Liu, S. C.; Fehsenfeld, F. C. (1990) Systematic
variations in the concentration of NOX (NO plus NO2) at Niwot Ridge, Colorado. J. Geophys. Res.
[Atmos.] 95: 1817-1836.
Parrish, D. D.; Hahn, C. J.; Williams, E. J.; Norton, R. B.; Fehsenfeld, F. C.; Singh, H. B.; Shelter, J. D.;
Gandrud, B. W.; Ridley, B. A. (1992) Indications of photochemical histories of Pacific air masses
from measurements of atmospheric trace species at Point Arena, California. J. Geophys. Res.
[Atmos.] 97: 15,883-15,901.
3-205
-------
Parrish, D. D.; Buhr, M. P.; Trainer, M.; Norton, R. B.; Shimshock, J. P.; Fehsenfeld, F. C.; Anlauf, K. G.;
Bottenheim, J. W.; Tang, Y. Z.; Wiebe, H. A.; Roberts, J. M.; Tanner, R. L.; Newman, L.;
Bowersox, V. C.; Olszyna, K. J.; Bailey, E. M.; Rodgers, M. O.; Wang, T.; Berresheim, H.;
Roychowdhury, U. K.; Demerjian, K. L. (1993) The total reactive oxidized nitrogen levels and the
partitioning between the individual species at six rural sites in eastern North America. J. Geophys.
Res. [Atmos.] 98: 2927-2939.
Pate, B.; Jayanty, R. K. M.; Peterson, M. R.; Evans, G. F. (1992) Temporal stability of polar organic
compounds in stainless steel canisters. J. Air Waste Manage. Assoc. 42: 460-462.
Paulson, S. E.; Seinfeld, J. H. (1992a) Development and evaluation of a photooxidation mechanism for isoprene.
J. Geophys. Res. [Atmos.] 97: 20,703-20,715.
Paulson, S. E.; Seinfeld, J. H. (1992b) Atmospheric photochemical oxidation of 1-octene: OH, Q and O(3P)
reactions. Environ. Sci. Technol. 26: 1165-1173.
Paulson, S. E.; Flagan, R. C.; Seinfeld, J. H. (1992a) Atmospheric photooxidation of isoprene part I: the
hydroxyl radical and ground state atomic oxygen reactions. Int. J. Cheni. Kinet. 24: 79-101.
Paulson, S. E.; Flagan, R. C.; Seinfeld, J. H. (1992b) Atmospheric photooxidation of isoprene part II: the ozone-
isoprene reaction. Int. J. Chem. Kinet. 24: 103-125.
Penkett, S. A.; Blake, N. J.; Lightman, P.; Marsh, A. R. W.; Anwyl, P.; Butcher, G. (1993) The seasonal
variation of nonmethane hydrocarbons in the free troposphere over the North Atlantic Ocean:
possible evidence for extensive reaction of hydrocarbons with the nitrate radical. J. Geophys. Res.
[Atmos.] 98: 2865-2885.
Perschke, H.; Broda, E. (1961) Determination of very small amounts of hydrogen peroxide. Nature (London) 190:
257-258.
Peterson, J. T. (1976) Calculated actinic fluxes (290-700 nm) for air pollution photochemistry applications.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Sciences
Research Laboratory; report no. EPA-600/4-76-025. Available from: NTIS, Springfield, VA; PB-
255819.
Pickering, K. E.; Thompson, A. M.; Dickerson, R. R.; Luke, W. T.; MacNaniara, D. P.; Greenberg, J. P.;
Zimmerman, P. R. (1990) Model calculations of tropospheric ozone production potential following
bserved convective events. J. Geophys. Res. [Atmos.] 95: 14,049-14,062.
Pielke, R. A. (1974) A three-dimensional numerical model of the Seabreeze over south Florida. Mon. Weather
Rev. 102: 115-139.
Pielke, R. A. (1984) Mesoscale meteorological modeling. Orlando, PL: Academic Press.
Pielke, R. A. (1989) Status of subregional and mesoscale models, v. 2: niesoscale meteorological models in the
United States. Fort Collins, CO: Electric Power Research Institute; report no. EN-6649, v. 2.
Pierce, T. E.; Schere, K. L.; Doll, D. C.; Heilman, W. E. (1990) Evaluation of the regional oxidant model
(version 2.1) using ambient and diagnostic simulations. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Atmospheric Research and Exposure Assessment Laboratory;
report no. EPA-600/3-90/046. Available from: NTIS, Springfield, VA; PB90-225293/HSU.
Pierotti, D. J. (1990) Analysis of trace oxygenated hydrocarbons in the environment. J. Atmos. Chem.
10: 373-382.
3-206
-------
Pierotti, D.; Wofsy, S. C.; Jacob, D.; Rasmussen, R. A. (1990) Isoprene and its oxidation products: methacrolein
and methyl vinyl ketone. J. Geophys. Res. [Atmos.] 95: 1871-1881.
Pierson, W. R.; Gertler, A. W.; Bradow, R. L. (1990) Comparison of the SCAQS tunnel study with other
on-road vehicle emission data. J. Air Waste Manage. Assoc. 40: 1495-1504.
Pitts, J. N., Jr.; Biermann, H. W.; Atkinson, R.; Winer, A. M. (1984a) Atmospheric implications of
simultaneous nighttime measurements of NO3 radicals and HONO. Geophys. Res. Lett. 11: 557-560.
Pitts, J. N., Jr.; Sanhueza, E.; Atkinson, R.; Carter, W. P. L.; Winer, A. M.; Harris, G. W.; Plum, C. N.
(1984b) An investigation of the dark formation of nitrous acid in environmental chambers. Int.
J. Chem. Kinet. 16: 919-939.
Placet, M.; Battye, R. E.; Fehsenfeld, F. C.; Bassett, G. W. (1991) Emissions involved in acidic deposition
processes. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume I:
emissions, atmospheric processes, and deposition. Washington, DC: The U.S. National Acid
Precipitation Assessment Program. (State of science and technology report no. 1).
Plait, U.; Perner, D. (1980) Direct measurement of atmospheric CH2O, HNQ, O, and SQ by differential optical
absorption in the near UV. J. Geophys. Res. C: Oceans Atmos. 85: 7453-7458.
Pleil, J. D.; Oliver, K. D.; McClenny, W. A. (1987) Enhanced performance of Nafion dryers in removing water
from air samples prior to gas chromatographic analysis. JAPCA 37: 244-248.
Plum, C. N.; Sanhueza, E.; Atkinson, R.; Carter, W. P. L.; Pitts, J. N., Jr. (1983) OH radical rate constants and
photolysis rates of D-dicarbonyls. Environ. Sci. Technol. 17: 479-483.
Pollack, A. K.; Stockenius, T. E.; Haney, J. L.; Stocking, T. S.; Fieber, J. L.; Moezzi, M. (1988) Analysis of
historical ozone concentrations in the Northeast: volume I. San Rafael, CA: Systems Applications,
Inc.; report no. SYSAPP-88/192a.
Possiel, N. C.; Cox, W. M. (1993) The relative effectiveness of NOX and VOC strategies in reducing northeast
U.S. ozone concentrations. In: Vanarsdale, A., ed. The role and importance of NOx emission
control: refereed papers from the international symposium; May-June 1990; Cambridge, MA. Water
Air Soil Pollut. 67: 161-179.
Possiel, N. C.; Doll, D. C.; Baugues, K. A.; Baldridge, E. W.; Wayland, R. A. (1990) Impacts of regional
control strategies on ozone in the northeastern United States. Presented at: 83rd annual meeting and
exhibition of the Air & Waste Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air &
Waste Management Association; report no 90-93.3.
Possiel, N. C.; Wayland, R. A.; Wilson, J. H., Jr.; Laich, E. J.; Mullen, M. A. (1993) Predicted impacts of
1990 CAAA controls on Northeast U.S. ozone levels [draft]. Washington, DC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards.
Prinn, R.; Cunnold, D.; Rasmussen, R.; Simnionds, P.; Alyea, P.; Crawford, A.; Fraser, P.; Rosen, R. (1990)
Atmospheric emissions and trends of nitrous oxide deduced from 10 years of ALE-GAGE data.
J. Geophys. Res. [Atmos]. 95: 18,369-18,385.
Prinn, R.; Cunnold, D.; Simnionds, P.; Alyea, P.; Boldi, R.; Crawford, A.; Fraser, P.; Gutzler, D.;
Hartley, D.; Rosen, R.; Rasmussen, R. (1992) Global average concentration and trend for hydroxyl
radicals deduced from ALE/GAGE trichloroethane (methyl chloroform) data for 1978-1990. J.
Geophys. Res. [Atmos.] 97: 2445-2461.
3-207
-------
Purdue, L. J. (1993) Continuous monitoring of VOC precursors. Research Triangle Park,
NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment
Laboratory; report no. EPA/600/A-93/045. Available from: NTIS, Springfield, VA; PB93-167211.
Purdue, L. J.; Hauser, T. R. (1980) Review of U.S. Environmental Protection Agency NQ monitoring
methodology requirements. In: Lee, S. D., ed. Nitrogen oxides and their effects on health.
Ann Arbor, MI: Ann Arbor Science Publishers, Inc.; pp. 51-76.
Purdue, L. J.; Reagan, J. A.; Lonneman, W. A.; Lawlass, T. C.; Drago, R. J.; Zalaquet, G. M.;
Holdren, M. W.; Smith, D. L.; Pate, A. D.; Buxton, B. E.; Spicer, C. W. (1992) Atlanta ozone
precursor monitoring study data report. Research Triangle Park, NC: U.S. Environmental Protection
Agency, Atmospheric Research and Exposure Assessment Laboratory; report
no. EPA-600/R-92-157. Available from: NTIS, Springfield, VA; PB92-220656/REB.
Rao, S. T. (1987) Application of the urban airshed model to the New York metropolitan area. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
report no. EPA-450/4-87-011. Available from: NTIS, Springfield, VA; PB87-201422/HSU.
Rao, S. T.; Sistla, G. (1993) Efficacy of nitrogen oxides and hydrocarbons emissions controls in ozone attainment
strategies as predicted by the Urban Airshed Model. In: Vanarsdale, A., ed. The role and importance
of NOx emission control: refereed papers from the international symposium; May-June 1990;
Cambridge, MA. Water Air Soil Pollut. 67: 95-116.
Rao, S. T.; Sistla, G.; Twaddell, R. (1989) Photochemical modeling analysis of emission control strategies in the
New York metropolitan area. Washington, DC: U.S. Environmental Protection Agency; report no.
EPA-230/2-89/026.
Rasniussen, P. A. (1989) Southern California Air Quality Study. Part I. Hydrocarbon collection and analyses. Part
II Air toxics collection and analyses. Sacramento, CA: California Air Resources Board; final report,
agreement no. A6-179-32.
Rauhut, M. M.; Bollyky, L. J.; Roberts, B. G.; Loy, M.; Whitman, R. H.; lannotta, A. V.; Semsel, A. M.;
Clarke, R. A. (1967) Chemiluminescence from reactions of electronegatively substituted aryl oxalates
with hydrogen peroxide and fluorescent compounds. J. Am. Cheni. Soc. 89: 6515-6522.
Reckner, L. R. (1974) Survey of users of the EPA-reference method for measurement of non-methane
hydrocarbons in ambient air. Washington, DC: U.S. Environmental Protection Agency, Office of
Research and Development; report no. EPA-650/4-75-008. Available from: NTIS, Springfield, VA;
PB-247515.
Regener, V. H. (1960) On a sensitive method for the recording of atmospheric ozone. J. Geophys. Res. 65: 3975-
3977.
Regener, V. H. (1964) Measurement of atmospheric ozone with the chemilurninescent method. J. Geophys. Res.
69: 3795-3800.
Rehme, K. A.; Puzak, J. C.; Beard, M. E.; Smith, C. P.; Paur, R. J. (1981) Evaluation of ozone calibration
procedures [project summary]. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Environmental Monitoring Systems Laboratory; report no. EPA-600/S4-80-050. Available from:
NTIS, Springfield, VA; PB81-118911.
Reynolds, S. D.; Roth, P. M.; Seinfeld, J. H. (1973) Mathematical modeling of photochemical air pollution—
I. formulation of the model. Atmos. Environ. 7: 1033-1061.
3-208
-------
Reynolds, S. D.; Liu, M.-K.; Hecht, T. A.; Roth, P. M.; Seinfeld, J. H. (1974) Mathematical modeling of
photochemical air pollution—III. evaluation of the model. Atmos. Environ. 8: 563-596.
Reynolds, S. D.; Tesche, T. W.; Reid, L. E. (1979) An introduction to the SAI airshed model and its usage.
San Rafael, CA: Systems Applications, Inc.; report no. SAI-EF79-31.
Richman, E. E., Jr.; Green, A. H.; Wright, R. S.; Sickles, J. E., II. (1989) Laboratory and field evaluations of
extrasensitive sulfur dioxide and nitrogen dioxide analyzers for acid deposition monitoring. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure
Assessment Laboratory; report no. EPA/600/3-90/034. Available from: NTIS, Springfield, VA;
PB90-201062.
Ridley, B. A.; Howlett, L. C. (1974) An instrument for nitric oxide measurements in the stratosphere. Rev. Sci.
Instruni. 45: 742-746.
Ridley, B. A.; Carroll, M. A.; Torres, A. L.; Condon, E. P.; Sachse, G. W.; Hill, G. P.; Gregory, G. L.
(1988a) An intercomparison of results from ferrous sulphate and photolytic converter techniques for
measurements of NOX made during the NASA GTE/CITE 1 aircraft program. J. Geophys. Res.
[Atmos.] 93: 15,803-15,811.
Ridley, B. A.; Carroll, M. A.; Gregory, G. L.; Sachse, G. W. (1988b) NO and NO, in the troposphere:
technique and measurements in regions of a folded tropopause. J. Geophys. Res. [Atmos.]
93: 15,813-15,830.
Ridley, B. A.; Carroll, M. A.; Dunlap, D. D.; Trainer, M.; Sachse, G. W.; Gregory, G. L.; Condon, E. P.
(1989) Measurements of NOX over the eastern Pacific ocean and southwestern United States during
the spring 1984 NASA GTE aircraft program. J. Geophys. Res. [Atmos.] 94: 5043-5067.
Ripperton, L. A.; Worth, J. J. B.; Vukovich, P. M.; Decker, C. E. (1977) Research Triangle Institute studies
of high ozone concentrations in nonurban areas. In: Dimitriades, B., ed. International conference on
photochemical oxidant pollution and its control—proceedings: volume I; September 1976; Raleigh,
NC. Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Sciences
Research Laboratory; pp. 413-424; report no. EPA-600/3-77-001a. Available from: NTIS,
Springfield, VA; PB-264232.
Roberts, J. M. (1990) The atmospheric chemistry of organic nitrates. Atmos. Environ. Part A 24: 243-287.
Roberts, J. M.; Fajer, R. W.; Springston, S. R. (1989) Capillary gas chromatographic separation of alkyl nitrates
and peroxycarboxylic nitric anhydrides. Anal. Chem. 61: 771-772.
Robinson, N. P.; Pierson, W. R.; Gertler, A. W.; Sagebiel, J. C. (1996) Comparison of MOBILE4.1 and
MOBILES predictions with measurements of vehicle emission factors in Fort McHenry and
Tuscarora mountain tunnels. In: Parrish, D.; Trainer, M.; Rao, S. T.; Solomon, P. A., eds.
A&WMA specialty conference on regional photochemical measurements and modeling, part 2;
November 1993; San Diego, CA. Atmos. Environ. 30: 2257-2267.
Rodgers, M. O.; Davis, D. D. (1989) A UV-photofragmentation/laser-induced fluorescence sensor for the
atmospheric detection of HONO. Environ. Sci. Technol. 23: 1106-1112.
Rogers, J. D. (1990) Ultraviolet absorption cross sections and atmospheric photodissociation rate constants of
formaldehyde. J. Phys. Chem. 94: 4011-4015.
Rogge, W. P.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine
organic aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
3-209
-------
Rogge, W. P.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993) Sources of fine
organic aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel tracks.
Environ. Sci. Technol. 27: 636-651.
Roselle, S. J.; Schere, K. L. (1990) Sensitivity of the EPA regional oxidant model to biogenic hydrocarbon
emissions. Presented at: 83rd annual meeting and exhibition of the Air & Waste Management
Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; paper
no. 90-94.4.
Roselle, S. J.; Pierce, T. E.; Schere, K. L. (1991) The sensitivity of regional ozone modeling to biogenic
hydrocarbons. J. Geophys. Res. [Atmos.] 96: 7371-7394.
Roselle, S. J.; Schere, K. L.; Chu, S.-H. (1992) Estimates of ozone response to various combinations of NQ and
VOC emission reductions in the eastern United States. Presented at: 1992 quadrennial ozone
symposium; June; Charlottesville, VA.
Roth, P. (1992) Using photochemical models in developing attainment strategies: perspectives and problems.
In: Proceedings of the Electric Power Research Institute photochemical modeling workshop; August;
Cambridge, MA. Palo Alto, CA: Electric Power Research Institute.
Roth, P. M.; Blanchard, C. L.; Reynolds, S. D. (1990) The role of grid-based, reactive air quality modeling
in policy analysis: perspectives and implications, as drawn from a case study. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment
Laboratory; report no. EPA/600/3-89/082. Available from: NTIS, Springfield, VA; PB90-187204.
Rowland, F. S. (1990) Stratospheric ozone depletion by chlorofluorocarbons. Ambio 19: 281-292.
Rowland, F. S. (1991) Stratospheric ozone depletion. Annu. Rev. Phys. Chem. 42: 731-768.
Russell, A. G.; McCue, K. P.; Cass, G. R. (1988a) Mathematical modeling of the formation of
nitrogen-containing air pollutants. 1. Evaluation of an Eulerian photochemical model. Environ. Sci.
Technol. 22: 263-271.
Russell, A. G.; McCue, K. P.; Cass, G. R. (1988b) Mathematical modeling of the formation of
nitrogen-containing pollutants. 2. Evaluation of the effect of emission controls. Environ. Sci.
Technol. 22: 1336-1347.
Sakamaki, P.; Hatakeyama, S.; Akimoto, H. (1983) Formation of nitrous acid and nitric oxide in the
heterogeneous dark reaction of nitrogen dioxide and water vapor in a smog chamber. Int. J. Chem.
Kinet. 15: 1013-1029.
Sakugawa, H.; Kaplan, I. R.; Tsai, W.; Cohen, Y. (1990) Atmospheric hydrogen peroxide: does it share a role
with ozone in degrading air quality? Environ. Sci. Technol. 24: 1452-1462.
Salas, L. J.; Singh, H. B. (1986) Measurements of formaldehyde and acetaldehyde in the urban ambient air.
Atmos. Environ. 20: 1301-1304.
Samson, P. J.; Ragland, K. W. (1977) Ozone and visibility reduction in the Midwest: evidence for large-scale
transport. J. Appl. Meteorol. 16: 1101-1106.
Samson, P. J.; Shi, B. (1988) A meteorological investigation of high ozone values in American cities. Ann Arbor,
MI: University of Michigan, Space Physics Research Laboratory, Department of Atmospheric,
Oceanic, and Space Sciences.
3-210
-------
Saylor, R. D.; Peters, L. K.; Mathur, R. (1991) The STEM-II regional-scale acid deposition and photochemical
oxidant model—III. a study of mesoscale acid deposition in the lower Ohio River Valley. Atmos.
Environ. Part A 25: 2873-2894.
Scheff, P. A.; Wadden, R. A (1993) Receptor modeling of volatile organic compounds. 1. Emission inventory and
validation. Environ. Sci. Technol. 27: 617-625.
Scheffe, R. D.; Morris, R. E. (1990) Assessment of ozone precursor control measurements using the urban
airshed model in EPA's 5-city UAM study. Presented at: 83rd annual meeting and exhibition of the
Air & Waste Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
Management Association; paper no. 90-93.4.
Scheffe, R. D.; Morris, R. E. (1991) Overview of EPA's 5-city UAM study. In: Berglund, R. L.;
Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
Association; pp. 611-621. (A&WMA transactions series no. TR-19).
Scheffe, R. D.; Morris, R. E. (1993) A review of the development and application of the Urban Airshed Model.
Atmos. Environ. Part B 27: 23-39.
Schere, K. L.; Wayland, R. A. (1989a) Development and evaluation of the regional oxidant model for the
northeastern United States. In: Schneider, T.; Lee, S. D.; Wolters, G. J. R.; Grant, L. D., eds.
Atmospheric ozone research and its policy implications: proceedings of the 3rd US-Dutch
international symposium; May 1988; Nijmegen, The Netherlands. Amsterdam, The Netherlands:
Elsevier Science Publishers; pp. 613-622. (Studies in environmental science 35).
Schere, K. L.; Wayland, R. A. (1989b) EPA regional oxidant model (ROM2.0): evaluation on 1980 NEROS data
bases. Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research
and Exposure Assessment Laboratory; report no. EPA/600/3-89/057. Available from: NTIS,
Springfield, VA; PB89-200828/HSU.
Schiff, H. L; Hastie, D. R.; Mackay, G. L; Iguchi, T.; Ridley, B. A. (1983) Tunable diode laser systems for
measuring trace gases in tropospheric air: a discussion of their use and the sampling and calibration
procedures for NO, NO2, and HNO3. Environ. Sci. Technol. 17: 352A-364A.
Schiff, H. L; Mackay, G. L; Castledine, C.; Harris, G. W.; Tran, Q. (1986) A sensitive direct measurement NO,
instrument. In: Proceedings of the 1986 EPA/APCA symposium on measurement of toxic air
pollutants; April; Raleigh, NC. Research Triangle Park, NC: U.S. Environmental Protection
Agency, Environmental Monitoring Systems Laboratory; pp. 834-844; report no. EPA-600/9-86-013.
(APCA publication VIP-7). Available from: NTIS, Springfield, VA; PB87-182713.
Schiff, H. L; Harris, G. W.; Mackay, G. I. (1987) Measurement of atmospheric gases by laser absorption
spectrometry. In: Johnson, R. W.; Gordon, G. E.; Calkins, W.; Elzerman, A. Z., eds. The
chemistry of acid rain: sources and atmospheric processes [papers presented at 191st meeting of the
American Chemical Society; April 1986; New York, NY]. Washington, DC: American Chemical
Society; pp. 274-288. (Comstock, M. J., ed. ACS symposium series: v. 349).
Schwartz, S. E. (1989) Acid deposition: unraveling a regional phenomenon. Science (Washington, DC)
243: 753-763.
Scott, W. E.; Stephens, E. R.; Hanst, P. L.; Doerr, R. C. (1957) Further developments in the chemistry of the
atmosphere. Proc. Am. Pet. Inst. Sect. 3: 37: 171-183.
Scott, G.; Seitz, W. R.; Ambrose, J. (1980) Improved determination of hydrogen peroxide by measurement of
peroxyoxalate chemiluminescence. Anal. Chim. Acta 115: 221-228.
3-211
-------
Seaman, N. L. (1990) Meteorological modeling applied to regional air-quality studies using four-dimensional data
assimilation. In: Proceedings of the IBM summer institute on environmental modeling; July;
Oberlech, Austria.
Seaman, N. L. (1992) SARMAP meteorological model development and testing: annual report. California Air
Resources Board.
Seila, R. L.; Lonneman, W. A.; Meeks, S. A. (1989) Determination of Q to Q2 ambient air hydrocarbons in 39
U.S. cities from 1984 through 1986. Research Triangle Park, NC: U.S. Environmental Protection
Agency, Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/3-89/058.
Available from: NTIS, Springfield, VA; PB89-214142/HSU.
Seinfeld, J. H. (1986) Atmospheric chemistry and physics of air pollution. New York, NY: John Wiley and Sons.
Seinfeld, J. H. (1988) Ozone air quality models: a critical review. JAPCA 38: 616-645.
Sevcik, J. (1975) Detectors in gas chromatography. New York, NY: American Elsevier Publishing Company, Inc.
Sexton, K. (1982) Evidence of an additive effect for small city plumes. Presented at: 75th annual meeting of the
Air Pollution Control Association; June; New Orleans, LA. Pittsburgh, PA: Air Pollution Control
Association; paper no. 82-31.4.
Sexton, F. W.; Michie, R. A., Jr.; McElroy, F. P.; Thompson, V. L. (1981) Technical assistance document for
the calibration and operation of automated ambient non-methane organic compound analyzers.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring
Systems Laboratory; report no. EPA-600/4-81-015. Available from: NTIS, Springfield, VA;
PB82-147406.
Sexton, F. W.; Michie, R. M., Jr.; McElroy, F. P.; Thompson, V. L. (1982) A comparative evaluation of seven
automated ambient nonmethane organic compound analyzers. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
EPA-600/4-82-046. Available from: NTIS, Springfield, VA; PB82-230798.
Sheih, C. M.; Wesely, M. L.; Walcek, C. J. (1986) A dry deposition module for regional acid deposition.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Sciences
Research Laboratory; report no. EPA/600/3-86/037. Available from: NTIS, Springfield, VA;
PB86-218104/HSU.
Shen, J.; Tanner, R. L.; Kelly, T. J. (1988) Development of techniques for measurement of gas-phase hydrogen
peroxide. Upton, NY: U.S. Department of Energy, Brookhaven National Laboratory; report no.
BNL-52138. Available from: NTIS, Springfield, VA; DE88009362.
Shepson, P. B.; Bottenheim, J. W.; Hastie, D. R.; Venkatram, A. (1992a) Determination of the relative ozone
and PAN deposition velocities at night. Geophys. Res. Lett. 19: 1121-1124.
Shepson, P. B.; Hastie, D. R.; So, K. W.; Schiff, H. L; Wong, P. (1992b) Relationships between PAN, PPN and
O3 at urban and rural sites in Ontario. Atmos. Environ. Part A 26: 1259-1270.
Sherman, C. A. (1978) A mass-consistent model for wind fields over complex terrain. J. Appl. Meteorol. 17: 312-
319.
Shreffler, J. H. (1992) A survey of data from the continuous sites of the 1990 Atlanta ozone precursor study.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and
Assessment Laboratory; report no. EPA-600/R-92/202. Available from: NTIS, Springfield, VA;
PB93-124618.
3-212
-------
Shreffler, J. H. (1993) Comparison of nonmethane organic compound concentration data collected by two methods
in Atlanta. Air Waste 43: 1576-1584.
Sickles, J. E., II; Michie, R. M. (1987) Evaluation of the performance of sulfation and nitration plates. Atmos.
Environ. 21: 1385-1391.
Sickles, J. E., II; Wright, R. S. (1979) Atmospheric chemistry of selected sulfur-containing compounds: outdoor
smog chamber study—phase 1. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Environmental Sciences Research Laboratory; pp. 45-49; report no. EPA-600/7-79-227. Available
from: NTIS, Springfield, VA; PB81-141525.
Sickles, J. E., II; Grohse, P. M.; Hodson, L. L.; Salmons, C. A.; Cox, K. W.; Turner, A. R.; Estes, E. D.
(1990) Development of a method for the sampling and analysis of sulfur dioxide and nitrogen dioxide
from ambient air. Anal. Cheni. 62: 338-346.
Sigsby, J. E., Jr.; Tejada, S.; Ray, W.; Lang, J. M.; Duncan, J. W. (1987) Volatile organic compound emissions
from 46 in-use passenger cars. Environ. Sci. Technol. 21: 466-475.
Sillman, S.; Samson, P. J. (1995) Impact of temperature on oxidant photochemistry in urban, polluted rural and
remote environments. J. Geophys. Res. [Atmos.] 100: 11,497-11,508.
Sillman, S.; Logan, J. A.; Wofsy, S. C. (1990) The sensitivity of ozone to nitrogen oxides and hydrocarbons in
regional ozone episodes. J. Geophys. Res. [Atmos.] 95: 1837-1851.
Sillman, S.; Samson, P. J.; Masters, J. M. (1993) Ozone production in urban plumes transported over water:
photochemical model and case studies in the northeastern and midwestern United States. J. Geophys.
Res. [Atmos.] 98: 12,687-12,699.
Simonaitis, R.; Olszyna, K. J.; Meagher, J. F. (1991) Production of hydrogen peroxide and organic peroxides in
the gas phase reactions of ozone with natural alkenes. Geophys. Res. Lett. 18: 9-12.
Singh, H. B. (1980) Guidance for the collection and use of ambient hydrocarbon species data in development of
ozone control strategies. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards; report no. EPA-450/4-80-008. Available from: NTIS,
Springfield, VA; PB80-202120.
Singh, H. B.; Kanakidou, M. (1993) An investigation of the atmospheric sources and sinks of methyl bromide.
Geophys. Res. Lett. 20: 133-136.
Singh, H. B.; Salas, L. J. (1983) Methodology for the analysis of peroxyacetyl nitrate (PAN) in the unpolluted
atmosphere. Atmos. Environ. 17: 1507-1516.
Singh, H. B.; Viezee, W. (1988) Enhancement of PAN abundance in the Pacific marine air upon contact with
selected surfaces. Atmos. Environ. 22: 419-422.
Singh, H. B.; Viezee, W.; Johnson, W. B.; Ludwig, F. L. (1980) The impact of stratospheric ozone on
tropospheric air quality. J. Air Pollut. Control Assoc. 30: 1009-1017.
Skov, H.; Hjorth, J.; Lohse, C.; Jensen, N. R.; Restelli, G. (1992) Products and mechanisms of the reactions of
the nitrate radical (NO3) with isoprene, 1,3-butadiene and 2,3-dimethyl-l,3-butadiene in air. Atmos.
Environ. Part A 26: 2771-2783.
Slemr, P.; Harris, G. W.; Hastie, D. R.; Mackay, G. L; Schiff, H. I. (1986) Measurement of gas phase hydrogen
peroxide in air by tunable diode laser absorption spectroscopy. J. Geophys. Res. 91: 5371-5378.
3-213
-------
Smith, R. A.; Drummond, I. (1979) Trace determination of carbonyl compounds in air by gas chromatography of
their 2,4-dinitrophenylhydrazones. Analyst (London) 104: 875-877.
Smith, O. P.; Kleindienst, T. E.; Hudgens, E. E. (1989) Improved high-performance liquid chromatographic
method for artifact-free measurements of aldehydes in the presence of ozone using
2,4-dinitrophenylhydrazine. J. Chromatogr. 483: 431-436.
Smith, D. P.; Kleindienst, T. E.; Hudgens, E. E.; Mclver, C. D.; Bufalini, J. J. (1991a) The photoxidation of
methyl tertiary butyl ether. Int. J. Chem. Kinet. 23: 907-924.
Smith, D. L.; Holdren, M. W.; McClenny, W. A. (1991b) Design and operational characteristics of the
Chrompack model 9000 as an automated gas chromatograph. In: Measurement of toxic and related
air pollutants: proceedings of the 1991 U.S. EPA/A&WMA international symposium, v. 1; May;
Durham, NC. Pittsburgh, PA: Air & Waste Management Association; pp. 398-402. (A&WMA
publication VIP-21).
Smith, D. P.; Kleindienst, T. E.; Hudgens, E. E.; Mclver, C. D.; Bufalini, J. J. (1992) Kinetics and mechanism
of the atmospheric oxidation of ethyl tertiary butyl ether. Int. J. Chem. Kinet. 24: 199-215.
Solomon, P. A.; Fall, T.; Salmon, L.; Cass, G. R.; Gray, H. A.; Davidson, A. (1989) Chemical characteristics
of PM10 aerosols collected in the Los Angeles area. JAPCA 39: 154-163.
Spicer, C. W.; Ward, G. P.; Kenny, D. V.; Leslie, N. P.; Billick, I. H. (1991) Measurement of oxidized nitrogen
compounds in indoor air. In: Measurement of toxic and related air pollutants: proceedings of the
1991 U.S. EPA/A&WMA international symposium, v. 1; May; Durham, NC. Pittsburgh, PA: Air &
Waste Management Association; pp. 103-108. (A&WMA publication VIP-21).
Spicer, C. W.; Buxton, B. E.; Holdren, M. W.; Kelly, T. J.; Rust, S. W.; Ramamurthi, M.; Smith, D. L.; Pate,
A. D.; Sverdrup, G. M.; Chuang, J. C.; Shah, J. (1993) Variability and source attribution of
hazardous urban air pollutants, Columbus field study. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Atmospheric Research and Exposure Assessment Laboratory;
contract no. 68D80082.
Staehelin, J.; Hoigne, J. (1982) Decomposition of ozone in water: rate of initiation by hydroxide ions and
hydrogen peroxide. Environ. Sci. Technol. 16: 676-681.
Staehr, W.; Lahmann, W.; Weitkamp, C. (1985) Range-resolved differential absorption lidar: optimization of
range and sensitivity. Appl. Opt. 24: 1950-1956.
Stasiuk, W. N., Jr.; Coffey, P. E. (1974) Rural and urban ozone relationships in New York State. J. Air Pollut.
Control Assoc. 24: 564-568.
Stauff, J.; Jaeschke, W.; Schlogl, G. (1972) Chemilumineszenz des "Dioxetandions" [Chemiluminescence of
"dioxetanedione"]. Z. Naturforsch. B: Anorg. Chem. Org. Chem. Biochem. Biophys. Biol.
27: 1434-1435.
Stauffer, D. R.; Seaman, N. L. (1990) Use of four-dimensional data assimilation in a limited-area mesoscale
model. Part I: Experiments with synoptic-scale data. Mon. Weather Rev. 118: 1250-1277.
Stauffer, D. R.; Warner, T. T.; Seaman, N. L. (1985) A Newtonian "nudging" approach to four dimensional data
assimilation: use of SESAME-IV data in a mesoscale model. Presented at: 7th conference on
numerical weather prediction; Montreal, Canada.
Stedman, D. H.; Bishop, G. A.; Peterson, J. E.; Guenther, P. L.; McVey, I. P.; Beaton, S. P. (1991) On-road
carbon monoxide and hydrocarbon remote sensing in the Chicago area. Final report. Springfield, IL:
3-214
-------
Illinois Department of Energy and Natural Resources, Office of Research and Planning; report no.
ILENR/RE-AQ-91/14.
Stephens, E. R. (1964) Absorptivities for infrared determination of peroxyacyl nitrates. Anal. Cheni. 36: 928-929.
Stephens, E. R. (1969) The formation, reactions, and properties of peroxyacyl nitrates (PANs) in photochemical
air pollution. In: Pitts, J. N., Jr.; Metcalf, R. L., eds. Advances in environmental science and
technology: v. 1. New York, NY: Wiley-Interscience; pp. 119-146.
Stephens, E. R.; Price, M. A. (1973) Analysis of an important air pollutant: peroxyacetyl nitrate. J. Chem. Educ.
50: 351-354.
Stephens, E. R.; Hanst, P. L.; Doerr, R. C.; Scott, W. E. (1956a) Reactions of nitrogen dioxide and organic
compounds in air. Ind. Eng. Chem. 48: 1498-1504.
Stephens, E. R.; Scott, W. E.; Hanst, P. L.; Doerr, R. C. (1956b) Recent developments in the study of the
organic chemistry of the atmosphere. Proc. Am. Pet. Inst. Sect. 3 36: 288-297.
Stephens, E. R.; Darley, E. P.; Taylor, O. C.; Scott, W. E. (1961) Photochemical reaction products in air
pollution. Int. J. Air Water Pollut. 4: 79-100.
Stephens, E. R.; Burleson, F. R.; Cardiff, E. A. (1965) The production of pure peroxyacyl nitrates. J. Air Pollut.
Control Assoc. 15: 87-89.
Stevens, R. K.; Hodgeson, J. A. (1973) Applications of chemiluminescent reactions to the measurement of air
pollutants. Anal. Chem. 45: 443A-449A.
Stevens, R. K.; Drago, R. J.; Mamane, Y. (1993) A long path differential optical absorption spectrometer and
EPA-approved fixed-point methods intercomparison. Atmos. Environ. Part B 27: 231-236.
Stockburger, L.; Knapp, K. T.; Ellestad, T. G. (1989) Overview and analysis of hydrocarbon samples during the
summer Southern California Air Quality Study. Presented at: 82nd annual meeting and exhibition of
the Air & Waste Management Association; June; Anaheim, CA. Pittsburgh, PA: Air & Waste
Management Association; paper no. 89-139.1.
Stockwell, W. R.; Calvert, J. G. (1983) The mechanism of the HO-SQ, reaction. Atmos. Environ. 17: 2231-2235.
Stockwell, W. R; Lurmann, F. W. (1989) Intercomparison of the ADOM and RADM gas-phase chemical
mechanisms. Palo Alto, CA: Electric Power Research Institute.
Stockwell, W. R.; Middleton, P.; Chang, J. S.; Tang, X. (1990) The second generation Regional Acid Deposition
Model chemical mechanism for regional air quality modeling. J. Geophys. Res. [Atmos.] 95: 16,343-
16,367.
Stump, F. D.; Knapp, K. T.; Ray, W. D.; Snow, R.; Burton, C. (1992) The composition of motor vehicle organic
emissions under elevated temperature summer driving conditions (75 to 105 DP). J. Air Waste
Manage. Assoc. 42: 152-158.
Sun, W.-Y. (1986) Air pollution in a convective boundary layer. Atmos. Environ. 20: 1877-1886.
Svensson, R.; Ljungstroem, E.; Lindqvist, O. (1987) Kinetics of the reaction between nitrogen dioxide and water
vapour. Atmos. Environ. 21: 1529-1539.
Sweet, C. W.; Vermette, S. J. (1992) Toxic volatile organic compounds in urban air in Ilinois. Environ. Sci.
Technol. 26: 165-173.
3-215
-------
Systems Applications International. (1993) Evaluations of the EPA regional oxidant model for the July 2-13, 1988
ozone episode. Systems Applications International; report no. SYSAPP-93/043.
Tanner, R. L. (1985) Chemical transformations in acid rain, v. I: new methodologies for sampling and analysis of
gas-phase peroxide. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Atmospheric Sciences Research Laboratory; report no. EPA-600/3-85-016. Available from: NTIS,
Springfield, VA; PB85-173425.
Tanner, R. L.; Meng, Z. (1984) Seasonal variations in ambient atmospheric levels of formaldehyde and
acetaldehyde. Environ. Sci. Technol. 18: 723-726.
Tanner, R. L.; Shen, J. (1990) Measurement of hydrogen peroxide in ambient air by inipinger and diffusion
scrubber. Aerosol Sci. Technol. 12: 86-97.
Tanner, R. L.; Daum, P. H.; Kelly, T. J. (1983) New instrumentation for airborne acid rain research.
Int. J. Environ. Anal. Chem. 13: 323-335.
Tanner, R. L.; Markovits, G. Y.; Ferreri, E. M.; Kelly, T. J. (1986) Sampling and determination of gas-phase
hydrogen peroxide following removal of ozone by gas-phase reaction with nitric oxide. Anal. Chem.
58: 1857-1865.
Tesche, T. W. (1983) Photochemical dispersion modeling: review of model concepts and applications studies.
Environ. Int. 9: 465-489.
Tesche, T. W. (1987) Photochemical modeling of 1984 SCCCAMP oxidant episodes: protocol for model
selection, adaptation, and performance evaluation. Washington, DC: U.S. Environmental Protection
Agency.
Tesche, T. W. (1992) Emissions modeling: status and new directions. In: Proceedings of the Electric Power
Research Institute photochemical modeling workshop; August; Cambridge, MA. Palo Alto, CA:
Electric Power Research Institute.
Tesche, T. W. (1993) Workplan for the development and application of the SARMAP modeling system.
Sacramento, CA: California Air Resources Board; report no. AG-90/TS35.
Tesche, T. W.; Haney, J. L.; Morris, R. E. (1987) Performance evaluation of four grid-based dispersion models
in complex terrain. Atmos. Environ. 21: 233-256.
Tesche, T. W.; Georgopoulos, P.; Lurman, F. L.; Roth, P. M.; Seinfeld, J. H.; Cass, G. (1990) Improvement of
procedures for evaluating photochemical models. Sacramento, CA: California Air Resources Board;
report no. ARB-R-91/453. Available from: NTIS, Springfield, VA; PB91-160374.
Tesche, T. W.; Roth, P. M.; Reynolds, S. D.; Lurmann, F. W. (1993) Scientific assessment of the urban airshed
model (UAM-IV). Washington, DC: American Petroleum Institute; API publication no. 4556.
Thompson, A.M. (1992) The oxidizing capacity of the Earth's atmosphere: probable past and future changes.
Science (Washington, DC) 256: 1157-1165.
Thompson, A. M.; Stewart, R. W.; Owens, M. A.; Herwehe, J. A. (1989) Sensitivity of tropospheric oxidants to
global chemical and climate change. Atmos. Environ. 23: 519-532.
Tilden, J. W.; Seinfeld, J. H. (1982) Sensitivity analysis of a mathematical model for photochemical air pollution.
Atmos. Environ. 16: 1357-1364.
3-216
-------
Topham, L. A.; Mackay, G. I.; Schiff, H. I. (1992) Performance assessment of the portable and lightweight
LOZ-3 chemiluminescence type ozone monitor. In: Measurement of toxic and related air pollutants:
proceedings of the 1992 U.S. EPA/A&WMA international symposium; May; Durham, NC.
Pittsburgh, PA: Air & Waste Management Association; pp. 745-749. (A&WMA publication VIP-
25).
Torres, A. L. (1985) Nitric oxide measurements at a nonurban eastern United States site: Wallops instrument
results from July 1983 GTE/CITE mission. J. Geophys. Res. [Atmos.] 90: 12,875-12,880.
Trainer, M.; Williams, E. J.; Parrish, D. D.; Buhr, M. P.; Allwine, E. J.; Westberg, H. H.; Fehsenfeld, F. C.;
Liu, S. C. (1987) Models and observations of the impact of natural hydrocarbons on rural ozone.
Nature (London) 329: 705-707.
Trainer, M.; Buhr, M. P.; Curran, C. M.; Fehsenfeld, F. C.; Hsie, E. Y.; Liu, S. C.; Norton, R. B.; Parrish,
D. D.; Williams, E. J. (1991) Observations and modeling of the reactive nitrogen photochemistry at
a rural site. J. Geophys. Res. [Atmos.] 96: 3045-3063.
Trainer, M.; Parrish, D. D.; Buhr, M. P.; Norton, R. B.; Fehsenfeld, F. C.; Anlauf, K. G.; Bottenheim, J. W.;
Tang, Y. Z.; Wiebe, H. A.; Roberts, J. M.; Tanner, R. L.; Newman, L.; Bowersox, V. C.;
Meagher, J. P.; Olszyna, K. J.; Rodgers, M. O.; Wang, T.; Berresheini, H.; Demerjian, K. L.;
Roychowdhury, U. K. (1993) Correlation of ozone with NOy in photochemically aged air. J.
Geophys. Res. [Atmos.] 98: 2917-1925.
Tripoli, G. J.; Cotton, W. R. (1982) The Colorado State University three-dimensional cloud/mesoscale model—
1982. Part I: General theoretical framework and sensitivity experiments. J. Rech. Atmos.
16: 185-219.
Tsalkani, N.; Toupance, G. (1989) Infrared absorptivities and integrated band intensities for gaseous peroxyacetyl
nitrate (PAN). Atmos. Environ. 23: 1849-1854.
Tuazon, E. C.; Atkinson, R. (1989) A product study of the gas-phase reaction of methyl vinyl ketone with the OH
radical in the presence of NOX. Int. J. Chem. Kinet. 21: 1141-1152.
Tuazon, E. C.; Atkinson, R. (1990a) A product study of the gas-phase reaction of isoprene with the OH radical in
the presence of NOX. Int. J. Chem. Kinet. 22: 1221-1236.
Tuazon, E. C.; Atkinson, R. (1990b) A product study of the gas-phase reaction of methacrolein with the
OH radical in the presence of NOX. Int. J. Chem. Kinet. 22: 591-602.
Tuazon, E. C.; Graham, R. A.; Winer, A. M.; Easton, R. R.; Pitts, J. N., Jr.; Hanst, P. L. (1978) A kilometer
pathlength Fourier-transform infrared system for the study of trace pollutants in ambient and
synthetic atmospheres. Atmos. Environ. 12: 865-875.
Tuazon, E. C.; Winer, A. M.; Graham, R. A.; Pitts, J. N., Jr. (1980) Atmospheric measurements of trace
pollutants by kilometer-pathlength FT-IR spectroscopy. Adv. Environ. Sci. Technol. 10: 259-300.
Tuazon, E. C.; Winer, A. M.; Graham, R. A.; Pitts, J. N., Jr. (1981a) Atmospheric measurements of trace
pollutants: long path Fourier transform infrared spectroscopy. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Environmental Sciences Research Laboratory; report
no. EPA-600/3-81-026. Available from: NTIS, Springfield, VA; PB81-179848.
Tuazon, E. C.; Winer, A. M.; Pitts, J. N., Jr. (1981b) Trace pollutant concentrations in aniultiday smog episode
in the California South Coast Air Basin by long path length Fourier transform infrared spectroscopy.
Environ. Sci. Technol. 15: 1232-1237.
3-217
-------
Tuazon, E. C.; Carter, W. P. L.; Aschmann, S. M.; Atkinson, R. (1991) Products of the gas-phase reaction
of methyl terf-butyl ether with the OH radical in the presence of NOX. Int. J. Chem. Kinet.
23: 1003-1015.
Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin:
a descriptive analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A
25: 207-215.
Tyndall, G. S.; Ravishankara, A. R. (1991) Atmospheric oxidation of reduced sulfur species. Int. J. Chem. Kinet.
23: 483-527.
United Kingdom Photochemical Oxidants Review Group. (1993) Ozone in the United Kingdom 1993. London,
United Kingdom: Department of Environment, Technical Policy Branch.
U.S. Congress. (1990) Clean Air Act amendments of 1990: conference report to accompany S. 1630. Washington,
DC: U.S. Government Printing Office; report 101-952.
U.S. Department of Health, Education, and Welfare. (1970) Air quality criteria for photochemical oxidants.
Washington, DC: National Air Pollution Control Administration; publication no. AP-63. Available
from: NTIS, Springfield, VA; PB-190262/BA.
U.S. Environmental Protection Agency. (1978) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental
Criteria and Assessment Office; report no. EPA-600/8-78-004. Available from: NTIS, Springfield,
VA; PB80-124753.
U.S. Environmental Protection Agency. (1985) Compilation of air pollutant emission factors. Volume I: stationary
point and area sources. Volume II: mobile sources. 4th ed. Research Triangle Park, NC: Office of
Air Quality Planning and Standards; Ann Arbor, MI: Office of Mobile Sources; report nos. AP-42-
ED-4-VOL-1 and AP-42-ED-4-VOL-2. Available from: NTIS, Springfield, VA; PB86-124906 and
PB87-205266.
U.S. Environmental Protection Agency. (1986a) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental
Criteria and Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS,
Springfield, VA; PB87-142949.
U.S. Environmental Protection Agency. (1986b) Guideline on air quality models (revised). Research Triangle
Park, NC: Office of Air Quality Planning and Standards; report no. EPA/450/2-78/027R. Available
from: NTIS, Springfield, VA; PB86-245248.
U.S. Environmental Protection Agency. (1989) The 1985 NAPAP emissions inventory (version 2): development of
the national utility reference file. Research Triangle Park, NC: Air and Energy Engineering
Research Laboratory; report no. EPA/600/7-89/013A. Available from: NTIS, Springfield, VA;
PB90-132341.
U.S. Environmental Protection Agency. (1990a) User's guide for the urban airshed model, volume I: user's
manual for UAM (CB-IV). Research Triangle Park, NC: Office of Air Quality Planning and
Standards; report no. EPA-450/4-90-007A. Available from: NTIS, Springfield, VA; PB91-131227.
U.S. Environmental Protection Agency. (1990b) User's guide for the urban airshed model, volume II: user's
manual for the UAM (CB-IV) modeling system. Research Triangle Park, NC: Office of Air Quality
Planning and Standards; report no. EPA-450/4-90-007B. Available from: NTIS, Springfield, VA;
PB91-131235.
3-218
-------
U.S. Environmental Protection Agency. (1990c) User's guide for the urban airshed model, v. Ill: user's manual
for the diagnostic wind model. Research Triangle Park, NC: Office of Air Quality Planning and
Standards; report no. EPA-450/4-90-007C. Available from: NTIS, Springfield, VA; PB91-131243.
U.S. Environmental Protection Agency. (1990d) User's guide for the urban airshed model, volume V: description
and operation of the ROM-UAM interface program system. Research Triangle Park, NC: Office of
Air Quality Planning and Standards; report no. EPA-450/4-90-007E. Available from: NTIS,
Springfield, VA; PB91-131268.
U.S. Environmental Protection Agency. (1991a) National air quality and emissions trends report, 1989. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/450/4-91/003.
Available from: NTIS, Springfield, VA; PB91-172247/XAB.
U.S. Environmental Protection Agency. (1991b) Guideline for regulatory application of the urban airshed model.
Research Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-450/4-
91-013. Available from: NTIS, Springfield, VA; PB92-108760/HSU.
U.S. Environmental Protection Agency. (1991c) Technical assistance document for sampling and analysis ozone
precursors. Research Triangle Park, NC: Atmospheric Research and Exposure Assessment
Laboratory; report no. EPA/600/8-91/215. Available from: NTIS, Springfield, VA; PB92-122795.
U.S. Environmental Protection Agency. (1992a) National air pollutant emission estimates, 1900-1991. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/454/R-92/013.
Available from: NTIS, Springfield, VA; PB93-157808/XAB.
U.S. Environmental Protection Agency. (1992b) User's guide for the urban airshed model, volume IV: user's
manual for the emissions preprocessor system 2.0. Part A: core FORTRAN system. Part B: interface
and emission display system. Research Triangle Park, NC: Office of Air Quality Planning and
Standards; report no. EPA-450/4-90-007D(R). Available from: NTIS, Springfield, VA; PB93-
122380.
U.S. Environmental Protection Agency. (1993a) Regional interim emission inventories (1987-1991),
v. I: development methodologies. Research Triangle Park, NC: Office of Air Quality Planning and
Standards; report no. EPA-454/R-93-021a. Available from: NTIS, Springfield, VA;
PB93-236107/XAB.
U.S. Environmental Protection Agency. (1993b) Regional interim emission inventories (1987-1991),
v. II: emission summaries. Research Triangle Park, NC: Office of Air Quality Planning and
Standards; report no. EPA-454/R-93-021b. Available from: NTIS, Springfield, VA; PB93-236115.
U.S. Environmental Protection Agency. (1993c) Air quality criteria for oxides of nitrogen. Research Triangle
Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; report nos. EPA/600/8-91/049aF-cF. 3v. Available from: NTIS, Springfield, VA;
PB95-124533, PB95-124525, and PB95-124517.
U.S. Environmental Protection Agency. (1993d) Volatile organic compound/paniculate matter speciation database
management system (SPECIATE), version 1.5 (for microcomputers). Research Triangle Park, NC:
Office of Air Quality Planning and Standards; report no. EPA/SW/DK-92/043. Available from:
NTIS, Springfield, VA; PB93-505238.
U.S. Environmental Protection Agency (1994) National air pollutant emission trends, 1900-1993. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-94-027.
Available from: NTIS, Springfield, VA; PB95-171989REB.
3-219
-------
Ulrickson, B. L. (1988) Mesoscale circulations in the Los Angeles Basin: a numerical modeling study
[Ph.D. dissertation]. Seattle, WA: University of Washington.
Valente, R. J.; Thornton, F. C. (1993) Emissions of NO from soil at a rural site in central Tennessee. J. Geophys.
Res. [Atmos.] 98: 16,745-16,753.
Van Valin, C. C.; Ray, J. D.; Boatman, J. P.; Gunter, R. L. (1987) Hydrogen peroxide in air during winter over
the south-central United States. Geophys. Res. Lett. 14: 1146-1149.
Venkatram, A.; Karamchandani, P. K.; Misra, P. K. (1988) Testing a comprehensive acid deposition model.
Atmos. Environ. 22: 737-747.
Vierkorn-Rudolph, B.; Rudolph, J.; Diederich, S. (1985) Determination of peroxyacetylnitrate (PAN)
in unpolluted areas. Int. J. Environ. Anal. Chem. 20: 131-140.
Viezee, W.; Singh, H. B. (1982) Contribution of stratospheric ozone to ground-level ozone concentrations—
a scientific review of existing evidence. Research Triangle Park, NC: U.S. Environmental Protection
Agency, Environmental Science Research Laboratory; grant CR-809330010.
Viezee, W.; Johnson, W. B.; Singh, H. B. (1979) Airborne measurements of stratospheric ozone intrusions into
the troposphere over the United States [final report]. Menlo Park, CA: SRI International; SRI project
6690.
Viezee, W.; Johnson, W. B.; Singh, H. B. (1983) Stratospheric ozone in the lower troposphere—II. assessment of
downward flux and ground-level impact. Atmos. Environ. 17: 1979-1993.
Wackier, D. J.; Bayly, P. V. (1988) The effectiveness of emission controls on reducing ozone levels in
Connecticut from 1976 through 1987. In: Wolff, G. T.; Hanisch, J. L.; Schere, K., eds. The
scientific and technical issues facing post-1987 ozone control strategies: transactions of an APCA
international specialty conference; November 1987; Hartford, CT. Pittsburgh, PA: Air Pollution
Control Association; pp. 398-415.
Wadden, R. A.; Uno, L; Wakamatsu, S. (1986) Source discrimination of short-term hydrocarbon samples
measured aloft. Environ. Sci. Technol. 20: 473-483.
Wakim, P. G. (1989) Temperature-adjusted ozone trends for Houston, New York and Washington 1981-1987.
Presented at: 82nd annual meeting and exhibition of the Air & Waste Management Association; June;
Anaheim, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 89-35.1.
Walega, J. G.; Stedman, D. H.; Shelter, R. E.; Mackay, G. L; Iguchi, T.; Schiff, H. I. (1984) Comparison
of a chemiluminescent and a tunable diode laser absorption technique for the measurement of nitrogen
oxide, nitrogen dioxide, and nitric acid. Environ. Sci. Technol. 18: 823-826.
Wallace, L. A.; Ott, W. R. (1982) Personal monitors: a state-of-the-art survey. J. Air Pollut. Control Assoc.
32: 601-610.
Wallington, T. J.; Japar, S. M. (1991) Atmospheric chemistry of diethyl ether and ethyl ten-butyl ether. Environ.
Sci. Technol. 25: 410-415.
Wang, W.; Warner, T. T. (1988) Use of four-dimensional data assimilation by Newtonian relaxation and
latent-heat forcing to improve a mesoscale-model precipitation forecast: a case study. Mon. Weather
Rev. 116: 2593-2613.
3-220
-------
Wang, S.-C.; Paulson, S. E.; Grosjean, D.; Flagan, R. C.; Seinfeld, J. H. (1992) Aerosol formation and growth
in atmospheric organic/NOx systems-I. Outdoor smog chamber studies of C7- and C8-hydrocarbons.
Atmos. Environ. Part A 26: 403-420.
Warneck, P. (1991) Chemical reactions in clouds. Fresenius' J. Anal. Chem. 340: 585-590.
Warneck, P. (1992) Chemistry and photochemistry in atmospheric water drops. Ber. Bunsen Ges. Phys. Chem.
96: 454-460.
Warneck, P.; Zerbach, T. (1992) Synthesis of peroxyacetyl nitrate in air by acetone photolysis. Environ. Sci.
Technol. 26: 74-79.
Warren, G. J.; Babcock, G. (1970) Portable ethylene chemiluminescence ozone monitor. Rev. Sci. Instrum.
41: 280-282.
Watanabe, I.; Stephens, E. R. (1978) Reexamination of moisture anomaly in analysis of peroxyacetyl nitrate.
Environ. Sci. Technol. 12: 222-223.
Watson, J. G.; Robinson, N. P.; Chow, J. C.; Henry, R. C.; Kim, B.; Nguyen, Q. T.; Meyer, E. L.;
Pace, T. G. (1990) Receptor model technical series, v. Ill (1989 revision): CMB7 user's manual.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards; report no. EPA-450/4-90-004. Available from: NTIS, Springfield, VA; PB90-
185067.
Wei, L.; Weihan, S. (1987) Automated method for determination of hydrogen peroxide in air and atmospheric
precipitation. Huanjing Huaxue 6: 23-29. (CA 109: 16299; 1988).
Wendel, G. J.; Stedman, D. H.; Cantrell, C. A.; Damrauer, L. (1983) Luminol-based nitrogen dioxide detector.
Anal. Chem. 55: 937-940.
Weschler, C. J.; Mandich, M. L.; Graedel, T. E. (1986) Speciation, photosensitivity, and reactions of transition
metal ions in atmospheric droplets. J. Geophys. Res. [Atmos.] 91: 5189-5204.
Wesely, M. L. (1988) Improved parameterizations for surface resistance to gaseous dry deposition in
regional-scale, numerical models. Research Triangle Park, NC: U.S. Environmental Protection
Agency, Atmospheric Sciences Research Laboratory; report no. EPA/600/3-88/025. Available from:
NTIS, Springfield, VA; PB88-225099/AS.
Westberg, H.; Lamb, B. (1985) Ozone production and transport in the Atlanta, Georgia, region. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Sciences Research
Laboratory; report no. EPA-600/3-85-013. Available from: NTIS, Springfield, VA; PB85-173839.
Westberg, H. H.; Rasmussen, R. A.; Holdren, M. (1974) Gas chromatographic analysis of ambient air for light
hydrocarbons using a chemically bonded stationary phase. Anal. Chem. 46: 1852-1855.
Westberg, H.; Allwine, K. J.; Elias, D. (1976) Vertical ozone distribution above several urban and adjacent rural
areas across the United States. In: Specialty conference on: ozone/oxidants—interactions with the
total environment; March; Dallas, TX. Pittsburgh, PA: Air Pollution Control Association; pp. 84-95.
Westberg, H. H.; Holdren, M. W.; Hill, H. H., Jr. (1982) Analytical methodology for the identification and
quantification of vapor phase organic pollutants. Atlanta, GA: Coordinating Research Council, Inc.;
report no. CRC/APRAC/CAPA-11-71. Available from: NTIS, Springfield, VA; PB82-255753.
Westberg, H.; Lonneman, W.; Holdren, M. (1984) Analysis of individual hydrocarbon species in ambient
atmospheres: techniques and data validity. In: Keith, L. H., ed. Identification and analysis of organic
3-221
-------
pollutants in air: [papers from a symposium at the 184th National American Chemical Society
meeting]; September 1982; Kansas City, MO. Woburn, MA: Butterworth Publishers; pp. 323-337.
Whitby, K. T.; Husar, R. B.; Liu, B. Y. H. (1972) The aerosol size distribution of Los Angeles smog. J. Colloid
Interface Sci. 39: 177-204.
Williams, E. J.; Guenther, A.; Fehsenfeld, F. C. (1992) An inventory of nitric oxide emissions from soils in the
United States. J. Geophys. Res. [Atmos.] 97: 7511-7519.
Winer, A. M. (1983) Investigation of the role of natural hydrocarbons in photochemical smog formation
in California: final report. Sacramento, CA: California Air Resources Board; contract
no. AO-056-32.
Winer, A. M.; Peters, J. W.; Smith, J. P.; Pitts, J. N., Jr. (1974) Response of commercial chemiluminescent NO-
NO2 analyzers to other nitrogen-containing compounds. Environ. Sci. Technol. 8: 1118-1121.
Winer, A. M.; Atkinson, R.; Arey, J.; Biermann, H. W.; Harger, W. P.; Tuazon, E. C.; Zielinska, B. (1987)
The role of nitrogenous pollutants in the formation of atmospheric mutagens and acid deposition.
Sacramento, CA: California Air Resources Board; report no. ARB-R-87/308. Available from: NTIS,
Springfield, VA; PB87-222949.
Winer, A. M.; Arey, J.; Atkinson, R.; Aschmann, S. M.; Long, W. D.; Morrison, C. L.; Olszyk, D. M. (1992)
Emission rates of organics from vegetation in California's Central Valley. Atmos. Environ. Part A
26: 2647-2659.
Wolff, G. T. (1993) On a NOx-focused control strategy to reduce O3. J. Air Waste Manage. Assoc.
43: 1593-1596.
Wolff, G. T.; Korsog, P. E. (1992) Ozone control strategies based on the ratio of volatile organic compounds
to nitrogen oxides. J. Air Waste Manage. Assoc. 42: 1173-1177.
Wolff, G. T.; Lioy, P. J. (1978) An empirical model for forecasting maximum daily ozone levels in the
northeastern U.S. J. Air Pollut. Control Assoc. 28: 1034-1038.
Wolff, G. T.; Lioy, P. J.; Meyers, R. E.; Cederwall, R. T.; Wight, G. D.; Pasceri, R. E.; Taylor, R. S. (1977a)
Anatomy of two ozone transport episodes in the Washington, D.C., to Boston, Mass., corridor.
Environ. Sci. Technol. 11: 506-510.
Wolff, G. T.; Lioy, P. J.; Wight, G. D.; Pasceri, R. E. (1977b) Aerial investigation of the ozone plume
phenomenon. J. Air Pollut. Control Assoc. 27: 460-463.
Wolff, G. T.; Kelly, N. A.; Ferman, M. A. (1982) Source regions of summertime ozone and haze episodes in the
eastern United States. Water Air Soil Pollut. 18: 65-81.
Wolff, G. T.; Ruthkosky, M. S.; Stroup, D. P.; Korsog, P. E. (1991) A characterization of the principal PM-10
species in Claremont (summer) and Long Beach (fall) during SCAQS. Atmos. Environ.
Part A 25: 2173-2186.
World Meteorological Organization. (1990a) Scientific assessment of stratospheric ozone: 1989. Volume 1.
Geneva, Switzerland: Global Ozone Research and Monitoring Project; pp. 331-332; report no. 20.
World Meteorological Organization. (1990b) Scientific assessment of stratospheric ozone: 1989. Volume II.
Appendices: AFEAS report. Geneva, Switzerland: Global Ozone Research Monitoring Project;
report no. 20.
3-222
-------
World Meteorological Organization. (1992) Scientific assessment of ozone depletion: 1991 [preprint]. Geneva,
Switzerland: World Meteorological Organization; report no. 25.
Wu, Y.-L.; Davidson, C. I.; Dolske, D. A.; Sherwood, S. I. (1992) Dry deposition of atmospheric contaminants:
the relative importance of aerodynamic, boundary layer, and surface resistances. Aerosol Sci.
Technol. 16: 65-81.
Wunderli, S.; Gehrig, R. (1991) Influence of temperature on formation and stability of surface PAN and ozone. A
two year field study in Switzerland. Atmos. Environ. Part A 25: 1599-1608.
Yamada, T.; Kao, C. J.; Bunker, S. (1989) Airflow and air quality simulations over the western mountainous
region with a four-dimensional data assimilation technique. Atmos. Environ. 23: 539-554.
Yanagisawa, Y.; Nishimura, H. (1982) A badge-type personal sampler for measurement of personal exposure
to NO2 and NO in ambient air. Environ. Int. 8: 235-242.
Yocke, M. A (1981) A three-dimensional wind model for complex terrain [Ph.D. dissertation]. Berkeley,
CA: University of California.
Yokouchi, Y.; Mukai, H.; Nakajima, K.; Ambe, Y. (1990) Semi-volatile aldehydes as predominant organic gases
in remote areas. Atmos. Environ. Part A 24: 439-442.
Yoshizumi, K.; Aoki, K.; Nouchi, I.; Okita, T.; Kobayashi, T.; Kamakura, S.; Tajima, M. (1984) Measurements
of the concentration in rainwater and of the Henry's Law constant of hydrogen peroxide. Atmos.
Environ. 18: 395-401.
Young, T. R.; Boris, J. P. (1977) A numerical technique for solving stiff ordinary differential equations associated
with the chemical kinetics of reactive-flow problems. J. Phys. Chem. 81: 2424-2427.
Zafiriou, O. C.; True, M. B. (1986) Interferences in environmental analysis of NO by NO plus O, detectors:
a rapid screening technique. Environ. Sci. Technol. 20: 594-596.
Zaitsu, K.; Ohkura, Y. (1980) New fluorogenic substrates for horseradish peroxidase: rapid and sensitive assays
for hydrogen peroxide and the peroxidase. Anal. Biochem. 109: 109-113.
Zalewsky, E.; Sistla, G.; Henry, R.; Rao, S. T. (1993) Trends in ozone and its precursors in the northeastern
United States. In: Vostal, J. J., ed. Tropospheric ozone: nonattainment and design value issues
[proceedings of a U.S. EPA/A&WMA international specialty conference]; October 1992; Boston,
MA. Pittsburgh, PA: Air & Waste Management Association; pp. 459-476. (A&WMA transactions
series no. 23).
Zeldin, M. D. (1993) Meteorological and air quality characterization of the 1987 Southern California Air Quality
Study. In: Southern California Air Quality Study data analysis: proceedings of an international
specialty conference; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management Association; pp.
3-8. (A&WMA publication VIP-26).
Zhang, D.-L.; Chang, H.-R.; Seaman, N. L.; Warner, T. T.; Fritsch, J. M. (1986) A two-way interactive nesting
procedure with variable terrain resolution. Mon. Weather Rev. 114: 1330-1339.
Zhang, S.-H.; Shaw, M.; Seinfeld, J. H.; Flagan, R. C. (1992) Photochemical aerosol formation fromD-pinene
andD-pinene. J. Geophys. Res. [Atmos.] 97: 20,717-20,729.
Zhou, X.; Mopper, K. (1990) Apparent partition coefficients of 15 carbonyl compounds between air and seawater
and between air and freshwater; implications for air-sea exchange. Environ. Sci. Technol. 24: 1864-
1869.
3-223
-------
Zika, R. G.; Saltzman, E. S. (1982) Interaction of ozone and hydrogen peroxide in water: implications for
analysis of H2O2 in air. Geophys. Res. Lett. 9: 231-234.
Zimmerman, P. R. (1979) Testing of hydrocarbon emissions from vegetation, leaf litter and aquatic surfaces,
and development of a methodology for compiling biogenic emission inventories: final report.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards; report no. EPA-450/4-79-004. Available from: NTIS, Springfield, VA; PB-296070.
Zuo, Y. ; Hoigne, J. (1993) Evidence for photochemical formation of HQ and oxidation of SO, in authentic fog
water. Science (Washington, DC) 260: 71-73.
Zweidinger, R. B.; Sigsby, J. E., Jr.; Tejada, S. B.; Stump, F. D.; Dropkin, D. L.; Ray, W. D.; Duncan, J. W.
(1988) Detailed hydrocarbon and aldehyde mobile source emissions from roadway studies. Environ.
Sci. Technol. 22: 956-962.
3-224
-------
4
Environmental Concentrations,
Patterns, and Exposure Estimates
4.1 Introduction
The effects of ozone (O3) on humans, animals, and vegetation have received
extensive examination and are discussed elsewhere in this document. As indicated in the
previous O3 criteria document (U.S. Environmental Protection Agency, 1986), most of the
human and welfare effects research has focused on evaluating those impacts on health or
vegetation of exposure to O3 that simulate ambient O3 exposures (e.g., matching the
occurrence of hourly average concentrations or more prolonged times of exposure). This
information on concentrations obtained from extensive monitoring in the United States can be
useful both for linking anthropogenic emissions of O3 precursors with the protection of health
and welfare (i.e., determining compliance with air standards) and for augmenting exposure
assessment and epidemiology studies. The major emphasis in this chapter, however, will be
on characterizing and summarizing the extensive O3-monitoring data collected under ambient
conditions. Although most of the O3 air quality data summarized were gathered for
compliance and enforcement purposes, the hourly averaged O3 information can be used for
determining patterns and trends and as inputs to exposure and health assessments (e.g., U.S.
Environmental Protection Agency, 1992a; Lefohn et al., 1990a). In the sections that follow,
the hourly averaged ambient O3 data have been summarized in different ways to reflect the
interests of those who wish to know more about the potential for O3 to affect humans and the
environment. This chapter is not an exposure assessment for ambient O3; rather, this chapter
elucidates the features of O3 concentration patterns and exposure possibilities.
Trend patterns for O3 over several periods of time are described in Section 4.2.
The trends for O3 have been summarized by the U.S. Environmental Protection Agency
(1994) for 1983 to 1993. In addition, trends analysis for specific regions of the United States
have been performed by several investigators. In some cases, attempts have been made to
adjust for meteorological variation. In Section 4.3, the hourly averaged concentration
information from several monitoring networks has been characterized for urban and rural
areas. The diurnal variation (Section 4.4) occurring at urban and rural locations, as well as
seasonal patterns, also are described. Specific focus is provided on O3 monitoring sites that
experience low maximum hourly average concentrations because these locations form the
"basis for comparison" for O3 concentrations and exposures. In Section 4.5, the seasonal
patterns of hourly average concentrations are discussed. The hourly average concentration
information is used in Section 4.6 to compare the spatial variations that occur in urban areas
with those in nonurban areas, as well as with those in high-elevation locations. For
4-1
-------
comparing indoor to outdoor O3 exposures or concentrations, information is provided in
Section 4.7 on the latest data on indoor/outdoor (I/O) ratios. Section 4.8 describes efforts to
estimate both human and vegetation exposure to O3. Examples are provided on how both
fixed-site monitoring information and human exposure models are used to estimate risks
associated with O3 exposure. A short discussion is provided on the importance of hourly
average concentrations, which are used in human health and vegetation experiments that
simulate "real world" exposures.
As indicated in the previous O3 criteria document (U.S. Environmental Protection
Agency, 1986), O3 is the only photochemical oxidant other than nitrogen dioxide (NO2) that is
routinely monitored and for which a comprehensive aerometric database exists. Data for
peroxyacetyl nitrate (PAN) and hydrogen peroxide (H2O2) have been obtained only as part of
special research investigations. Consequently, no data on nationwide patterns of occurrence
are available for these non-O3 oxidants; nor are extensive data available on the correlations of
levels and patterns of these oxidants with those of O3. Sections 4.9 and 4.10 summarize the
available data for these other oxidants. Section 4.11 describes the co-occurrence patterns of
O3 with NO2; sulfur dioxide (SO2); and acidic aerosols, precipitation, and cloudwater.
4.1.1 Characterizing Ambient Ozone Concentrations
It is important to distinguish among concentration, exposure, and dose when using
air quality data to assess human health and vegetation effects. For this document, the
following definitions apply:
1. The "concentration" of a specific air pollutant is the amount of that material
per unit volume of air. Air pollution monitors measure pollutant
concentrations, which may or may not provide accurate exposure estimates.
2. The term "exposure" is defined as the concentration of a pollutant encountered
by the subject (animal, human, or plant) for a duration of time. Exposure
implies that such an encounter leads to intake (i.e., through the respiratory tract
or stomata).
3. The term "dose" is defined as that mass of pollutant delivered to an inner
target. This term has numerous quantitative descriptions (e.g., micrograms of
O3 per square centimeter of lung epithelium per minute), so the context of the
use of this term within the document must be considered. Human dosimetry is
discussed in Chapter 8.
The dose incurred by an organism (e.g., plant, animal, or human) is a more
complicated measure involving the concentration, the exposure duration, and the concurrent
state of the organism's susceptibility. These distinctions become important because the
concentration of an airborne contaminant that is measured in an empty room or at a stationary
outdoor monitor is not in fact an exposure. A measured concentration functions as an
alternative to an exposure only to the degree to which it represents concentrations actually
experienced by individuals.
Concentrations of airborne contaminants for vegetation are considered to represent
an exposure when a plant is subjected to them over a specified time period. As indicated in
Chapter 5 (see Section 5.5), dose has been defined historically by air pollution vegetation
researchers as ambient air quality concentration multiplied by time (O'Gara, 1922). However,
a more rigorous definition was required. Runeckles (1974) introduced the concept of
"effective dose" as the amount or concentration of pollutant that is adsorbed by vegetation, in
4-2
-------
contrast to that which is present in the ambient air. Fowler and Cape (1982) developed this
concept further and proposed that the "pollutant adsorbed dose" be defined in units of grams
per square meter (of ground or leaf area) and could be obtained as the product of
concentration, time, and stomatal (or canopy) conductance for the gas in question. Taylor
et al. (1982) suggested internal flux (milligrams per square meter per hour) as a measure of
the dose to which plants respond. In this chapter, dose will be taken to signify, for the
purposes of vegetation, that amount of pollutant absorbed by the plant.
In order to characterize the specific doses responsible for affecting human health
and vegetation, there has to be a linkage between exposure and actual dose. Unfortunately, it
is difficult to predict this relationship, even with the available models. For example, the
sensitivity of vegetation to O3 as a function of time of day, period of growth, or edaphic
conditions can determine the severity of response. For example, high O3 concentrations may
cause minimum injury or damage to plants, whereas more moderate O3 concentrations may
cause a greater degree of injury or damage (Showman, 1991). Because not enough is known
to quantify the links between exposure and dosage, and routine monitoring for O3 is
summarized as hourly average concentrations (i.e., potential exposure), most of the
information provided in this chapter is characterized in terms of concentration and exposure.
As indicated in Chapter 5, for many years, air pollution specialists have explored
alternative mathematical approaches for summarizing ambient air quality information in
biologically meaningful forms that can serve as alternatives for characterizing dose.
For vegetation, as indicated in Chapter 5 (Section 5.5), extensive research has
focused on identifying indicators of concentration and duration (exposure) that are firmly
founded on biological principles. Many of these indicators have been based on research
results indicating that the magnitude of vegetation responses to air pollution is determined
more as a function of the magnitude of the concentration than of the length of the exposure
(U.S. Environmental Protection Agency, 1992b). Short-term (1- to 8-h), high
O3 concentrations (>0.1 ppm) have been identified by many researchers as being more
important than long-term, low O3 concentrations for induction of visible injury to vegetation
(see Chapter 5 for further discussion).
Long-term, average concentrations were used initially as an exposure indicator to
describe O3 concentrations over time when assessing vegetation effects (Heck et al., 1982).
Based on the view presented in the previous criteria document (U.S. Environmental Protection
Agency, 1986) that higher concentrations of O3 should be given more weight than lower
concentrations (see Section 5.5 for further details), the following specific concerns about the
use of a long-term average to summarize exposures of O3 began appearing in the literature:
the use of a long-term average failed to consider the impact of peak concentrations and of
duration; a large number of hourly data sets within the commonly used 7-h window (0900 to
1559 hours), although diversely distributed and implying potentially diverse exposure
potentials, were characterized by the same 7-h seasonal mean; and high hourly average
concentrations (e.g., values greater than 0.1 ppm) occurred outside of a fixed 7-h window.
In summarizing the hourly average concentrations in this chapter, specific attention
is given to the relevance of the exposure indicators used. For example, for human health
considerations, concentration (or exposure) indicators such as the daily maximum 1-h average
concentrations, as well as the number of daily maximum 4-h or 8-h average concentrations,
are used to characterize information in the population-oriented locations. For vegetation,
several different types of exposure indicators are used. For example, much of the National
Crop Loss Assessment Network (NCLAN) exposure information is summarized in terms of
4-3
-------
the 7-h average concentrations. However, because peak-weighted, cumulative indicators (i.e.,
exposure parameters that sum the products of hourly average concentrations multiplied by
time over an exposure period) have shown considerable promise in relating exposure and
vegetation response (see Section 5.5), several exposure indicators that use either a threshold
or a sigmoidal weighting scheme are discussed in this chapter to provide insight concerning
the O3 exposures that are experienced at a select number of rural monitoring sites in the
United States. The peak-weighted, cumulative exposure indicators used in this chapter are
SUM06 and SUM08 (the sums of all hourly average concentrations equal to or greater than
0.06 and 0.08 ppm, respectively) and W126 (the sum of the hourly average concentrations
that have been weighted according to a sigmoid function [see Lefohn and Runeckles, 1987]
that theoretically is based on a hypothetical vegetation response).
The exposure indicators used for human health considerations are in concentration
units (i.e., parts per million), whereas the indicators used for vegetation are in both parts per
million (e.g., 7-h seasonal average concentrations) and parts per million per hour (e.g.,
SUM06, SUM08, W126). The magnitude of the peak-weighted, cumulative indicators at
specific sites can be compared with those values experienced at areas that experience low
hourly average maximum concentrations. In some cases, to provide more detailed
information about the distribution patterns for a specific O3 exposure regime, the percentile
distribution of the hourly average concentrations (in parts per million) is given. For further
clarification of the determination and rationale for the exposure indicators that are used for
assessing human health and vegetation effects, the reader is encouraged to read Chapters 5
(Section 5.5) and 7.
4.1.2 The Identification and Use of Existing Ambient Ozone Data
Information is readily available from the database supported by a network of
monitoring stations that were established to determine compliance with the National Ambient
Air Quality Standards (NAAQS) for O3. Most of the data presented in this chapter were
obtained from data stored in the U.S. Environmental Protection Agency's (EPA's)
computerized Aerometric Information Retrieval System (AIRS) and were collected after 1978.
As pointed out in the previous criteria document for O3 and other photochemical oxidants
(U.S. Environmental Protection Agency, 1986), there was some difficulty in interpreting the
O3 data obtained at most sites across the United States prior to 1979 because of calibration
problems.
In the United States, O3 hourly average concentrations are monitored routinely
through the National Air Monitoring Network, consisting of three types of sites. The
National Air Monitoring Station (NAMS) sites are located in areas where the concentrations
of O3 and subsequent potential human exposures are expected to be high. Criteria for these
sites have been established by regulation to meet uniform standards of siting, quality
assurance, equivalent analytical methodology, sampling intervals, and instrument selection to
assure consistency among the reporting agencies. For O3, NAMS sites are located only in
urban areas with populations exceeding 200,000. The other two types of sites are State and
Local Air Monitoring Stations and Special Purpose Monitors, which meet the same rigid
criteria for the NAMS sites but may be located in areas that do not necessarily experience
high concentrations in populated areas.
For O3, the reporting interval is 1 h, with the instruments operating continuously
and producing an integrated hourly average measurement. In many cases, EPA summarizes
4-4
-------
air quality data by an O3 "season". Table 4-1 summarizes the O3 season for the District of
Columbia and each of the states in the United States.
Table 4-1. Ozone Monitoring Season by State
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
D.C.
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
aAir Quality Control
bAQCR Numbers 1,
Begin
March
April
January
March
January
March
April
April
April
January
March
January
April
April
April
April
April
April
January
April
April
April
April
April
March
April
Region (AQCR)
2, 3, 6, 8, 9, and
End
November
October
December
November
December
September
October
October
October
December
November
December
October
October
October
October
October
October
December
October
October
October
October
October
November
October
State
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas3
Texasb
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Numbers 4, 5, 7, 10, and 11.
12.
Begin
June
April
January
April
April
January
April
April
May
April
March
April
April
April
April
June
April
January
March
May
April
April
April
April
April
April
End
September
October
December
October
October
December
October
October
September
October
November
October
October
October
October
September
October
December
October
September
October
October
October
October
October
October
Source: Code of Federal Regulations (1991).
4-5
-------
In this chapter, data are analyzed for the purpose of providing focus on specific
issues of exposure-response relationships that are considered in the later effects chapters. The
analyses proceed from a national picture of peak annual averages in Metropolitan Statistical
Areas (MSAs), through national 10- and 3-year trends, to characteristic seasonal and diurnal
patterns at selected stations, and then a brief examination of the incidence of episodic 1-h
levels. Although there are O3 data collected from monitoring stations not listed in AIRS, the
major source of information was derived from ambient air concentrations from monitoring
sites operated by the State and local air pollution agencies who report their data to AIRS.
Because meteorology affects the identification of trends, methodologies that adjust for
meteorology are described below.
To obtain a better understanding of the potential effect of ambient
O3 concentrations on human health and vegetation, hourly average concentration information
was summarized for urban versus rural (forested and agricultural) areas in the United States.
A land use characterization of "rural" does not imply that any specific location is isolated
from anthropogenic influences. For example, Logan (1989) has noted that hourly average
O3 concentrations above 0.08 ppm are common in rural areas of the eastern United States in
spring and summer, but are unusual in remote western sites. Consequently, for the purposes
of comparing exposure regimes that may be characteristic of clean locations in the United
States with those that are urban influenced (i.e., located in either urban or rural locations), this
chapter characterizes data collected from those stations whose locations appear to be isolated
from large-scale anthropogenic influences.
Long-term (multiyear) patterns and trends are available only from stationary
ambient monitors; data on indoor concentrations are collected predominantly in selected
settings during comparatively short-term studies. Data from the indoor and outdoor
environments are reviewed here separately.
4.2 Trends in Ambient Ozone Concentrations
Ozone concentrations and, thus, exposure change from year to year. High
O3 levels occurred in 1983 and 1988 in some areas of the United States. These levels more
than likely were attributable, in part, to hot, dry, stagnant conditions. However, O3 levels in
1992 were the lowest of the 1983 to 1992 period (U.S. Environmental Protection Agency,
1993). These low levels may have been due to meteorological conditions that were less
favorable for O3 formation and to recently implemented control measures. Nationally, the
summer of 1992 was the third coolest summer on record (U.S. Environmental Protection
Agency, 1993). The U.S. Environmental Protection Agency (1993) has reported a 21%
improvement in O3 levels between 1983 and 1992, which, in part, may be attributed to
relatively high O3 levels in 1983, compared to the low O3 exposure years from the period
1989 through 1992. However, new statistical techniques accounting for meteorological
influences have been used by EPA and they appear to suggest an improvement (independent
of meteorological considerations) of 10% for the 10-year period, 1983 to 1992 (U.S.
Environmental Protection Agency, 1993).
The EPA summarizes trends for the NAAQS for the most current 3- and 10-year
periods. In order to be included in the 10-year trend analysis in the annual National Air
Quality and Emissions Trend Report (U.S. Environmental Protection Agency, 1993), a
4-6
-------
station must report valid data for at least 8 of the last 10 years. A companion analysis of the
most recent 3 years requires valid data in all 3 years. Analysis in the above report covers the
periods 1983 to 1992 and 1990 to 1992, respectively; 509 sites met the 10-year period
criteria, and 672 sites are included in the 1990 to 1992 database. The NAMS sites comprise
196 of the long-term trends sites and 222 of the sites in the 3-year database.
Figure 4-1 displays the 10-year composite average trend for the second highest
daily maximum hourly average concentration during the O3 season for the 509 trend sites and
the subset of 196 NAMS sites. The 1992 composite average for the 509 trend sites is 21%
lower than the 1983 average and 20% lower for the subset of 196 NAMS sites. The 1992
value is the lowest composite average of the past 10 years (U.S. Environmental Protection
Agency, 1993). The 1992 composite average is significantly less than all the previous nine
years, 1983 to 1991. As discussed in U.S. Environmental Protection Agency (1992a), the
relatively high O3 concentrations in 1983 and 1988 likely were attributable in part to hot, dry
stagnant conditions in some areas of the country that were especially conducive to
O3 formation.
a.
a.
8
o
O
0.18-
0.16-
0.10-
0.08-
0.06-
0.04-
0.02-
0.00
NAAQS
^
=*^
All Sites (509)
NAMS Sites (196)
I I I I I I I I I I
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
Figure 4-1. National trend in the composite average of the second highest maximum
1-h ozone concentration at both National Air Monitoring Stations (NAMS)
and all sites with 95% confidence intervals, 1983 to 1992.
Source: U.S. Environmental Protection Agency (1993).
From 1991 to 1992, the composite mean of the second highest daily maximum 1-h
O3 concentrations decreased 7% at the 672 sites and 6% at the subset of 222 NAMS sites.
4-7
-------
Also, from 1991 to 1992, the composite average of the number of estimated instances of
O3 exceeding the standard decreased by 23% at the 672 sites, and by 19% at the 222 NAMS
sites. Nationwide volatile organic compound (VOC) emissions decreased 3% from 1991 to
1992 (U.S. Environmental Protection Agency, 1993).
The composite average of the second daily maximum concentrations decreased in
8 of the 10 EPA regions from 1991 to 1992, and remained unchanged in Region VII. Except
for Region VII, the 1992 regional composite means are lower than the corresponding 1990
levels. Although meteorological conditions in the east during 1993 were more conducive to
O3 formation than those in 1992, the composite mean level for 1993 was the second lowest
composite average for the decade (1984 to 1993) (U.S. Environmental Protection Agency,
1994).
Investigators have explored methods for investigating techniques for adjusting
O3 trends for meteorological influences (Stoeckenius and Hudischewskyj, 1990; Wakim, 1990;
Shively, 1991; Korsog and Wolff, 1991; Lloyd et al., 1989; Davidson, 1993; Cox and Chu,
1993). Stoeckenius and Hudischewskyj (1990) used a classification method to group days
into categories according to the magnitude of O3 and the similarity of meteorological
conditions within each defined group. Adjusted O3 statistics for each year were computed
from the meteorologically grouped data, and the yearly frequency of occurrence of each group
relative to its long-term frequency was described. Wakim (1990) used standard regression
analysis to quantify the effect of daily meteorology on O3. Adjusted O3 statistics were
calculated by adding the expected O3 statistic for a year with typical meteorology to the
average of the regression residuals obtained for the adjusted year. Shively (1991) described a
model in which the frequency of exceedance of various O3 thresholds was modeled as a
nonhomogeneous Poisson process where the parameter is a function of time and
meteorological variables. Kolaz and Swinford (1990) categorized O3 days as "conducive" or
"nonconducive", based on selected meteorological conditions within the Chicago, IL, area.
Within these categories, the meteorological intensity of days conducive to daily exceedances
of the NAAQS for O3 was calculated and used to establish long-term trends in the annual
exceedance rate.
Cox and Chu (1993) modeled the daily maximum O3 concentration using a
Weibull distribution with fixed-shape and scale parameters, the logarithm of which varies as a
linear function of several meteorological variables and a yearly index. The authors tested for
a statistically significant trend term to determine if an underlying meteorologically adjusted
trend could be detected. Overall, the measured and modeled predicted percentiles tracked
closely in the northern latitudes but performed less adequately in southern coastal and desert
areas. The results suggested that meteorologically adjusted upper percentiles of the
distribution of daily maximum 1-h O3 are decreasing in most urban areas over the period
1981 through 1991. The median rate of change was -1.1% per year, indicating that O3 levels
have decreased approximately 11% over this time period. The authors reported that trends
estimated by ignoring the meteorological component appear to underestimate the rate of
improvement in O3 primarily because of the uneven year-to-year distribution of
meteorological conditions favorable to O3 formation.
Lefohn et al. (1993a) focused on a potentially useful method for identifying
monitoring sites whose improvement in the level of O3 concentrations may be attributed more
to the implementation of abatement control strategies than meteorological changes. As has
been pointed out previously, meteorology plays an important role in affecting the
O3 concentrations that are contained in the tail of the 1-h distributions, as indicated by the
4-8
-------
successful predictive application of the exponential-tail model to distributions (California Air
Resources Board, 1992). Because meteorology plays such an important role in affecting the
tail of the 1-h distribution at a specific site, changes in "attainment" status are not expected to
affect changes in the entire distribution pattern and, thus, the average diurnal pattern. Lefohn
et al. (1993b) investigated the change in the annual average diurnal pattern as changes in
O3 levels occurred. The authors reported that, although the amplitude of the diurnal patterns
changed, there was little evidence for consistent changes in the shape of the annual diurnal
patterns (Figure 4-2). In a follow-up to this analysis, Lefohn et al. (1993a) reported that
25 of the 36 sites that changed compliance status across years showed no statistically
significant change in the shape of the average diurnal profile (averaged by O3 season).
In addition, the authors reported that for 71% (10 of 14) of the sites in Southern California
and Dallas-Fort Worth, TX, that showed improvement in O3 levels (i.e., reductions in the
number of exceedances over the years), but still remained in "nonattainment," a statistically
significant change in the shape of the seasonally averaged diurnal profile occurred (e.g.,
Figure 4-3). Thus, the authors noted that, for the Southern California and Dallas-Fort Worth
sites, changes were observed in the seasonally averaged diurnal profiles, whereas for the sites
moving between attainment and nonattainment status, such a change in shape generally was
not observed. Lefohn et al. (1993a) pointed out that it was possible that meteorology played
a more important role in affecting attainment status than did changes in emission levels.
Historically, the long-term O3 trends in the United States characterized by EPA
have emphasized air quality statistics that are closely related to the NAAQS. A report by the
National Academy of Sciences (NAS) (National Research Council, 1991) stated that the
principal measure currently used to assess O3 trends is highly sensitive to meteorological
fluctuations and is not a reliable measure of progress in reducing O3 over several years for a
given area. The NAS report recommended that "more statistically robust methods be
developed to assist in tracking progress in reducing ozone." The NAS report also points out
that most of the trends analyses are developed from violations of standards based on lower
concentration cutoffs or using percentile distributions. Because of the interest by EPA in
tracking trends in the quality of the air that people breathe when outdoors, most of the above
measures have some association with the existing NAAQS, in the form of either threshold
violations or O3 concentrations.
Several of the alternative examples provided in the NAS report were described
previously by Curran and Frank (1991). Several of the examples mentioned in the NAS
report involved threshold violations: the number of days on which the maximum
O3 concentration was above 0.12 ppm (Jones et al., 1989; Kolaz and Swinford, 1990; Wakim,
1990); the number of times during the year that the daily summary statistics exceeded
0.080 or 0.105 ppm (Stoeckenius, 1991), or the number of days in California when the
O3 concentration exceeded 0.2 ppm (Zeldin et al., 1991). Several other O3 concentration
measures are described in this report.
As an alternative to the way in which EPA historically has implemented its trends
analysis, U.S. Environmental Protection Agency (1992a) used percentiles in the range of the
50th percentile (or median) to the 95th percentile. The U.S. Environmental Protection
Agency (1992a) reported that the pattern for the 10-year trends (1982 to 1991), using the
various alternative O3 summary statistics, were somewhat similar. There was a tendency for
the curves to become flatter in the lower percentiles. The peak years of 1983 and 1988 were
still evident in the trend lines for each indicator. The increase of 8% recorded in the annual
4-9
-------
(a)
Montgomery Co., AL
Annual
Concord, CA
Annual
1 3 5 7 9 11 13 15 17 19 21 23
1 3 5 7 9 11 13 15 17 19 21 23
-1987
-1988 o 1989
-1990
-1987
1988 o 1989
-1990
Louisville, KY
Annual
0.08 •
f 0.06 -
&
1 0.04 -
0.02 -
(d) Dade Co., FL
Annual
1 3 5 7 9 11 13 15 17 19 21 23
1 3 5 7 9 11 13 15 17 19 21 23
-1987-»-1988 o 1989
-1990
-1987-»-1988 o 1989
-1990
Figure 4-2. The annually averaged composite diurnal curves for the following sites that
changed from nonattainment to attainment status: (a) Montgomery County,
AL; (b) Concord, CA; (c) Louisville, KY; and (d) Dade County, FL; for the
period 1987 to 1990. The darkened curve in each figure identified the year in
which the greatest number of daily maximum 4-h average concentrations
>0.08 ppm occurred.
Source: Lefohn et al. (1993b).
second-highest daily maximum 1-h concentration from 1987 to 1988 also was seen in the
95th and 90th percentile concentrations. The lower percentile indicators had smaller increases
of 3 to 4%. The percent change between 1982 and 1991 for each of the summary statistics
follows: annual daily maximum 1-h concentration, -11%; annual second daily maximum 1-h
concentration, -8%; 95th percentile of the daily maximum 1-h concentrations, -5%; 90th
percentile, -4%; 70th percentile, -1%; 50th percentile, or median of the daily maximum 1-h
concentrations, +1%; and the annual mean of the daily maximum 1-h concentrations, -1%.
Besides EPA, additional investigators have assessed trends at several locations in
the United States (e.g., Kuntasai and Chang, 1987; Gallopoulos et al., 1988; Korsog and
Wolff, 1991; Lloyd et al., 1989; Rao et al., 1992; Davidson, 1993). For example, Kuntasai
and Chang (1987) performed a basin-wide air quality trend analysis for the South Coast Air
4-10
-------
0.08
7 9 11 13 15 17 19 21 23
120
£ 0
Hour
1980 1982 1984 1986 1988 1990
Year
1980 * 1981--°--- 1990™A- 1991
Figure 4-3. A summary of the (a) seasonal (January to December) averaged composite
ozone diurnal curve and (b) integrated exposure W126 index for the
Los Angeles, CA, site for the period 1980 to 1991.
Source: Lefohn et al. (1993a).
Basin of California using multistation composite daily maximum 1-h average ambient
concentrations for the third quarter from 1968 to 1985. Basin-wide ambient O3 concentrations
appeared to show downward trends for the period 1970 to 1985, but because of high
fluctuations, it was difficult to delineate trends for shorter periods. The meteorology-adjusted
O3 showed a more consistent downward trend than did unadjusted O3. Korsog and Wolff
(1991) examined trends from 1973 to 1983 at eight major population centers in the
northeastern United States, using a robust statistical method. The 75th percentile was used by
the authors in determining trends. The data were collected over a 3-mo (June through
August) period. The surface temperature and upper air temperature variables were found to
be the best predictors of O3 behavior. Two regression procedures were performed to remove
the variability of meteorological conditions conducive to high O3 (i.e., O3 concentrations
>0.08 ppm). The results of the analysis showed that there had been a decrease of a few ppb
on a yearly basis for the majority of the sites investigated by the authors.
Lloyd et al. (1989) investigated the improvement in O3 air quality from 1976 to
1987 in the South Coast Air Basin. The authors reported that when the trend in total
exceedance hours of a consistent set of basin air monitoring stations was considered, the
improvement over the period of investigation was substantial. The authors reported that the
number of station hours at or above the Stage I Episode Level (0.2 ppm, 1-h average) had
decreased by about two-thirds over the period 1976 to 1987. Davidson (1993) reported on the
number of days on which O3 concentrations at one or more stations in the South Coast Air
Basin exceeded the federal standard and the number of days reaching Stage I episode levels,
for the months of May through October in the years 1976 to 1991. The author reported that
the number of basin days exceeding the federal standard declined at an average annual rate of
4-11
-------
2.27 days/year over the period. In addition, the number of basin days with Stage I episodes
declined at an average annual rate of 4.70 days/year over the period 1976 to 1991. Rao et al.
(1992) demonstrated the use of some statistical methods for examining trends in ambient
O3 air quality downwind of major urban areas. The authors examined daily maximum 1-h
O3 concentrations measured over New Jersey, metropolitan New York City, and Connecticut
for the period 1980 to 1989. The analyses indicated that although there has been an
improvement in O3 air quality downwind of New York City, there has been little change in
O3 levels upwind of New York City during this 10-year period.
Lefohn and Runeckles (1987) proposed a sigmoidal weighting function that was
used in developing a cumulative integrated exposure index (W126):
(4-1)
,
[1 + M x exp("A x Ci)]
where: w; = weighting factor for concentration i,
M and A are positive arbitrary constants, and
q = concentration i.
Lefohn et al. (1988b) reported the use of the sigmoidally weighted index with constants,
M and A, 4,403 and 126 ppm"1, respectively. The authors referred to the index as W126.
The values were subjectively determined to develop a weighting function that (1) included
hourly average concentrations as low as 0.04 ppm, (2) had an inflection point near
0.065 ppm, and (3) had an equal weighting of one for hourly average concentrations at
approximately 0.10 ppm and above. To determine the value of the index, the sigmoidal
weighting function at c; was multiplied by the hourly average concentration, c;, and summed
over all relevant hours. The index included the lower, less biologically effective
concentrations in the integrated exposure summation. The weighting function has been used
to describe the relationship between O3 exposure and vegetation response (e.g., Lefohn et al.,
1988b, 1992a).
Lefohn and Shadwick (1991), using the W126 sigmoidally weighted exposure
index, assessed trends in O3 exposures at rural sites in the United States over 5- and 10-year
periods (1984 to 1988 and 1979 to 1988, respectively) for forestry and agricultural regions of
the United States. Although the statistical analysis did not explore the effects on trends of the
lower O3 exposure period 1989 to 1992, the analysis did reflect the effect of the higher
O3 exposure years (1983 and 1988). The hot, dry summer of 1988 was associated with the
highest O3 exposures in both the forest and agricultural regions of the eastern United States.
To compare the exposure index values across years, a correction for missing data was applied
for each pollutant. The corrections were determined for each site on a monthly basis. The
Kendall's K statistic (Mann-Kendall test) was used to identify linear trends. Estimates of the
rate of change (slope) for the index were calculated. Table 4-2 summarizes the results of the
analysis. For sites distributed by forestry regions, there were more positive than negative
slope estimates for the 5-year analysis of sites in the southern, midwestern, and Mid- Atlantic
regions. For the 10-year analysis, the above was true except for the Mid- Atlantic seasonal
analysis, where there was one positive and one negative significant
trend. In the southern region, 38% of the sites showed significant trends. For the sites in
Table 4-2. Summary by Forestry and Agricultural Regions for
4-12
-------
Ozone Trends Using the W126 Exposure Parameter Accumulated
on a Seasonal Basis3
Region
South
Midwest
West
Pacific
Northwest
Plains
Northeast
Mid-Atlantic
Rocky
Mountains
All
Region
Pacific
Mountain
Northern
Plains
Lake States
Corn Belt
Northeast
Appalachian
Southeast
Delta State
Southeastern
Plains
All
Not
53
38
10
4
3
14
12
5
139
Not
14
5
3
10
20
26
27
16
9
9
139
5-Year
Significant15
(16)
(1)
(0)
(2)
(0)
(0)
(0)
(2)
(21)
5 -Year
Significant15
(2)
(2)
(0)
(0)
(1)
(0)
(9)
(5)
(0)
(2)
(21)
Trends
-
0
0
0
0
0
1
0
0
1
Trends
-
0
0
0
0
0
1
0
0
0
0
1
Forestry
Significant
+
14
7
o
J
0
0
0
3
1
28
Agricultural
Significant
+
3
1
0
1
3
3
14
1
2
0
28
10-Year
Trends
Significant
Not Significant
13
20
4
2
2
7
4
2
54
10-Year
-
1
1
2
0
0
1
1
0
6
Trends
+
7
6
1
0
0
1
1
1
17
Significant
Not Significant
6
2
2
5
11
11
8
4
4
1
54
-
2
0
0
0
1
2
0
1
0
0
6
+
1
1
0
1
2
2
8
0
1
1
17
aSee Appendix A for abbreviations and acronyms.
bNumbers in parentheses in the "Not Significant" column under "5-Year Trends" are the number of sites with
exactly 3 years of data.
Source: Lefohn and Shadwick (1991).
4-13
-------
the northeastern region, few sites showed a significant trend. There were considerably fewer
sites in the remaining regions than in the four forestry regions above. Hence, for these
regions, no significance was assigned to the differences in the number of negative and
positive slope estimates in the tables. Similar to the results reported for the forestry regions,
most of the sites in the agricultural regions showed no O3 trends. However, in the
Appalachian agricultural region, as many as 50% of the sites showed a pronounced indication
of a trend. A predominance of positive significant trends for both the 5- and 10-year analyses
was observed. In the other agricultural regions, there were approximately an equal number of
positive and negative significant 5- and 10-year trends. The O3 results produced patterns that
were not pronounced enough to draw more than tentative conclusions for the 10-year analysis.
For the 5-year analysis, there was still not a strong indication of an O3 trend. However, when
significant trends were observed, they were almost always positive. This can be attributed to
eastern O3 levels that were generally higher in 1988 than in previous years.
4.3 Surface Ozone Concentrations
4.3.1 Introduction
Ozone is measured at levels above the minimum detectable level at all monitoring
locations in the world (Lefohn et al., 1990a). As dicussed earlier in Chapter 3, the concept of
a "natural" background of O3 is complex. Concentrations of background O3 can vary with
temperature, wind speed and direction, vertical motion, geographic location including latitude
and altitude, and season of the year. This background O3 can be attributed the following
sources: (1) downward transport of stratospheric O3 through the free troposphere to near
ground level; (2) in situ O3 production from methane emitted from swamps and wetlands
reacting with natural NOX emitted from soils, lighting strikes, and from downward transport of
NO from the stratosphere into the troposphere; and (3) in situ production of O3 from the
reactions of biogenic VOCs with natural NOX (National Research Council, 1991). A fourth
source to be considered is the O3 production resulting from long range transport of O3 from
distant pollutant sources (see Chapter 3).
The occasional occurrence of stratospheric injection of O3, at specific times and in
certain locations, is accepted and may be responsible for some of the rare occurrences of
elevated levels that have been observed at some high- and low-elevation remote sites.
A summer season average contribution of approximately 5 to 10 ppb for surface-level
O3 concentration from stratospheric intrusion has been estimated (Altshuller, 1989).
For purposes of comparing how O3 levels have changed over time, it would be
interesting to know how current levels compare to previous, historical natural background
levels. However, estimations of background O3 concentrations are difficult to make. The
definition of background and the use of O3 measurements are subject to much uncertainty.
It is difficult, if not impossible, to determine whether any geographic location on earth is free
from human influence (Finlayson-Pitts and Pitts, 1986). The natural precursor emissions can
be responsible for the production of the O3 concentrations observed at remote sites
(Chameides et al., 1988; Zimmerman, 1979; Trainer et al., 1987). Citing indirect evidence for
the possible importance of natural emissions, Lindsay et al. (1989) have emphasized that
additional research is required to assess the role that natural hydrocarbons might play in urban
and regional O3 episodes.
4-14
-------
It is possible for urban emissions, as well as O3 produced from urban area
emissions, to be transported to more rural downwind locations. This can result in elevated
O3 concentrations at considerable distances from urban centers (Wolff et al., 1977; Husar
et al., 1977; Wight et al., 1978; Vukovich et al., 1977; Wolff and Lioy, 1980; Pratt et al.
1983; Logan, 1985; Altshuller, 1986; U.S. Environmental Protection Agency, 1986; Kelly
et al., 1986; Pinkerton and Lefohn, 1986; Lefohn et al., 1987a; Logan, 1989; Lefohn and
Lucier, 1991; Taylor and Hanson, 1992). For example, on over 40% of the 98 days that the
maximum 1-h O3 concentrations exceeded 0.12 ppm, the highest value was measured
downwind of St. Louis at one of the rural sites, which was located approximately 50 km from
downtown St. Louis (Altshuller, 1986). Urban O3 concentration values often are depressed
because of titration by NOX (Stasiuk and Coffey, 1974). Reagan (1984) and Lefohn et al.
(1987a) have observed this phenomenon where O3 concentrations at center-city sites were
lower than some rural locations. Because of the absence of chemical scavenging, O3 tends to
persist longer in nonurban than in urban areas (U.S. Environmental Protection Agency, 1986;
Coffey et al., 1977; Wolff et al., 1977; Isaksen et al., 1978).
The distribution of O3 or its precursors at a rural site near an urban source is
affected by wind direction (i.e., whether the rural site is located up- or downwind from the
source) (Kelly et al., 1986; Lindsay and Chameides, 1988). Thus, it may be difficult to apply
land-use designations to the generalization of exposure regimes that may be experienced in
urban versus rural areas. Because of this, it is difficult to identify a set of unique
O3 distribution patterns that adequately describe the hourly average concentrations experienced
at monitoring sites in rural locations (Lefohn et al., 1991).
4.3.2 Urban Area Concentrations
Figure 4-4 shows the highest second daily maximum 1-h average O3 concentrations
in 1991 across the United States. The highest second daily maximum 1-h O3 concentrations
by MSA for the years 1989 to 1991 are summarized in Table 4-3. The highest
O3 concentrations are observed in Southern California, but high levels of O3 also occur in the
Texas Gulf Coast, the Northeast Corridor, and other heavily populated regions of the United
States, but with a much lower frequency.
Lefohn (1992a) reported that, for many urban sites that experience high second
daily maximum 1-h average values (i.e, >0.125 ppm), most are associated with only a few
episodes. Monitoring sites in polluted regions tend to experience frequent hourly average
O3 concentrations at or near minimum detectable levels. The percentile summary information
for some of these sites shows that, although some of the highest hourly average
concentrations occur at these locations, their occurrence is infrequent (Table 4-4). For
example, O3 monitoring sites at Delmar, CA; Stratford and Madison, CT; Baton Rouge, LA;
Bayonne, NJ; New York City and Babylon, NY; Harris County, TX; and Bayside, WI, exhibit
maximum hourly average concentrations above 0.125 ppm; however, only 1% of the hourly
average concentrations generally exceed 0.100 ppm. Although for human health
considerations, the occurrence of a second daily maximum hourly average concentration
>0.125 ppm is important, Table 4-4 illustrates that, for most of the sites listed (except for
several sites in California), such high hourly average concentrations occur less than 1% of the
time and are associated with occasional episodes.
As indicated in Section 4.1, interest has been expressed in characterizing
O3 exposure regimes for sites experiencing daily maximum 8-h concentrations above specific
4-15
-------
Figure 4-4. United States map of the highest second daily maximum 1-h average ozone
concentration by Metropolitan Statistical Area, 1991.
Source: U.S. Environmental Protection Agency (1992a).
thresholds (e.g., 0.08 or 0.10 ppm). Table 4-5 summarizes the highest second daily maximum
8-h average O3 concentrations by MSA for the years 1989 to 1991. The data have been
reported for the O3 season as summarized in Table 4-1. In some cases, high concentrations
occur in the fall and winter periods as well as in the summertime. Analyses documented the
occurrence, at some sites, of multihour periods within a day of O3 at levels of potential health
effects. Although most of these analyses were made using monitoring data collected from
sites in or near nonattainment areas, the analysis of Berglund et al. (1988) showed that at five
sites, two in New York state, two in rural California, and one in rural Oklahoma, an
alternative O3 standard of an 8-h average of 0.10 ppm would be exceeded even though the
existing 1-h standard would not be. Berglund et al. (1988) described the occurrence at these
five sites (none of which was in or near a nonattainment area) of O3 concentrations showing
only moderate peaks but exhibiting multihour levels above 0.10 ppm. Lefohn et al. (1993b)
identified those areas in the United States for the period 1987 to 1989 where more than one
occurrence of an 8-h daily maximum average concentration of 0.08 ppm was experienced, but
an hourly average concentration equal to or greater than 0.12 ppm never occurred.
A follow-up to the points made above is whether an improvement in O3 levels may
produce distributions of 1-h O3 that result in a broader diurnal profile than those seen in
4-16
-------
Table 4-3. The Highest Second Daily Maximum One-Hour Ozone
Concentration (ppm) by Metropolitan Statistical Area (MSA) for
the Years 1989 to 1991
MSA
1989 1990 1991
MSA
1989 1990 1991
Akron, OH 0.14 0.11 0.13
Albany-Schenectady-Troy, NY 0.10 0.11 0.10
Albuquerque, NM 0.10 0.10 0.09
Allentown-Bethlehem, PA-NJ 0.10 0.11 0.12
Altoona, PA 0.10 0.10 0.11
Anaheim-Santa Ana, CA 0.24 0.21 0.20
Anderson, IN 0.10
Anderson, SC 0.09
Ann Arbor, MI 0.10 0.09 0.11
Appleton-Oshkosh-Neenah, WI 0.10 0.08 0.09
Asheville, NC 0.08 0.09 0.08
Atlanta, GA 0.12 0.15 0.13
Atlantic City, NJ 0.12 0.16 0.14
Augusta, GA-SC 0.10 0.11 0.10
Aurora-Elgin, IL 0.11 0.09 0.13
Austin, TX 0.11 0.11 0.10
Bakersfield, CA 0.16 0.16 0.16
Baltimore, MD 0.13 0.14 0.16
Baton Rouge, LA 0.16 0.18 0.14
Beaumont-Port Arthur, TX 0.15 0.15 0.13
Beaver County, PA 0.10 0.10 0.11
Bellingham, WA 0.05 0.08 0.07
Benton Harbor, MI 0.12
Bergen-Passaic, NJ 0.12 0.13 0.14
Billings, MT 0.08
Birmingham, AL 0.12 0.13 0.11
Boston, MA 0.12 0.11 0.13
Boulder-Longmont, CO 0.11 0.10 0.10
Bradenton, FL 0.10 0.10 0.10
Brazoria, TX 0.15 0.13
Bridgeport-Milford, CT 0.18 0.16 0.15
Brockton, MA 0.13 0.12 0.15
Buffalo, NY 0.11 0.11 0.11
Canton, OH 0.12 0.11 0.12
Cedar Rapids, IA 0.08 0.07 0.08
Champaign-Urbana-Rantoul, IL 0.09 0.09 0.08
Charleston, SC 0.09 0.10 0.09
Charleston, WV 0.10 0.12 0.12
Charlotte-Gastonia-Rock Hill, NC-SC 0.13 0.12 0.12
Chattanooga, TN-GA 0.11 0.12 0.10
Chicago, IL 0.12 0.11 0.13
Chico, CA 0.10 0.12 0.09
Cincinnati, OH-KY-IN 0.12 0.15 0.14
Cleveland, OH 0.12 0.12 0.13
Colorado Springs, CO 0.09 0.09 0.09
Columbia, SC 0.10 0.11 0.11
Columbus, GA-AL 0.09 0.11 0.10
Columbus, OH 0.11 0.11 0.12
Corpus Christi, TX 0.10 0.10 0.11
Cumberland, MD-WV 0.09 0.10
Dallas, TX 0.13 0.14 0.12
Danbury, CT 0.13 0.15 0.14
Davenport-Rock Island-Moline, IA-IL 0.11 0.10 0.10
Dayton-Springfield, OH 0.15 0.12 0.12
Decatur, IL 0.09 0.09 0.10
Denver, CO 0.11 0.11 0.11
Des Moines, IA 0.08 0.07 0.07
Detroit, MI 0.14 0.12 0.13
Duluth, MN-WI 0.06
Eau Claire, WI 0.06
El Paso, TX 0.14 0.14 0.13
Elmira, NY 0.09 0.10 0.10
Erie, PA 0.12 0.10 0.11
Eugene-Springfield, OR 0.08 0.09 0.09
Evansville, IN-KY 0.12 0.11 0.12
Fayetteville, NC 0.11 0.10 0.10
Flint, MI 0.10 0.10 0.10
Fort Collins, CO 0.09 0.10 0.09
Ft. Lauderdale-Hollywood-Pompano, FL 0.12 0.10 0.10
Fort Myers-Cape Coral, FL 0.10 0.08 0.08
Fort Wayne, IN 0.12 0.09 0.10
Fort Worth-Arlington, TX 0.13 0.14 0.15
Fresno, CA 0.15 0.15 0.16
Galveston-Texas City, TX 0.14 0.15 0.15
Gary-Hammond, IN 0.11 0.12 0.12
Grand Rapids, MI 0.13 0.14 0.15
Greeley, CO 0.10 0.11 0.10
Green Bay, WI 0.09 0.09 0.10
Greensboro-Winston Salem-High Point, NC 0.10 0.12 0.11
Greenville-Spartanburg, SC 0.10 0.11 0.11
Hamilton-Middletown, OH 0.11 0.13 0.12
Harrisburg-Lebanon-Carlisle, PA 0.11 0.12 0.11
Hartford, CT 0.14 0.15 0.15
Hickory, NC 0.09
Honolulu, HI 0.05 0.05 0.05
Houma-Thibodaux, LA 0.11 0.12 0.10
Houston, TX 0.23 0.22 0.20
Huntington-Ashland, WV-KY-OH 0.12 0.14 0.14
Huntsville, AL 0.09 0.09 0.11
Indianapolis, IN 0.12 0.11 0.11
Iowa City, IA 0.09 0.09 0.06
Jackson, MS 0.09 0.10 0.09
Jacksonville, FL 0.11 0.11 0.10
Jamestown-Dunkirk, NY 0.08 0.10
Janesville-Beloit, WI 0.12 0.09 0.11
Jersey City, NJ 0.12 0.18 0.14
Johnson City-Kingsport-Bristol, TN-WV 0.11 0.12 0.12
Johnstown, PA 0.10 0.10 0.11
Joliet, IL 0.10 0.09 0.12
Kalamazoo, MI 0.08
Kansas City, MO-KS 0.11 0.11 0.12
Kenosha, WI 0.13 0.11 0.15
Knoxville, TN 0.10 0.12 0.11
Lafayette, LA 0.10 0.11 0.08
Lafayette, IN 0.09 0.10
Lake Charles, LA 0.13 0.13 0.12
Lake County, IL 0.13 0.10 0.12
Lancaster, PA 0.10 0.10 0.12
4-17
-------
Table 4-3 (cont'd). The Highest Second Daily Maximum One-Hour
Ozone Concentration (ppm) by Metropolitan Statistical
Area (MSA) for the Years 1989 to 1991
MSA
1989 1990 1991
MSA
1989 1990 1991
Lansing-East Lansing, MI 0.10 0.10 0.11
Las Graces, NM 0.11 0.10 0.10
Las Vegas, NV 0.11 0.11 0.09
Lawrence-Haverhill, MA-NH 0.12 0.10 0.13
Lexington-Fayette, KY 0.11 0.11 0.10
Lima, OH 0.10 0.10 0.10
Lincoln, NE 0.06 0.07 0.07
Little Rock-North Little Rock, AR 0.09 0.10 0.10
Longview-Marshall, TX 0.10 0.13 0.11
Lorain-Elyria, OH 0.12 0.09 0.10
Los Angeles-Long Beach, CA 0.33 0.27 0.31
Louisville, KY-IN 0.11 0.13 0.13
Lynchburg, VA 0.10 0.09
Madison, WI 0.10 0.08 0.11
Manchester, NH 0.10 0.10 0.10
Medford, OR 0.09 0.10 0.07
Melbourne-Titusville-Palm Bay, FL 0.10 0.09 0.09
Memphis, TN-AR-MS 0.12 0.12 0.11
Miami-Hialeah, FL 0.12 0.11 0.12
Middlesex-Somerset-Hunterdon, NJ 0.13 0.15 0.13
Middletown, CT 0.17 0.16 0.17
Milwaukee, WI 0.15 0.13 0.18
Minneapolis-St. Paul, MN-WI 0.10 0.10 0.09
Mobile, AL 0.10 0.11 0.09
Modesto, CA 0.13 0.12 0.11
Monmouth-Ocean, NJ 0.14 0.14 0.15
Montgomery, AL 0.08 0.10 0.09
Muskegon, MI 0.14 0.13 0.15
Nashua, NH 0.09 0.10 0.11
Nashville, TN 0.14 0.13 0.12
Nassau-Suffolk, NY 0.15 0.14 0.18
New Bedford, MA 0.12 0.13 0.13
New Haven-Meriden, CT 0.15 0.16 0.18
New London-Norwich, CT-RI 0.14 0.16 0.14
New Orleans, LA 0.11 0.11 0.11
New York, NY 0.13 0.16 0.18
Newark, NJ 0.13 0.13 0.14
Niagara Falls, NY 0.10 0.10 0.10
Norfolk-Virginia Beach-Newport News, VA 0.10 0.11 0.11
Oakland, CA 0.13 0.12 0.12
Oklahoma City, OK 0.11 0.11 0.11
Omaha, NE-IA 0.10 0.08 0.08
Orlando, FL 0.11 0.12 0.10
Owensboro, KY 0.10 0.11 0.09
Oxnard-Ventura, CA 0.17 0.15 0.16
Parkerburg-Marietta, WV-OH 0.12 0.11 0.12
Pascogoula, MS 0.10 0.11 0.10
Pensacola, FL 0.09 0.12 0.11
Peoria, IL 0.11 0.09 0.10
Philadelphia, PA-NJ 0.16 0.14 0.16
Phoenix, AZ 0.11 0.14 0.12
Pittsburgh, PA 0.13 0.11 0.12
Pittsfield, MA 0.09 0.11 0.10
Portland, ME 0.13 0.13 0.14
Portland, OR-WA 0.09 0.15 0.11
Portsmouth-Dover-Rochester, NH-ME 0.11 0.10 0.13
Poughkeepsie, NY 0.08 0.12 0.13
Providence, RI 0.13 0.14 0.16
Provo-Orem, UT 0.11 0.09 0.08
Racine, WI 0.14 0.11 0.14
Raleigh-Durham, NC 0.11 0.12 0.11
Reading, PA 0.11 0.11 0.12
Redding, CA 0.09 0.09 0.08
Reno, NV 0.10 0.14 0.09
Richmond-Petersburg, VA 0.11 0.12 0.12
Riverside-San Bernardino, CA 0.28 0.30 0.25
Roanoke, VA 0.10 0.09 0.10
Rochester, NY 0.11 0.11 0.11
Rockford, IL 0.10 0.09 0.09
Sacramento, CA 0.14 0.16 0.16
St. Louis, MO-IL 0.13 0.13 0.12
Salinas-Seaside-Monterey, CA 0.11 0.09 0.09
Salt Lake City-Ogden, UT 0.15 0.12 0.11
San Antonio, TX 0.11 0.10 0.11
San Diego, CA 0.19 0.17 0.18
San Francisco, CA 0.09 0.06 0.07
San Jose, CA 0.13 0.12 0.12
San Juan, PR 0.06 0.07 0.08
Santa Barbara-Santa Maria-Lompoc, CA 0.16 0.13 0.10
Santa Cruz, CA 0.08 0.08 0.10
Santa Fe, NM 0.05 0.08 0.08
Santa Rosa-Petaluma, CA 0.10 0.08 0.10
Sarasota, FL 0.10 0.10 0.10
Scranton-Wilkes-Barre, PA 0.11 0.11 0.13
Seattle, WA 0.09 0.13 0.11
Sharon, PA 0.11 0.10 0.11
Sheboygan, WI 0.11 0.11 0.16
Shreveport, LA 0.12 0.12 0.11
South Bend-Mishawaka, IN 0.10 0.10 0.11
Spokane, WA 0.07 0.08
Springfield, IL 0.11 0.10 0.10
Springfield, MO 0.09 0.08 0.08
Springfield, MA 0.13 0.12 0.13
Stamford, CT 0.16 0.14 0.15
Steubenville-Weirton, OH-WV 0.11 0.09 0.12
Stockton, CA 0.11 0.12 0.11
Syracuse, NY 0.10 0.11 0.11
Tacoma, WA 0.09 0.13 0.09
Tallahassee, FL 0.07 0.05
Tampa-St. Petersburg-Clearwater, FL 0.10 0.11 0.11
Terre Haute, IN 0.11 0.11 0.10
Toledo, OH 0.11 0.10 0.12
Trenton, NJ 0.14 0.14 0.15
Tucson, AZ 0.10 0.10 0.09
Tulsa, OK 0.12 0.12 0.12
Utica-Rome, NY 0.09 0.10 0.10
Vallejo-Fairfield-Napa, CA 0.11 0.10 0.11
Vancouver, WA 0.09 0.11 0.10
4-18
-------
Table 4-3 (cont'd). The Highest Second Daily Maximum One-Hour
Ozone Concentration (ppm) by Metropolitan Statistical
Area (MSA) for the Years 1989 to 1991
MSA
Victoria, TX
Vineland-Millville-Bridgeton, NJ
Visalia-Tulare-Porterville, CA
Washington, DC-MD-VA
W. Palm Beach-Boca Raton-Delray, FL
Wheeling, WV-OH
Wichita, KS
Williamsport, PA
1989
0.10
0.13
0.15
0.13
0.11
0.11
0.09
0.08
1990
0.07
0.13
0.14
0.13
0.09
0.11
0.10
0.09
1991
0.10
0.12
0.12
0.14
0.09
0.11
0.10
0.10
MSA
Wilmington, DE-NJ-MD
Wilmington, NC
Worcester, MA
York, PA
Youngstown- Warren, OH
Yuba City, CA
Yuma, AZ
1989
0.13
0.10
0.10
0.11
0.01
1990
0.14
0.09
0.12
0.12
0.10
0.09
0.09
1991
0.15
0.14
0.11
0.12
0.10
0.09
high-oxidant urban areas where O3 regimes contain hourly average concentrations with
sharper peaks. The result would be an increase in the number of exceedances of daily
maximum 8-h average concentrations >0.08 ppm, when compared to those sites experiencing
sharper peaks. Lefohn et al. (1993b), using aerometric data at specific sites, observed how
O3 concentrations change when the sites change compliance status. One of the parameters
examined was 4-h daily maxima. The number of exceedances for a specific daily maximum
average concentration tended to decrease as fewer exceedances of the current 1-h standard
were observed at a given site. The number of occurrences of the daily maximum 4-h average
concentration >0.08 ppm and the number of exceedances of the current form of the standard
had a positive, weak correlation (r = 0.51). Lefohn et al. (1993a,b) reported few changes in
the shape of the average diurnal patterns as sites changed attainment status; this may have
explained why Lefohn et al. (1993b) could not find evidence that the number of occurrences
of the daily maximum 4-h average concentration >0.08 ppm increased when the sites
experienced few high hourly average concentrations.
There has been considerable interest in possibly substituting one index for another
when attempting to relate O3 exposure with an effect. For example, using O3 ambient air
quality data, McCurdy (1988) compared the number of exceedances of 0.12 ppm and the
number of occurrences of the daily maximum 8-h average concentrations >0.08 ppm and
reported that a positive correlation (r = 0.79) existed between the second-highest 1-h daily
maximum in a year and the expected number of days with an 8-h daily maximum average
concentration >0.08 ppm O3. In this case, the predictive strength of using one O3 exposure
index to predict another is not strong.
Similar to analysis performed by McCurdy (1988), all of the hourly averaged data
from rural agricultural and forested sites in the AIRS database were summarized into
maximum 3-mo SUM06, second highest daily maximum hourly average concentration, and
second highest daily maximum 8-h average concentration exposure indices per year for the
period 1980 to 1991. For the rural agricultural sites, the correlation coefficients between the
3-mo SUM06 and the second highest daily maximum hourly average concentration and the
second highest daily maximum 8-h average concentration were 0.650 and 0.739, respectively
(Figure 4-5). For the rural forested sites, the correlation coefficients between the 3-mo
SUM06 and the second highest daily maximum hourly average concentration and the second
highest daily maximum 8-h average concentration were 0.585 and 0.683, respectively
(Figure 4-6).
4-19
-------
Table 4-4. Summary of Percentiles of Hourly Average Concentrations (ppm)
for the April-to-October Period3
AIRS Site Name Year
060370016 Glendora, CA 1989
1990
1991
060595001 La Habra, CA 1989
1990
1991
060710005 San Bernardino County, CA 1989
1990
1991
060731001 Del Mar, CA 1989
1990
1991
090013007 Stratford, CT 1989
1990
•£- 1991
O 090093002 Madison, CT 1989
1990
1991
220330003 Baton Rouge, LA 1989
1990
1991
340170006 Bayonne, NJ 1989
1990
1991
360610063 New York, NY 1989
1990
1991
361030002 Babylon, NY 1989
1990
1991
Min.
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.001
0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.001
0.001
0.001
0.000
0.000
0.002
0.001
0.000
0.001
10
0.000
0.000
0.000
0.000
0.000
0.000
0.020
0.020
0.020
0.020
0.020
0.020
0.008
0.010
0.007
0.008
0.008
0.007
0.001
0.000
0.002
0.001
0.001
0.002
0.015
0.014
0.015
0.004
0.006
0.005
30
0.020
0.010
0.010
0.010
0.010
0.010
0.050
0.040
0.040
0.040
0.030
0.040
0.024
0.023
0.019
0.022
0.023
0.023
0.009
0.011
0.010
0.008
0.009
0.011
0.028
0.029
0.029
0.015
0.017
0.018
50
0.030
0.030
0.020
0.030
0.020
0.020
0.060
0.060
0.060
0.040
0.040
0.050
0.036
0.033
0.030
0.033
0.033
0.034
0.021
0.023
0.020
0.021
0.022
0.024
0.040
0.039
0.041
0.027
0.027
0.030
70
0.060
0.050
0.050
0.040
0.040
0.040
0.090
0.080
0.080
0.050
0.050
0.050
0.046
0.044
0.042
0.043
0.043
0.045
0.034
0.038
0.031
0.036
0.036
0.038
0.051
0.051
0.056
0.039
0.040
0.044
90
0.120
0.110
0.100
0.070
0.070
0.070
0.140
0.120
0.120
0.070
0.060
0.060
0.064
0.059
0.060
0.059
0.063
0.065
0.059
0.063
0.054
0.059
0.058
0.065
0.073
0.074
0.082
0.060
0.060
0.067
95
0.150
0.140
0.140
0.090
0.090
0.090
0.160
0.150
0.140
0.080
0.070
0.070
0.077
0.068
0.074
0.070
0.075
0.082
0.069
0.079
0.067
0.074
0.073
0.082
0.086
0.090
0.096
0.073
0.075
0.081
99
0.220
0.200
0.200
0.140
0.140
0.130
0.200
0.180
0.190
0.120
0.100
0.100
0.115
0.100
0.110
0.103
0.107
0.123
0.094
0.109
0.092
0.099
0.106
0.110
0.110
0.116
0.123
0.101
0.105
0.111
Max
0.340
0.290
0.320
0.260
0.210
0.210
0.270
0.330
0.270
0.250
0.170
0.150
0.202
0.176
0.157
0.149
0.197
0.193
0.168
0.187
0.134
0.147
0.185
0.167
0.134
0.175
0.177
0.156
0.146
0.217
Number of Observations
4,874
4,888
4,907
4,875
4,887
4,899
4,871
4,899
4,905
4,814
5,060
5,017
4,673
3,853
4,794
4,272
4,477
4,814
4,964
5,000
4,905
4,815
4,939
4,943
4,825
4,707
4,910
4,407
4,876
4,873
-------
Table 4-4 (cont'd). Summary of Percentiles of Hourly Average Concentrations (ppm)
for the April-to-October Period3
AIRS Site Name
482010024 Harris County, TX
550790085 Bayside, WI
Year
1989
1990
1991
1989
1990
1991
Min.
0.000
0.000
0.000
0.002
0.002
0.002
10
0.000
0.000
0.000
0.006
0.009
0.008
30
0.010
0.010
0.000
0.024
0.025
0.025
50
0.020
0.020
0.020
0.035
0.034
0.035
70
0.030
0.040
0.030
0.046
0.044
0.047
90
0.060
0.070
0.060
0.066
0.061
0.070
95
0.070
0.090
0.080
0.077
0.071
0.081
99
0.110
0.130
0.110
0.101
0.094
0.113
Max
0.230
0.220
0.170
0.151
0.130
0.189
Number of Observations
4,728
4,274
4,322
4,376
4,395
4,303
"See Appendix A for abbreviations and acronyms.
-------
Table 4-5. The Highest Second Daily Maximum Eight-Hour
Average Ozone Concentration (ppm) by Metropolitan Statistical
Area (MSA) for the Years 1989 to 1991
MSA
Akron, OH
Albany-Schenectady-Troy, NY
Albuquerque, NM
Alexandria, LA
Allentown-Bethlehem, PA-NJ
Altoona, PA
Anaheim-Santa Ana, CA
Anderson, IN
Anderson, SC
Ann Arbor, MI
Appleton-Oshkosh-Neenah, WI
Asheville, NC
Atlanta, GA
Atlantic City, NJ
Augusta, GA-SC
Aurora-Elgin, IL
Austin, TX
Bakersfield, CA
Baltimore, MD
Baton Rouge, LA
Beaumont-Port Arthur, TX
Beaver County, PA
Bellingham, WA
Benton Harbor, MI
Bergen-Passaic, NJ
Billings, MT
Biloxi-Gulfport, TX
Birmingham, AL
Bismark, ND
Bloomington-Normal, IL
Boston, MA
Boulder-Longmont, CO
Bradenton, FL
Brazoria, TX
Bridgeport-Milford, CT
Brockton, MA
Buffalo, NY
Canton, OH
Carson City, NV
Cedar Rapids, IA
Champaign-Urbana-Rantoul, IL
Charleston, SC
Charleston, WV
Charlotte-Gastonia-Rock Hill, NC-SC
Charlottesville, VA
Chattanooga, TN-GA
Chicago, IL
Chico, CA
Cincinnati, OH-KY-IN
Cleveland, OH
Colorado Springs, CO
Columbia, SC
Columbus, GA-AL
Columbus, OH
Corpus Christi, TX
Cumberland, MD-WV
Dallas, TX
Danbury, CT
1989
0.109
0.087
0.078
0.077
0.091
0.077
0.146
0.084
0.093
0.091
0.083
0.096
0.104
0.078
0.088
0.099
0.124
0.103
0.095
0.110
0.095
0.038
0.093
0.56
0.088
0.086
0.081
0.109
0.082
0.086
0.139
0.110
0.085
0.098
0.070
0.078
0.084
0.094
0.087
0.089
0.076
0.091
0.101
0.081
0.106
0.099
0.072
0.079
0.068
0.097
0.083
0.101
0.098
1990
0.097
0.091
0.089
0.076
0.098
0.090
0.135
0.087
0.078
0.074
0.125
0.135
0.092
0.077
0.103
0.120
0.111
0.134
0.100
0.085
0.068
0.097
0.105
0.062
0.071
0.105
0.084
0.075
0.101
0.114
0.106
0.096
0.098
0.057
0.080
0.084
0.083
0.100
0.089
0.094
0.084
0.083
0.119
0.096
0.065
0.094
0.075
0.098
0.085
0.070
0.115
0.105
1991
0.102
0.089
0.077
0.074
0.112
0.094
0.110
0.081
0.096
0.082
0.064
0.102
0.112
0.081
0.100
0.084
0.118
0.127
0.100
0.101
0.095
0.059
0.098
0.106
0.089
0.088
0.061
0.095
0.118
0.083
0.074
0.107
0.121
0.107
0.097
0.099
0.066
0.077
0.074
0.099
0.094
0.091
0.083
0.106
0.074
0.115
0.101
0.068
0.083
0.083
0.112
0.075
0.076
0.095
0.116
MSA
Davenport-Rock Island-Moline, IA-IL
Dayton-Springfield, OH
Decatur, IL
Denver, CO
Des Moines, IA
Detroit, MI
Duluth, MN-WI
Eau Claire, WI
El Paso, TX
Elmira, NY
Erie, PA
Eugene-Springfield, OR
Evansville, IN-KY
Fayetteville, NC
Flint, MI
Fort Collins, CO
Ft. Lauderdale-Hollywood-Pompano, FL
Fort Myers-Cape Coral, FL
Fort Wayne, IN
Fort Worth-Arlington, TX
Fresno, CA
Galveston-Texas City, TX
Gary-Hammond, IN
Grand Rapids, MI
Greeley, CO
Green Bay, WI
Greensboro- Winston Salem-High Point, NC
Greenville-Spartanburg, SC
Hamilton-Middletown, OH
Harrisburg-Lebanon-Carlisle, PA
Hartford, CT
Hickory, NC
Honolulu, HI
Houma-Thibodaux, LA
Houston, TX
Huntington-Ashland, WV-KY-OH
Huntsville, AL
Indianapolis, IN
Iowa City, IA
Jackson, MS
Jacksonville, FL
Jamestown-Dunkirk, NY
Janesville-Beloit, WI
Jersey City, NJ
Johnson City-Kingsport-Bristol, TN-WV
Johnstown, PA
Joliet, IL
Kalamazoo, MI
Kansas City, MO-KS
Kenosha, WI
Knoxville, TN
Lafayette, LA
Lafayette, IN
Lake Charles, LA
Lake County, IL
Lancaster, PA
Lansing-East Lansing, MI
Las Cruces, NM
1989
0.102
0.122
0.084
0.089
0.073
0.103
0.073
0.083
0.075
0.092
0.061
0.097
0.089
0.093
0.076
0.089
0.084
0.105
0.098
0.116
0.102
0.102
0.119
0.080
0.095
0.083
0.088
0.095
0.091
0.114
0.020
0.082
0.121
0.102
0.072
0.097
0.078
0.086
0.090
0.097
0.105
0.083
0.082
0.082
0.090
0.113
0.088
0.080
0.077
0.088
0.092
0.085
0.093
0.074
1990
0.071
0.096
0.077
0.086
0.051
0.091
0.051
0.049
0.087
0.080
0.088
0.077
0.094
0.088
0.086
0.076
0.078
0.070
0.091
0.116
0.105
0.096
0.122
0.107
0.080
0.074
0.100
0.091
0.111
0.108
0.109
0.080
0.037
0.084
0.141
0.109
0.080
0.099
0.084
0.083
0.084
0.068
0.081
0.128
0.100
0.090
0.070
0.089
0.093
0.105
0.086
0.092
0.084
0.082
0.089
0.083
0.082
1991
0.086
0.107
0.087
0.080
0.056
0.111
0.080
0.094
0.093
0.070
0.107
0.085
0.090
0.077
0.064
0.062
0.096
0.116
0.119
0.094
0.101
0.124
0.081
0.079
0.087
0.085
0.105
0.100
0.112
0.042
0.077
0.115
0.124
0.082
0.100
0.060
0.075
0.077
0.082
0.090
0.117
0.080
0.099
0.091
0.071
0.089
0.118
0.091
0.075
0.090
0.096
0.102
0.096
0.087
0.074
4-22
-------
Table 4-5 (cont'd). The Highest Second Daily Maximum
Eight-Hour Average Ozone Concentration (ppm) by Metropolitan
Statistical Area (MSA) for the Years 1989 to 1991
MSA
Las Vegas, NV
Lawrence-Haverhill, MA-NH
Lewiston-Auburn, ME
Lexington-Fayette, KY
Lima, OH
Lincoln, NE
Little Rock-North Little Rock, AR
Longview-Marshall, TX
Lorain-Elyria, OH
Los Angeles-Long Beach, CA
Louisville, KY-IN
Lynchburg, VA
Madison, WI
Manchester, NH
Medford, OR
Melbourne-Titusville-Palm Bay, FL
Memphis, TN-AR-MS
Miami-Hialeah, FL
Middlesex-Somerset-Hunterdon, NJ
Middletown, CT
Milwaukee, WI
Minneapolis-St. Paul, MN-WI
Mobile, AL
Modesto, CA
Monmouth-Ocean, NJ
Montgomery, AL
Muskegon, MI
Nashua, NH
Nashville, TN
Nassau-Suffolk, NY
New Bedford, MA
New Haven-Meriden, CT
New London-Norwich, CT-RI
New Orleans, LA
New York, NY
Newark, NJ
Niagara Falls, NY
Norfolk- Virginia Beach-Newport News, VA
Oakland, CA
Oklahoma City, OK
Omaha, NE-IA
Orlando, FL
Owensboro, KY
Oxnard- Ventura, CA
Parkersburg-Marietta, WV-OH
Pascagoula, MS
Pensacola, FL
Peoria, IL
Philadelphia, PA-NJ
Phoenix, AZ
Pittsburgh, PA
Pittsfield, MA
Portland, ME
Portland, OR-WA
Portsmouth-Dover-Rochester, NH-ME
Poughkeepsie, NY
Providence, RI
1989
0.084
0.104
0.089
0.097
0.088
0.057
0.077
0.076
0.096
0.188
0.096
0.089
0.084
0.063
0.082
0.099
0.087
0.108
0.119
0.115
0.090
0.079
0.101
0.118
0.066
0.139
0.072
0.093
0.099
0.104
0.108
0.128
0.075
0.111
0.108
0.082
0.089
0.091
0.089
0.075
0.096
0.096
0.147
0.094
0.082
0.080
0.087
0.118
0.086
0.107
0.075
0.124
0.071
0.107
0.079
0.107
1990
0.082
0.077
0.090
0.097
0.086
0.060
0.083
0.089
0.082
0.170
0.093
0.083
0.079
0.098
0.076
0.082
0.100
0.076
0.111
0.117
0.100
0.077
0.098
0.106
0.107
0.081
0.100
0.095
0.102
0.115
0.101
0.122
0.127
0.086
0.119
0.107
0.092
0.095
0.091
0.090
0.075
0.082
0.104
0.119
0.088
0.092
0.098
0.075
0.110
0.096
0.100
0.094
0.109
0.111
0.086
0.085
0.112
1991
0.075
0.106
0.101
0.088
0.091
0.060
0.089
0.086
0.091
0.178
0.119
0.079
0.089
0.087
0.055
0.069
0.093
0.072
0.111
0.125
0.118
0.079
0.062
0.091
0.122
0.071
0.119
0.110
0.107
0.121
0.106
0.128
0.115
0.079
0.133
0.119
0.095
0.089
0.083
0.089
0.073
0.075
0.077
0.129
0.104
0.077
0.082
0.088
0.123
0.094
0.106
0.095
0.134
0.092
0.123
0.101
0.127
MSA
Provo-Orem, UT
Racine, WI
Raleigh-Durham, NC
Reading, PA
Redding, CA
Reno, NV
Richmond-Petersburg, VA
Riverside-San Bernardino, CA
Roanoke, VA
Rochester, NY
Rockford, IL
Sacramento, CA
St. Louis, MO-IL
Salinas-Seaside-Monterey, CA
Salt Lake City-Ogden, UT
San Antonio, TX
San Diego, CA
San Francisco, CA
San Jose, CA
San Juan, PR
Santa Barbara-Santa Maria-Lompoc, CA
Santa Cruz, CA
Santa Fe, NM
Santa Rosa-Petaluma, CA
Sarasota, FL
Scranton-Wilkes-Barre, PA
Seattle, WA
Sharon, PA
Sheboygan, WI
Shreveport, LA
South Bend-Mishawaka, IN
Spokane, WA
Springfield, IL
Springfield, MO
Springfield, MA
Stamford, CT
Steubenville-Weirton, OH-WV
Stockton, CA
Syracuse, NY
Tacoma, WA
Tallahassee, FL
Tampa-St. Petersburg-Clearwater, FL
Terre Haute, IN
Toledo, OH
Trenton, NJ
Tucson, AZ
Tulsa, OK
Utica-Rome, NY
Vallejo-Fairfield-Napa, CA
Vancouver, WA
Victoria, TX
Vineland-Millvile-Bridgeton, NJ
Visalia-Tulare-Porterville, CA
Washington, DC-MD-VA
W. Palm Beach-Boca Raton-Delray, FL
Wheeling, WV-OH
Wichita, KS
1989
0.094
0.110
0.099
0.095
0.080
0.081
0.094
0.196
0.077
0.094
0.085
0.105
0.105
0.082
0.114
0.100
0.139
0.064
0.094
0.043
0.129
0.066
0.049
0.083
0.085
0.088
0.078
0.098
0.103
0.098
0.089
0.085
0.075
0.123
0.113
0.094
0.086
0.090
0.077
0.072
0.088
0.087
0.093
0.119
0.074
0.093
0.082
0.076
0.058
0.093
0.122
0.114
0.106
0.081
0.086
0.079
1990
0.070
0.090
0.094
0.101
0.100
0.109
0.101
0.193
0.075
0.097
0.073
0.125
0.098
0.074
0.086
0.080
0.135
0.056
0.075
0.042
0.129
0.058
0.069
0.061
0.083
0.096
0.099
0.095
0.088
0.102
0.089
0.060
0.082
0.061
0.113
0.112
0.075
0.093
0.093
0.094
0.085
0.095
0.084
0.112
0.084
0.094
0.094
0.074
0.080
0.056
0.106
0.103
0.110
0.067
0.089
0.089
1991
0.071
0.118
0.091
0.109
0.093
0.075
0.097
0.189
0.078
0.103
0.081
0.124
0.107
0.071
0.086
0.085
0.128
0.054
0.086
0.044
0.075
0.067
0.076
0.076
0.080
0.111
0.087
0.094
0.103
0.087
0.093
0.060
0.087
0.069
0.117
0.115
0.098
0.090
0.098
0.077
0.083
0.089
0.107
0.131
0.080
0.097
0.091
0.078
0.042
0.086
0.108
0.104
0.114
0.059
0.093
0.081
4-23
-------
Table 4-5 (cont'd). The Highest Second Daily Maximum
Eight-Hour Average Ozone Concentration (ppm) by Metropolitan
Statistical Area (MSA) for the Years 1989 to 1991
MSA
1989 1990 1991
MSA
1989 1990 1991
Williamsport, PA
Wilmington, DE-NJ-MD
Wilmington, NC
Worcester, MA
0.065 0.072 0.087
0.105 0.110 0.118
0.086
0.097 0.089 0.107
York, PA
Youngstown-Warren, OH
Yuba City, CA
Yuma, AZ
0.091 0.108 0.103
0.088 0.085 0.101
0.084 0.076 0.084
0.080 0.075 0.070
0.40 -i
(a)
r = 0.670
50 100
3-mo SUM06 (ppm-h)
150
0.25-i
0.00
r = 0.739
50 100
3-mo SUM06 (ppm-h)
150
Figure 4-5. The relationship between (a) the second highest daily maximum hourly
average ozone (OJ concentration and the maximum 3-mo SUM06 value and
(b) the second highest daily maximum 8-h average O3 concentration and the
maximum 3-mo SUM06 value for specific site years at rural agricultural sites
for the 1980-to-1991 period.
4-24
-------
0.20-1
r = 0.585
50 100
3-mo SUM06 (ppm-h)
150
0.15-1
r = 0.683
50 100
3-mo SUM06 (ppm-h)
150
Figure 4-6. The relationship between (a) the second highest daily maximum hourly
average ozone (OJ concentration and the maximum 3-mo SUM06 value and
(b) the second highest daily maximum 8-h average O3 concentration and the
maximum 3-mo SUM06 value for specific site years at rural forested sites for
the 1980-to-1991 period.
One of the difficulties in attempting to use correlation analysis between indices for
rationalizing the substitution of one exposure index for another to predicting an effect (e.g.,
SUM06 versus the second highest daily maximum hourly average concentration) is the
introduction of the error associated with estimating levels of one index from another. Lefohn
et al. (1989) have recommended that if a different exposure index (e.g., second highest daily
maximum hourly average concentration) is to be compared to, for example, the SUM06 for
4-25
-------
adequacy in predicting crop loss, then the focus should be on how well the two exposure
indices predict crop loss using the effects model that is a function of the most relevant index,
and not on how well the indices predict one another. Using data from both urban and rural
O3 monitoring sites in the midwestern United States that were located near agricultural or
forested areas, Lefohn et al. (1989) reported a large amount of scatter between the second
highest daily maximum hourly average concentration and the SUM06 indices. This large
scatter indicated considerable uncertainty when attempting to predict a value for SUM06,
given a specific second highest daily maximum hourly average concentration value. The
authors reported that for a given second highest daily maximum hourly average concentration,
the SUM06 values varied over a large range. Lefohn et al. (1989) concluded that such large
uncertainty would introduce additional uncertainty when attempting to use the predicted
exposure index to estimate an effect. The authors concluded that less error would
be introduced if either of the two indices were used directly in the development of an
exposure-response model.
As pointed out by the U.S. Environmental Protection Agency (1986), a familiar
measure of O3 air quality is the number or percentage of days on which some specific
concentration is equalled or exceeded. This measure, however, does not shed light on one of
the more important questions regarding the effects of O3 on both people and plants: what is
the possible significance of high concentrations lasting 1 h or longer and then recurring on
2 or more successive days?
The recurrence of high O3 concentrations on consecutive days was examined in
four cities (one site in each city) by the U.S. Environmental Protection Agency (1986). The
numbers of multiday events were tallied by length of event (i.e., how many consecutive days)
using data for the daylight hours (0600 to 2000 hours) in the second and third quarters of
1979 through 1981. These sites were selected because they included areas known to
experience high O3 concentrations (e.g., California), and because they represent different
geographic regions of the country (west, southwest, and east).
Because of the importance of episodes and respites, the U.S. Environmental
Protection Agency (1986) commented on the occurrences of the length of episodes and the
time between episodes. The agency concluded that its analysis showed variations among sites
in the lengths of episodes as well as the respite periods. In its discussion, the U.S.
Environmental Protection Agency (1986) defined a day or series of days on which the daily
1-h maximum reached or exceeded the specified level as an "exposure"; the intervening day
or days when that level was not reached was called a "respite". Four O3 concentrations were
selected: 0.06, 0.12, 0.18, and 0.24 ppm. At the Dallas site, for example, the value equalled
or exceeded 0.06 ppm for more than 7 days in a row. The Pasadena site experienced 10 such
exposures, but these 10 exposure events spanned 443 days; in Dallas, the 11 exposures
involved only 168 days. At the lowest concentration (>0.06 ppm), the Dallas station recorded
more short-term (<7 days) exposures (45) involving more days (159) than the Pasadena
station (14 exposures over 45 days) because the daily 1-h maximum
statistic in Pasadena remained above 0.06 ppm for such protracted periods. At concentrations
>0.12 ppm, the lengthy exposures at the Pasadena site resolved into numerous shorter
exposures, whereas in Dallas the exposures markedly dwindled in number and duration.
4-26
-------
4.3.3 Nonurban Area Concentrations
4.3.3.1 Sites That Experience Low Maximum Hourly Average Concentrations
It is important to establish reference points that can be used to describe the
distribution of hourly average concentrations at monitoring sites experiencing low maximum
concentrations. Doing so will make it possible to confirm that the hourly average
concentrations in control chambers utilized for research experiments associated with human
health and vegetation effects are similar to those experienced under ambient conditions. For
example, there has been concern expressed that O3 concentrations in charcoal-filtration
chambers used by NCLAN did not simulate the levels in those areas of the United States that
experience low maximum O3 concentrations (see Chapter 5, Section 5.5). Heuss (1982)
expressed concern that the O3 levels in the charcoal-filtration chambers were lower than those
at sites experiencing low maximum hourly average concentrations, and that the resulting
agricultural loss estimates derived from the NCLAN models may have been too high.
Two possible approaches for establishing reference points have been discussed
(Lefohn et al., 1990a). One method is to estimate, using mathematical models and historical
data, unpolluted background levels prior to disturbance by human influence. However, there
are difficulties with this approach. Background can be defined as the unpolluted conditions in
preindustrial times (i.e., absolutely unpolluted air in which there was no human interference).
Alternatively, a background also can be defined as the condition currently existing at any
location that is presently free from human influence. However, almost all geographic
locations on the earth have been impacted by human influences (e.g., Finlayson-Pitts and
Pitts, 1986; Hong et al., 1994) (see Section 4.3.1).
It is unlikely that a single value or even a fixed range of O3 background values can
apply uniformly across North America or elsewhere in the northern or southern hemispheres.
Any attempt to quantify the historical background is subject to much uncertainty for the
following reasons:
1. Little is known with certainty about the nature of past unpolluted
conditions.
2. Even if all anthropogenic emissions of O3 precursors were
eliminated, it is unlikely that O3 in, for example, eastern North
America, would return to preindustrial levels. Since
preindustrial times, major land use changes have occurred.
Because substantial amounts of natural emissions of
O3 precursors are derived from soils and vegetation, especially
during the warmer months, it is probable that these land use
changes have modified the emissions of O3 precursors and, thus,
changed the concentrations of O3.
Although not representing natural O3 background, attempts have been made, using
historical data, to estimate O3 concentrations in the late 1800s and early 1900s. Model
simulations and limited observations suggest that tropospheric O3 has increased in the
northern hemisphere since the preindustrial times and future increases may be possible
(Bojkov, 1986; Volz and Kley, 1988; Thompson, 1992). However, several investigators have
discussed the possible confounding influences that led to a great deal of uncertainty associated
with characterizing the O3 concentrations measured in the late 1800s and early 1900s (Lisac
and Grubisic, 1991; Lefohn et al., 1992c; Cartalis and Varotsos, 1994). Using data collected
over the past 30 years, consistent trends in tropospheric O3 have not been observed across the
4-27
-------
northern hemisphere (World Meteorological Organization/United Nations Environment
Program, 1994). Using ozonesonde data, observations show that tropospheric O3 has
increased above some locations in the northern hemisphere. However, in the 1980s, the
trends were variable and either small or nonexistent (Logan, 1994). European measurements
at some surface sites indicate an increase in O3 concentration since earlier this century.
Because of the uncertainty associated with measurements taken in the late 1880s and early
1990s, it is difficult to compare O3 concentration levels experienced in the United States with
today's levels. However, it can be concluded that O3 levels measured around 100 years ago
were lower than the values observed today at most sites in the United States, and that
consistently increasing trends in O3 concentration measured at surface levels have not been
observed annually at sites monitoring O3 in the United States.
An alternative approach (adopted by Lefohn et al., 1990a, to establish reference
points) is to examine O3 hourly average concentration data from those monitoring sites in the
world that experienced low maximum hourly average values. When data from these sites,
including several in North America, were characterized, and the range of hourly average
concentrations compared, it was found that the distributions of hourly average concentrations
for sites with the lowest hourly average concentrations were similar. The authors believed the
data from these sites could be used for establishing reference points that could be compared
with more polluted areas.
Is it appropriate to use sites that experience the lowest hourly average
concentrations in the United States today as reference levels, or should O3 background levels
that may have existed 100 years ago be used? Although it might be argued that all sites in
the United States have been affected to such a level that today's values are not relevant
because an increase in O3 concentrations has occurred since the late 1800s, two key points
argue against this line of reasoning. First, as indicated above, although O3 levels have
increased since the last century, consistent trends in tropospheric O3 have not been observed
across the northern hemisphere. Thus, at some monitoring sites, O3 levels may not show
increasing trends. Second, all the increases in background levels of O3 may not necessarily
be associated with changes in anthropogenic emissions. Since preindustrial times, major land
use changes, resulting from human-induced activities, have occurred. Changes in natural
emissions of O3 precursors associated with soils and vegetation in the last 100 years may be
associated with these land use changes, with the result that some of the increased levels in
O3 concentrations may be attributed to sources other than anthropogenic emissions. Thus,
using historical data that contain large uncertainties in the estimation of O3 concentrations as
a reference point, may yield unrealistically low hourly average concentrations.
Based on a review of available data, the U.S. Environmental Protection Agency
(1989) has indicated that a reasonable estimate of O3 background concentration near sea level
in the United States today, for an annual average, is from 0.020 to 0.035 ppm. This estimate
included a 0.005- to 0.015-ppm contribution from stratospheric intrusions and a 0.01-ppm
contribution from photochemically affected biogenic nonmethane hydrocarbons. In addition,
the U.S. Environmental Protection Agency (1989) estimated that an additional 0.010 ppm is
possible from the photochemical reaction of biogenic methane. A more conservative
approach would be to associate the sum of O3 concentrations from these two processes with
the differences between 0.020 to 0.035 ppb and the stratospheric intrusion contribution.
For calculating annual average concentrations, the estimate made by the U.S.
Environmental Protection Agency (1989) may be valid. Pratt et al. (1983), using data from
low-elevation rural sites in Minnesota and North Dakota, reported that annual average
4-28
-------
concentrations for an O3 monitoring site in LaMoure County, ND (400 m), for 1978 through
1981, ranged from 0.030 to 0.035 ppm, whereas an O3 monitoring site in Traverse County,
MN (311 m), had a range of 0.029 to 0.035 ppm. Bower et al. (1989) reported that the
remote northern Scotland site, Strath Vaich (270 m), had a 1987 to 1988 annual average
O3 concentration of 0.031 ppm.
The U.S. Environmental Protection Agency (1989) has estimated that background
O3 concentrations for a 1-h daily maximum at sea level in the United States during the
summer are in the range of 0.03 to 0.05 ppm. However, the actual value may vary from site
to site. Using measurements at a remote site in South Dakota, Kelly et al. (1982) estimated
the background O3 in air masses entering the midwestern and eastern United States to be
0.03 to 0.05 ppm. Lefohn and Foley (1992) reported that hourly average O3 concentration
data available for sites that experience low maximum hourly average values in the western
United States indicate that, in almost all cases, the maximum hourly average concentrations
were in the range of 0.060 to 0.075 ppm, and, at some of the sites, there were infrequent
occurrences of hourly average concentrations below 0.02 ppm (i.e., lack of scavenging).
These observations were similar to those reported for several O3 monitoring sites experiencing
low maximum hourly average O3 concentrations for other locations in the world (Lefohn et
al., 1990a; Pedersen and Lefohn, 1994).
Some vegetation researchers have used the seasonal average of the daily 7-h
(0900 to 1559 hours) average as the exposure parameter in exposure-response models (Heck
et al., 1982). For quantifying the effects of air pollution on crops and trees, some
investigators have used controlled environment and field methods with charcoal-filtration
systems (Olszyk et al., 1989). In both the design of the experiments and the analysis of the
data, the 7-h (0900 to 1559 hours) seasonal mean reference point for O3 was assumed to be
0.025 ppm. The 0.025 ppm concentration was used to estimate crop loss across the United
States (Adams et al., 1985, 1989). For sites experiencing low maximum hourly average
concentrations in the western United States, except for several years of O3 measurements at
Olympic National Park (Table 4-6), the 7-mo (April to October) average of the 7-h daily
average concentrations ranged from 0.025 to 0.045 ppm (Altshuller and Lefohn, 1996).
Ozone hourly average concentrations were characterized at several sites located in
both the western and south-central United States that experienced low maximum hourly
average concentrations (Table 4-6). Redwood National Park, CA; Olympic National Park,
WA; Glacier National Park, MT; Denali National Park, AK; Badlands, SD; Great Sand Dunes
National Monument, CO; Theodore Roosevelt National Park, ND; and Quachita National
Forest, AR, experienced no hourly average concentration >0.08 ppm for the period April to
October (Altshuller and Lefohn, 1996). Except for 1988, the year in which Yellowstone
National Park, WY, experienced a major forest fire, the Wyoming site experienced no hourly
average concentrations >0.08 ppm. Logan (1989) has noted that O3 hourly average
concentrations above 0.08 ppm rarely are exceeded at remote western sites. In almost all
cases for the above sites, the maximum hourly average concentration was <0.075 ppm. There
have been some questions raised to whether the distributions experienced at those sites
exhibiting low maximum hourly average concentrations in the western United States were
representative of sites in the eastern and midwestern United States because of differences in
biogenic precursors. The O3 monitoring site in the
4-29
-------
Table 4-6. Seasonal (April to October) Percentile Distribution of Hourly
Ozone Concentrations, Number of Hourly Mean Ozone Occurrences >0.08 and >0.10,
Seasonal Seven-Hour Average Concentrations, W126, and SUM06 Values for Sites
Experiencing Low Hourly Average Concentrations with Data Capture >75%
(Concentrations in ppm)a
Location Site/AIRS ID Year
Redwood, CA Redwood NP 1988
060150002 1989
1990
1991
1992
1993
Olympic, WA Olympic NP 1982
530090012 1984
1986
1989
1990
.£>. 1991
CO 1993
0
Glacier, MT Glacier NP 1989
300298001 1990
1991
1992
Yellowstone, WY Yellowstone NP 1988
560391010 1989
1990
1991
1992
1993
022900003 1990
1991
1992
1993
Min.
0.002
0.000
0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.002
0.000
0.004
0.001
0.000
0.003
0.005
0.003
0.002
10
0.011
0.010
0.011
0.012
0.010
0.010
0.000
0.000
0.000
0.003
0.005
0.006
0.004
0.003
0.003
0.001
0.001
0.020
0.018
0.015
0.020
0.018
0.018
0.017
0.018
0.016
0.017
30
0.018
0.017
0.018
0.019
0.017
0.017
0.010
0.010
0.010
0.010
0.012
0.014
0.010
0.015
0.014
0.014
0.013
0.029
0.027
0.023
0.030
0.029
0.028
0.024
0.024
0.023
0.023
Percentiles
50 70
0.023
0.022
0.023
0.025
0.021
0.022
0.010
0.010
0.020
0.015
0.018
0.019
0.016
0.026
0.026
0.027
0.025
0.037
0.036
0.029
0.037
0.036
0.036
0.029
0.028
0.028
0.028
0.029
0.027
0.028
0.031
0.026
0.027
0.020
0.020
0.020
0.022
0.023
0.024
0.021
0.036
0.035
0.036
0.033
0.044
0.044
0.036
0.042
0.042
0.042
0.034
0.034
0.034
0.033
90
0.038
0.034
0.035
0.038
0.035
0.035
0.030
0.020
0.040
0.030
0.030
0.033
0.029
0.046
0.044
0.046
0.043
0.054
0.052
0.043
0.048
0.051
0.047
0.040
0.041
0.044
0.041
95
0.041
0.038
0.038
0.041
0.039
0.038
0.030
0.020
0.040
0.035
0.034
0.036
0.034
0.050
0.047
0.049
0.048
0.058
0.057
0.046
0.051
0.056
0.050
0.043
0.043
0.047
0.043
99
0.046
0.042
0.043
0.045
0.045
0.042
0.040
0.030
0.040
0.046
0.043
0.044
0.041
0.058
0.052
0.056
0.055
0.070
0.063
0.053
0.057
0.064
0.054
0.048
0.047
0.050
0.048
Max
0.060
0.047
0.053
0.054
0.055
0.054
0.060
0.050
0.060
0.065
0.064
0.056
0.064
0.067
0.066
0.062
0.077
0.098
0.071
0.061
0.064
0.075
0.060
0.050
0.057
0.054
0.055
No. of
Obs.
4,825
4,624
4,742
4,666
4,679
4,666
4,704
4,872
4,776
4,220
4,584
4,677
4,595
4,770
5,092
5,060
4,909
4,257
4,079
4,663
4,453
4,384
4,399
3,978
4,809
4,800
4,773
Hours
>0.08 >0.10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
17
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
Seasonal
7-h
0.026
0.024
0.025
0.027
0.023
0.025
0.020
0.015
0.025
0.021
0.022
0.025
0.022
0.036
0.036
0.036
0.033
0.043
0.042
0.034
0.042
0.042
0.041
0.030
0.030
0.031
0.030
W126
(ppm-h)
1.8
1.0
1.2
1.7
1.1
1.1
7.4
1.6
13.7
0.7
0.8
0.9
0.7
5.9
4.1
5.3
4.1
14.0
11.0
3.8
7.7
10.7
6.5
2.1
2.7
3.7
2.6
SUMO6
(ppm-h)
0.1
0.0
0.0
0.0
0.0
0.0
0.1
0.0
0.1
0.1
0.3
0.0
0.3
1.8
1.3
0.7
1.0
8.9
6.7
0.5
1.2
6.3
0.2
0.0
0.0
0.0
0.0
-------
Table 4-6 (con't). Seasonal (April to October) Percentile Distribution of Hourly
Ozone Concentrations, Number of Hourly Mean Ozone Occurrences >00.08 and >0.10,
Seasonal Seven-Hour Average Concentrations, W126, and SUM06 Values for Sites
Experiencing Low Hourly Average Concentrations with Data Capture >75%
(Concentrations in ppm)a
Location
Badlands, SD
Great Sand Dunes, CO
Theodore Roosevelt,
ND, North Unit
Point Reyes, CA
Arches, UT
Montgomery Co., AR
Site/ AIRS ID
Badlands NP
460711010
Sand Dunes NM
080030002
Theodore
Roosevelt NP
380530002
Point Reyes NP
060410002
Arches NP
490190101
Quachita NF
050970001
Year
1988
1989
1990
1991
1988
1989
1990
1991
1984
1985
1986
1989
1992
1993
1989
1990
1991
1992
1989
1990
1991
1992
1993
Min.
0.005
0.007
0.006
0.005
0.010
0.011
0.010
0.008
0.000
0.000
0.004
0.004
0.005
0.004
0.007
0.006
0.006
0.007
0.000
0.000
0.000
0.000
0.000
10
0.022
0.020
0.019
0.020
0.028
0.031
0.030
0.029
0.017
0.019
0.017
0.023
0.019
0.018
0.021
0.017
0.019
0.018
0.031
0.020
0.002
0.005
0.008
30
0.032
0.028
0.027
0.028
0.035
0.037
0.037
0.037
0.025
0.026
0.027
0.032
0.027
0.025
0.026
0.022
0.025
0.024
0.040
0.025
0.010
0.015
0.019
Percentiles
50 70
0.038
0.034
0.032
0.034
0.039
0.041
0.041
0.043
0.032
0.032
0.033
0.039
0.033
0.031
0.031
0.025
0.030
0.028
0.045
0.028
0.016
0.024
0.027
0.045
0.041
0.037
0.040
0.044
0.045
0.045
0.048
0.039
0.038
0.039
0.045
0.039
0.037
0.036
0.029
0.034
0.033
0.050
0.031
0.023
0.030
0.035
90
0.053
0.049
0.044
0.047
0.050
0.050
0.052
0.055
0.047
0.046
0.047
0.054
0.047
0.045
0.042
0.036
0.040
0.041
0.057
0.036
0.035
0.041
0.047
95
0.056
0.053
0.048
0.050
0.053
0.052
0.055
0.058
0.050
0.049
0.050
0.058
0.050
0.048
0.045
0.040
0.043
0.045
0.059
0.039
0.041
0.045
0.052
99
0.061
0.060
0.054
0.056
0.058
0.057
0.062
0.065
0.059
0.054
0.056
0.065
0.056
0.055
0.058
0.046
0.048
0.050
0.065
0.045
0.051
0.053
0.061
Max
0.072
0.071
0.063
0.066
0.076
0.063
0.070
0.077
0.068
0.061
0.062
0.073
0.063
0.064
0.080
0.063
0.072
0.066
0.084
0.056
0.064
0.067
0.070
No. of
Obs.
4,791
4,840
4,783
4,584
4,827
4,436
4,624
4,130
4,923
4,211
4,332
4,206
4,332
4,281
4,577
4,856
4,588
4,794
4,260
4,639
4,835
4,902
4,844
Hours
>00.08 >0.10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
2
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
Seasonal
7-h
0.043
0.040
0.037
0.038
0.043
0.044
0.044
0.046
0.038
0.038
0.039
0.046
0.040
0.038
0.033
0.028
0.031
0.031
0.048
0.030
0.027
0.033
0.036
W126
(ppm-h)
13.3
9.2
4.8
6.2
11.1
10.8
13.6
17.0
7.0
5.0
5.5
14.2
6.1
4.6
4.8
1.8
3.0
3.0
21.2
1.7
1.6
3.0
6.7
SUMO6
(ppm-h)
5.2
3.1
0.8
0.7
1.9
1.2
5.1
9.0
2.8
0.1
0.4
11.0
0.8
0.7
2.5
0.4
0.6
0.3
12.6
0.0
0.1
0.9
3.7
"See Appendix A for abbreviations and acronyms.
-------
Quachita National Forest in Arkansas experienced distributions of hourly average
concentrations similar to some of the western sites.
Evans et al. (1983) summarized O3 hourly averaged data collected at eight stations
located in eight national forests across the United States. The first three stations began
operations in 1976 (Green Mountain National Forest, VT; Kisatchie National Forest, LA; and
Custer National Forest, MT); the second three in 1978 (Chequamegon National Forest, WI;
Mark Twain National Forest, MO; and Croatan National Forest, NC); and the last two in
1979 (Apache National Forest, AZ; and Ochoco National Forest, OR). For the period 1979 to
1983, hourly maximum average concentrations at the sites (Custer, Ochoco, and Apache
national forests) were similar to the hourly average concentrations determined for six of the
seven sites characterized by Lefohn and Foley (1992). In almost all cases, none of the sites
experienced hourly average concentrations >0.08 ppm, and the maximum hourly average
concentrations were in the range of 0.060 to 0.075 ppm. Table 4-7 summarizes the percentile
distributions for the three national forest sites.
Several sites experiencing low maximum hourly average concentrations were
characterized by Lefohn et al. (1990a), using various exposure indices. One of the indices
used was W126 (see Section 4.1); the W126 values, calculated over an annual period, are
provided in Table 4-8. The W126 values for Theodore Roosevelt National Park were in the
range of 6.48 to 8.03 ppm-h. The maximum hourly average concentration reported at the site
was 0.068 ppm. The W126 values at the Custer and Ochoco national forests sites ranged
from 5.79 to 22.67 ppm-h. The maximum hourly average concentrations measured at each
site were 0.075 and 0.080 ppm, respectively. The W126 values calculated for the Custer and
Ochoco national forest sites showed greater variability from year to year than the values
calculated for the South Pole, Barrow, and Theodore Roosevelt National Park sites.
As the W126 values increased, the magnitude of the year-to-year variability also
increased. For 2 years of data, the W126 values calculated for the White River U-4 Oil
Shale, UT, site were 19.98 and 32.10 ppm-h. The maximum hourly concentration recorded
was 0.079 ppm. The W126 values calculated for the Apache National Forest site ranged from
10.24 to 81.39 ppm-h. The highest hourly average concentration was 0.090 ppm.
The 7-h (0900 to 1559 hours) average concentration has been used by vegetation
researchers to characterize O3 exposures experienced in plant chamber experiments
(see Chapter 5). Because O3 concentrations are highest during the warm-season months and,
at many low-elevation sites, during daylight hours, the 7-mo seasonal, 7-h (0900 to
1559 hours) average concentration is higher than annual average values. Most remote sites
outside North America experience seasonal 7-h averages of 0.025 ppm (Table 4-9) (Lefohn
et al., 1990a). The seasonal average of the daily 7-h average values for the South Pole,
Antarctica, range from 0.024 to 0.027 ppm. The values range from 0.022 to 0.026 ppm at
Barrow, AK. In the continental United States and southern Canada, values range from
approximately 0.028 to 0.050 ppm (Lefohn et al., 1990a). At an O3 monitoring site at the
Theodore Roosevelt National Park, the range of the 7-mo (April to October) average of the 7-
h daily average concentrations was from 0.038 to 0.046 ppm (Table 4-6). These 7-mo
seasonal averages appear to be representative of values that may occur at other sites located
in the United States and other locations in the northern hemisphere. In earlier investigations,
Lefohn (1984) reported 3-mo (June to August), 7-h averages of 0.048, 0.044, and 0.059 ppm
at remote sites at Custer, Ochoco, and Apache national forests, respectively.
4-32
-------
Table 4-7. Seasonal (April to October) Percentile Distribution of Hourly Ozone
Concentrations, Number of Hourly Mean Ozone Occurrences >0.08 and >0.10,
Seasonal Seven-Hour Average Concentrations, and W126 Values for Three "Clean"
National Forest Sites with Data Capture >75%
(Concentrations in ppm)a
Percentiles
Site AIRS ID Year
CusterNF, MT 300870101 1978
1979
1980
1983
Ochoco NF, OR 410130111 1980
1981
1982
1983
Apache NF, AZ 040110110 1981
1982
1983
Min.
0.000
0.010
0.010
0.010
0.010
0.010
0.010
0.010
0.010
0.015
0.004
10
0.010
0.025
0.025
0.025
0.030
0.025
0.025
0.025
0.025
0.030
0.025
30
0.020
0.035
0.035
0.035
0.035
0.030
0.030
0.035
0.030
0.040
0.035
50
0.035
0.040
0.040
0.040
0.040
0.035
0.035
0.035
0.035
0.045
0.040
70
0.040
0.045
0.050
0.045
0.045
0.040
0.040
0.040
0.040
0.050
0.045
90
0.050
0.050
0.055
0.050
0.055
0.045
0.045
0.045
0.045
0.055
0.055
95
0.055
0.055
0.060
0.055
0.055
0.045
0.050
0.050
0.050
0.060
0.055
99
0.060
0.060
0.065
0.060
0.065
0.055
0.055
0.055
0.055
0.065
0.065
Max
0.070
0.075
0.070
0.065
0.080
0.075
0.065
0.060
0.065
0.075
0.070
No. of
Obs.
4,759
5,014
4,574
4,835
4,759
4,459
4,697
4,423
4,806
4,714
4,788
Hours
>0.08
0
0
0
0
5
0
0
0
0
0
0
>0.10
0
0
0
0
0
0
0
0
0
0
0
Seasonal
7-h (ppm)
0.033
0.043
0.043
0.042
0.044
0.035
0.038
0.039
0.039
0.047
0.042
W126
(ppm-h)
8.3
13.2
19.7
10.7
16.5
4.7
7.6
6.8
7.6
21.9
14.6
CO
CO "See Appendix A for abbreviations and acronyms.
-------
Table 4-8. The Value of the W126 Sigmoidal Exposure Parameter
Calculated Over the Annual Period
(Units in ppm-h)a
Site
South Pole, Antarctica
Bitumount, Alberta, Canada
Barrow, AK
Theodore Roosevelt NP, ND
Custer NF, MT
Ochoco NF, OR
Birch Mountain, Alberta, Canada
White River Oil Shale Project, UT
Fortress Mountain, Alberta, Canada
Apache NF, AZ
Mauna Loa, HI
Whiteface Mountain, NY
Hohenpeissenberg, FRG
American Samoa
Elevation (m)
2,835
350
11
727
1,006
1,364
850
1,600
2,103
2,424
3,397
1,483
975
82
1976 1977 1978 1979 1980
2
2.99
2
14.08 22
19
19.73
81
27
86.50 68
61.28 25.04 35.64 21.76 18
0
.65
.60
.67
.54
.39
.48
.30
.53
.28
1981
3.
2.
5.
10.
45.
33.
29.
0.
72
60
79
24
68
75
53
24
1982
3.
3.
9.
19.
27.
33.
32.
49.
0.
.01
.15
.10
.98
.18
.68
.03
.00
.25
1983 1984
2.
2.
12.
8.
32.
17.
48.
37.
19.
0.
.41 3.54
.36 2.79
8.03
.18
.02
.10
.48
.90 19.18
.82 42.94
.85 40.43
.28 0.30
1985 1986 1987
2.76 4.09
2.03 2.46 3.69
6.69 6.48b
25.04 83.89
32.66 24.48
41.36 32.07 58.33
0.26 0.30 0.32
aSee Appendix A for abbreviations and acronyms.
bCollection did not occur during the months of October, November, and December.
Source: Lefohn et al. (1990a).
-------
Table 4-9. The Value of the Ozone Season (Seven-Month) Average of
the Daily Seven-Hour (0900 to 1559 Hours) Concentration (ppm)a
Site Elevation (m) 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987
South Pole, Antarctica 2,835 0.025 0.027 0.026 0.027 0.024 0.025
Bitumount, Alberta, Canada 350 0.028
Barrow, AK 11 0.022 0.025 0.024 0.024 0.022 0.026 0.022 0.026
Theodore Roosevelt NP, ND 727 0.038 0.039 0.039b
Custer NF, MT 1,006 0.043 0.044 0.042
Ochoco NF, OR 1,364 0.043 0.035 0.038 0.038
Birch Mountain, Alberta, Canada 850 0.036
White River Oil Shale Project, UT 1,600 0.045 0.045
Fortress Mountain, Alberta, Canada 2,103 0.041 0.050
Co Apache NF, AZ 2,424 0.054 0.039 0.047 0.040
cn
MaunaLoa, HI 3,397 0.035 0.039 0.034 0.038 0.035 0.035 0.034
Whiteface Mountain, NY 1,483 0.049 0.046 0.040 0.034 0.041 0.044 0.043 0.043 0.045
Hohenpeissenberg, FRG 975 0.047 0.040 0.044 0.040 0.037 0.043 0.047 0.040 0.043
American Samoa 82 0.010 0.010 0.011 0.009 0.012 0.010 0.011
"See Appendix A for abbreviations and acronyms.
bCollection did not occur during the months of October, November, and December.
Source: Lefohn et al. (1990a).
-------
4.3.3.2 Urban-Influenced Nonurban Areas
It is difficult to identify a set of unique O3 distribution patterns that adequately
describes the hourly average concentrations experienced at monitoring sites in nonurban
locations because, as indicated earlier, many nonurban sites in the United States are
influenced by local sources of pollution or long-range transport of O3 or its precursors.
Unlike the clean sites characterized by Lefohn and Jones (1986), Lefohn et al. (1990a), and
Lefohn and Foley (1992), urban-influenced nonurban sites sometimes show frequent hourly
average concentrations near the minimum detectable level, but almost always show
occurrences of hourly average concentrations above 0.1 ppm. The frequent occurrence of
hourly average concentrations near the minimum detectable level is indicative of scavenging
processes (i.e., NOX); the presence of high hourly average concentrations can be attributable to
the influence of either local generation or the long-range transport of O3. For example, Evans
et al. (1983) reported that the Green Mountain (VT) and Mark Twain national forest (MO)
sites were influenced by long-range transport of O3. The U.S. Environmental Protection
Agency (1986) reported that the maximum hourly average concentrations at Green Mountain
(for the period 1977 to 1981) and Mark Twain (for the period 1979 to 1983) were 0.145 and
0.155 ppm, respectively. Using hourly averaged data from the AIRS database for a select
number of rural monitoring sites, Table 4-10 summarizes the percentiles of the hourly average
O3 concentrations, the number of occurrences of the hourly average concentration >0.10 ppm,
and the 3-mo sum of all hourly average concentrations >0.06 ppm.
As part of a comprehensive air monitoring project sponsored by the Electric Power
Research Institute (EPRI), O3 measurements were made by the chemiluminescence method
from 1977 through 1979 at nine "nonurban" Sulfate Regional Experiment (SURE) Program
sites and Eastern Regional Air Quality Study sites in the eastern United States. On the basis
of diurnal NOX patterns that indicated the influence of traffic emissions, five of the sites were
classed as "suburban", and the other four were classed as "rural". The O3 data from these
nine stations are summarized in Table 4-11. The sites are influenced either by local sources
or by transport of O3 or its precursors. The maximum hourly average concentrations
generally are higher than 0.125 ppm, and the occurrence of hourly average concentrations
near minimum detectable levels indicates NOX scavenging processes.
As part of its effort to provide long-term estimates of dry acidic deposition across
the United States, the National Dry Deposition Network (NDDN) operated more than 50 sites,
which included 41 in the eastern United States and 9 in the western United States, that
routinely recorded hourly average O3 concentrations. Figure 4-7 shows the locations of the
NDDN sites. Edgerton and Lavery (1992) have summarized the O3 concentrations at some of
the sites for the period 1988 to 1990. Table 4-12 summarizes the 7-h (0900 to 1559 hours)
growing season average concentration (May to September) for selected sites in the Midwest
and the East. Fifty-nine percent of the monitoring sites listed in the table have been classified
as agricultural and 36% as forested; one site was classified as commercial. As noted by the
U.S. Environmental Protection Agency (1992a), 1988 was an exceptionally high
O3 concentration year, when compared with 1989 and 1990. The number of hourly
O3 concentrations >0.08 ppm is presented in Table 4-12. Edgerton and Lavery (1992) have
summarized O3 hourly average concentration data for several sites using the cumulative
4-36
-------
Table 4-10. Summary of Percentiles, Number of Hourly Occurrences >0.10 ppm,
and Three-Month SUM06 Values for Selected Rural Ozone Monitoring Sites
in 1989 (April to October)
(Concentrations in ppm)a
AIRS Site Name
RURAL AGRICULTURAL
170491001 Effingham County, IL
180970042 Indianapolis, IN
240030014 Anne Arandel, MD
310550032 Omaha, NE
420070003 New Brighton, PA
5 1 06 1 0002 Fauquier County, VA
RURAL FOREST
060430004 Yosemite NP, CA
360310002 Essex County, NY
470090101 Smoky Mountain NP, TN
511870002 Shenandoah NP (Dickey Ridge), VA
RURAL OTHER
040132004 Scottsdale, AZ
350431001 Sandoval County, NM
370810011 Guilford County, NC
371470099 Farmville, NC
550270001 Horicon, WI
551390007 Oshkosh, WI
Min.
0.000
0.001
0.000
0.002
0.000
0.000
0.000
0.016
0.000
0.004
0.000
0.000
0.004
0.000
0.002
0.002
10
0.009
0.006
0.006
0.021
0.008
0.009
0.008
0.031
0.025
0.027
0.006
0.010
0.010
0.010
0.019
0.016
30
0.023
0.021
0.021
0.030
0.021
0.021
0.022
0.040
0.036
0.037
0.018
0.020
0.023
0.023
0.029
0.028
50
0.036
0.034
0.032
0.037
0.032
0.033
0.035
0.049
0.044
0.045
0.031
0.030
0.034
0.034
0.037
0.038
Percentiles (ppm)
70 90
0.046
0.046
0.045
0.047
0.043
0.045
0.049
0.056
0.053
0.054
0.045
0.040
0.046
0.044
0.047
0.048
0.063
0.063
0.064
0.062
0.062
0.061
0.065
0.066
0.065
0.065
0.062
0.060
0.063
0.062
0.062
0.063
95
0.070
0.072
0.073
0.067.0
.070
0.069
0.072
0.072
0.070
0.071
0.071
0.060
0.070
0.070
0.070
0.070
99
0.081
0.085
0.090
0.075
0.087
0.084
0.083
0.086
0.081
0.082
0.084
0.070
0.083
0.083
0.088
0.084
Max
0.104
0.103
0.120
0.098
0.102
0.122
0.111
0.106
0.098
0.100
0.107
0.090
0.113
0.100
0.111
0.121
Number of Number of Max Uncorrected
Hourly Occurrences 3-mo SUM06 Value
Values >0. 10 (ppm-h)
4,600
4,592
4,360
4,160
5,055
5,050
4,853
4,792
4,764
4,454
5,070
5,059
4,853
4,833
4,142
4,206
1
3
10
0
4
5
3
4
0
1
4
0
2
2
11
3
25.3
25.4
25.5
24.9
29.4
24.6
37.6
45.6
35.9
33.5
31.7
25.1
27.7
26.4
24.6
27.9
"See Appendix A for abbreviations and acronyms.
-------
03
Table 4-11. Summary of Percentiles of Hourly Average Concentrations for Electric Power
Research Institute Sulfate Regional Experiment (SURE) Program Sites/Eastern Regional Air
Quality Study (ERAQS) Ozone Monitoring Sites
(Units in ppm)
SURE/ERAQS Name
Montague, MA
Scranton, PA
Indian River, DE
Duncan Falls, OH
Rockport, IN
Giles County, TN
Roanoke, IN
Research Triangle Park, NC
Lewisburg, WV
Year
1978
1979
1978
1979
1978
1979
1978
1979
1978
1979
1978
1979
1978
1979
1978
1979
1978
1979
Min.
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.000
10
0.002
0.000
0.015
0.011
0.010
0.008
0.005
0.010
0.008
0.008
0.000
0.000
0.004
0.004
0.001
0.001
0.020
0.013
30
0.018
0.013
0.031
0.022
0.024
0.020
0.022
0.021
0.021
0.019
0.018
0.014
0.019
0.017
0.017
0.012
0.034
0.022
Percentiles
50 70
0.032
0.025
0.040
0.030
0.035
0.031
0.034
0.029
0.032
0.028
0.032
0.024
0.032
0.026
0.032
0.024
0.045
0.029
0.043
0.035
0.048
0.040
0.049
0.042
0.049
0.042
0.044
0.038
0.046
0.036
0.044
0.038
0.049
0.037
0.056
0.039
90
0.061
0.056
0.062
0.061
0.072
0.063
0.071
0.060
0.066
0.055
0.066
0.055
0.067
0.061
0.076
0.058
0.072
0.056
95
0.075
0.070
0.073
0.074
0.085
0.073
0.081
0.069
0.078
0.064
0.075
0.065
0.079
0.074
0.087
0.068
0.079
0.065
99
0.119
0.103
0.094
0.097
0.103
0.092
0.110
0.086
0.101
0.083
0.087
0.081
0.106
0.098
0.108
0.084
0.091
0.080
Max
0.202
0.149
0.126
0.132
0.134
0.138
0.144
0.110
0.145
0.104
0.110
0.130
0.160
0.133
0.142
0.131
0.115
0.099
Number of
Observations
7,138
8,485
5,461
8,313
6,874
8,527
5,125
7,595
6,849
8,391
6,034
8,439
5,874
8,001
7,081
8,652
7,019
7,849
-------
Upper
Northeast
South
Central
Figure 4-7. The location of National Dry Deposition Network monitoring sites as of
December 1990.
Source: Edgerton and Lavery (1992).
integrated exposure index, W126, as proposed by Lefohn and Runeckles (1987). Based on
evidence presented in the literature relating O3 exposure with agricultural yield reduction, the
index was proposed as a way to weight the higher hourly average concentrations greater than
the lower values. The data in the table illustrate the large differences in cumulative exposure
between those that occurred in 1988 and those that were experienced in 1989 and 1990. The
percentile of the hourly average concentrations is summarized in Table 4-13. Although
several of the monitoring sites are located in fairly remote locations in the eastern United
States (based on land use characterization) the maximum hourly average concentrations reflect
the transport of O3 or its precursors into the area.
Taylor et al. (1992) have summarized the O3 concentrations that were experienced
at 10 EPRI Integrated Forest Study sites in North America. The authors reported that in 1988
all sites experienced maximum hourly average concentrations >0.08 ppm. In almost all cases,
the sites experienced multiple occurrences above 0.08 ppm. This implies that,
4-39
-------
Table 4-12. Seven-Hour Growing Season Mean, W126 Values, and Number of Hourly Ozone
Concentrations >80 ppb for Selected Eastern National Dry Deposition Network Sites3
Subregion
NORTHEAST
Connecticut Hill
Washington's Crossing
Pennsylvania State University
Laurel Hill State Park
Beltsville
Cedar Creek State Park
UPPER NORTHEAST
Whiteface Mountain
Ashland
MIDWEST
Argonne National Lab
Vincennes
Oxford
UPPER MIDWEST
Unionville
Perkinstown
SOUTH CENTRAL
Sand Mountain
Georgia Station
Perryville
Research Triangle Park
Coweeta
Edgar Evins State Park
Horton Station
SOUTHERN PERIPHERY
Caddo Valley
Sumatra
State
NY
NJ
PA
PA
MD
WV
NY
ME
IL
IN
OH
MI
WI
AL
GA
KY
NC
NC
TN
VA
AR
FL
Site
110
144
106
117
116
119
105
135
146
140
122
124
134
152
153
129
101
137
127
120
150
156
Land Classb
RF
RA
RA
RF
RA
RA
RF
RA
RA
RA
RA
RA
RA
RA
RA
RA
RC
RF
RF
RF
RF
RF
7-h Mean (ppb)
1988 1989 1990
55.0
59.0
62.7
—
59.8
43.5
—
61.1
62.0
65.3
—
—
—
65.2
62.3
55.6
62.3
—
—
48.3
52.8
46.0
48.4
54.6
44.9
45.7
37.9
51.4
51.1
53.5
51.5
44.2
52.6
48.1
50.8
50.8
41.0
47.2
51.4
46.2
39.8
45.3
52.4
51.0
48.6
55.5
48.2
42.3
35.3
46.3
50.9
51.7
47.4
38.8
63.6
62.6
47.9
56.1
54.4
49.5
46.3
1988
75.5
63.5
68.8
—
50.4
37.8
—
59.1
68.1
91.8
—
—
—
103.6
62.3
44.3
127.6
—
—
W126 (ppm-h)
1989 1990
40.3
46.0
25.4
29.1
45.4
19.6
25.3
9.1
29.6
36.4
48.4
35.4
19.0
40.6
28.1
39.7
31.7
16.1
26.9
61.2
18.5
17.8
36.8
43.7
42.7
31.0
45.7
24.3
31.2
8.7
21.6
35.8
46.4
30.7
11.6
68.7
69.7
21.3
44.5
70.6
21.0
20.0
1988
86.8
65.6
75.5
—
56.3
40.9
—
69.4
78.5
103.2
—
—
—
99.6
71.0
150.7
—
—
SUM06 (ppm-h)
1989 1990
47.2
52.1
28.1
30.8
48.6
19.0
25.2
5.4
35.0
40.3
55.3
41.6
18.3
33.2
21.8
38.9
35.5
24.4
64.0
15.6
16.5
35.8
48.4
45.0
32.1
49.4
23.2
29.0
5.8
25.7
41.2
51.7
31.6
7.6
83.4
77.7
49.2
82.8
25.2
17.4
SUM08 (ppm-h)
1988 1989 1990
44.3
32.3
41.6
—
27.0
17.0
—
32.3
36.7
56.8
—
—
—
39.7
20.5
60.2
—
—
5.5
21.2
5.0
7.0
22.9
4.1
2.6
0.6
10.4
8.7
15.8
9.0
0.2
3.4
4.6
5.2
6.6
1.5
8.5
0.2
1.0
3.3
21.3
11.6
8.0
21.9
6.5
8.4
0.8
3.6
11.8
17.2
7.1
0.0
24.0
28.4
8.2
9.2
2.3
0.9
"See Appendix A abbreviations for and acronyms.
bRA = Rural agricultural; RF = Rural forest; RC = Rural commercial; — = No data or insufficient data.
Source: Edgerton and Lavery (1992).
-------
Table 4-13. Summary of Percentiles for National Dry Deposition Network Monitoring Sites
(Units in ppm)
Site No.
RURAL
106
116
119
122
124
129
134
135
140
144
146
Name
AGRICULTURAL SITES
Pennsylvania State University, PA
Beltsville, MD
Cedar Creek, WV
Oxford, OH
Unionville, MI
Perryville, KY
Perkinstown, WI
Loring AFB/Ashland, ME
Vincennes, IN
Washington Crossing, NJ
Argonne National Laboratory, IL
Year
1988
1989
1990
1989
1990
1988
1989
1990
1988
1989
1990
1989
1990
1988
1989
1989
1990
1989
1990
1988
1989
1990
1989
1990
1988
1989
1990
Min.
0.000
0.000
0.000
0.002
0.000
0.000
0.001
0.001
0.001
0.001
0.000
0.003
0.004
0.002
0.001
0.007
0.006
0.002
0.002
0.000
0.000
0.000
0.000
0.001
0.000
0.000
0.000
10
0.013
0.010
0.015
0.003
0.001
0.008
0.006
0.007
0.019
0.017
0.015
0.021
0.020
0.024
0.020
0.023
0.020
0.017
0.014
0.007
0.009
0.009
0.006
0.008
0.004
0.005
0.004
30
0.026
0.022
0.027
0.014
0.015
0.017
0.013
0.014
0.032
0.029
0.028
0.031
0.029
0.038
0.033
0.032
0.028
0.026
0.023
0.024
0.025
0.025
0.021
0.021
0.019
0.019
0.017
Percentiles
50 70
0.036
0.033
0.038
0.029
0.027
0.029
0.024
0.024
0.044
0.039
0.037
0.038
0.036
0.049
0.043
0.038
0.035
0.032
0.029
0.036
0.036
0.035
0.033
0.032
0.032
0.029
0.028
0.049
0.043
0.048
0.044
0.041
0.044
0.037
0.038
0.058
0.050
0.048
0.047
0.044
0.062
0.052
0.046
0.041
0.039
0.036
0.052
0.047
0.045
0.046
0.043
0.046
0.041
0.039
90
0.073
0.059
0.065
0.068
0.067
0.069
0.056
0.057
0.083
0.069
0.067
0.063
0.061
0.080
0.066
0.057
0.050
0.049
0.046
0.076
0.064
0.062
0.067
0.065
0.073
0.061
0.057
95
0.086
0.066
0.074
0.081
0.080
0.082
0.065
0.067
0.096
0.077
0.077
0.071
0.069
0.094
0.072
0.062
0.056
0.055
0.051
0.089
0.072
0.073
0.078
0.079
0.085
0.070
0.065
99
0.114
0.082
0.090
0.096
0.103
0.108
0.082
0.085
0.117
0.092
0.092
0.086
0.084
0.110
0.086
0.071
0.065
0.063
0.068
0.104
0.085
0.089
0.100
0.104
0.103
0.088
0.077
Max Number of Observations
0.143
0.104
0.120
0.131
0.137
0.134
0.172
0.116
0.221
0.109
0.116
0.113
0.105
0.143
0.102
0.085
0.074
0.103
0.088
0.120
0.112
0.110
0.159
0.148
0.146
0.126
0.097
4,716
5,089
5,056
5,062
4,597
4,938
5,044
5,025
4,746
5,073
5,077
5,041
5,065
4,061
4,787
5,029
5,063
5,067
5,080
4,908
5,065
5,084
5,053
5,058
5,037
5,055
5,033
-------
Table 4-13 (cont'd). Summary of Percentiles for National Dry Deposition Network Monitoring Sites
(Units in ppm)
Site No.
RURAL
152
153
RURAL
105
110
117
120
127
137
150
156
Name
AGRICULTURAL SITES (cont'd)
Sand Mountain, AL
Georgia Station, GA
FOREST SITES
Whiteface Mountain, NY
Ithaca, NY
Laurel Hill, PA
Horton Station, VA
Edgar Evins State Park, TN
Coweeta, NC
Caddo Valley, AR
Sumatra, FL
Year
1989
1990
1989
1990
1988
1989
1990
1988
1989
1990
1988
1989
1990
1988
1989
1990
1989
1990
1988
1989
1990
1989
1990
1989
1990
Min.
0.000
0.000
0.002
0.002
0.000
0.003
0.005
0.005
0.002
0.001
0.001
0.000
0.001
0.010
0.002
0.004
0.000
0.001
0.001
0.001
0.000
0.002
0.002
0.001
0.000
10
0.020
0.021
0.014
0.021
0.016
0.022
0.018
0.025
0.025
0.022
0.012
0.009
0.009
0.031
0.032
0.032
0.017
0.019
0.010
0.007
0.008
0.005
0.004
0.012
0.011
30
0.031
0.035
0.025
0.034
0.026
0.030
0.028
0.034
0.036
0.033
0.025
0.020
0.020
0.045
0.043
0.044
0.028
0.032
0.022
0.016
0.018
0.016
0.015
0.022
0.023
Percentiles
50 70
0.041
0.045
0.034
0.044
0.034
0.038
0.036
0.043
0.044
0.041
0.036
0.031
0.030
0.057
0.050
0.052
0.037
0.041
0.034
0.025
0.029
0.028
0.029
0.030
0.033
0.051
0.057
0.045
0.056
0.044
0.047
0.046
0.055
0.052
0.049
0.050
0.043
0.042
0.067
0.059
0.059
0.047
0.052
0.047
0.037
0.043
0.041
0.041
0.040
0.043
90
0.065
0.074
0.062
0.073
0.062
0.059
0.060
0.080
0.065
0.063
0.076
0.061
0.060
0.084
0.070
0.071
0.062
0.067
0.065
0.055
0.059
0.057
0.057
0.057
0.057
95
0.072
0.080
0.069
0.084
0.074
0.066
0.069
0.090
0.071
0.069
0.092
0.069
0.071
0.096
0.076
0.075
0.067
0.073
0.072
0.061
0.064
0.063
0.065
0.065
0.063
99
0.082
0.093
0.082
0.102
0.098
0.078
0.086
0.103
0.081
0.081
0.119
0.087
0.086
0.114
0.085
0.084
0.077
0.085
0.094
0.071
0.072
0.075
0.077
0.075
0.072
Max Number of Observations
0.097
0.117
0.118
0.144
0.129
0.093
0.115
0.126
0.101
0.093
0.156
0.110
0.109
0.145
0.103
0.097
0.090
0.109
0.145
0.094
0.085
0.102
0.094
0.098
0.118
4,509
5,068
3,540
4,814
5,051
4,698
5,016
4,827
5,064
5,075
5,007
4,697
5,032
5,012
4,976
5,066
5,060
5,027
4,182
4,275
5,046
5,046
5,078
4,700
4,444
RURAL COMMERCIAL SITE
101
Research Triangle Park, NC
1988
1989
0.000
0.000
0.004
0.004
0.020
0.019
0.035
0.030
0.050
0.042
0.072
0.063
0.084
0.071
0.111
0.083
0.137
0.121
5,030
4,893
-------
although the sites were located in remote forested areas, the sites experienced elevated
O3 concentrations that were more than likely due to long-range transport of O3 or its
precursors.
Ozone concentrations on a seasonal basis in the Shenandoah National Park exhibit
some features in common with both urban and rural areas. During some years, maximum
hourly average concentrations exceed 0.12 ppm, although some sites in the park exhibit a lack
of hourly average concentrations near minimum detectable level. Taylor and Norby (1985)
have characterized O3 episodes, which they defined as any day in which a 1-h mean
O3 concentration was >0.08 ppm. Based on a 4-year monitoring period in the park, the
probability was 80% that any given episode during the growing season would last 2 or more
days, whereas the probabilities of episodes lasting for periods greater than 3, 4, and 5 days
were 30, 10, and 2%, respectively. Single-day O3 episodes were infrequent. Taylor and
Norby (1985) noted that, given the frequency of respites, there was a 50% probability that a
second episode would occur within 2 weeks.
Because of a lack of air quality data collected at rural and remote locations, it has
been necessary to use interpolation techniques to estimate O3 exposures in nonurban areas. In
the absence of actual O3 data, interpolation techniques have been applied to the estimation of
O3 exposures across the United States (Reagan, 1984; Lefohn et al., 1987a; Knudsen and
Lefohn, 1988). "Kriging", a mathematical interpolation technique, has been used to provide
estimates of seasonal O3 values for the NCLAN for 1978 through 1982 (May to September of
each year) (Reagan, 1984). These values, along with updated values, coupled with exposure-
response models were used to predict agriculturally related economic benefits anticipated by
lower O3 levels in the United States (Adams et al., 1985; Adams et al., 1989).
Kriging is a statistical tool developed by Matheron (1963) and named in honor of
D.G. Krige. Although originally developed specifically for ore reserve estimation, kriging has
been used for other spatial estimation applications, such as analyzing and modeling air quality
data (Grivet, 1980; Faith and Sheshinski, 1979). At its simplest, kriging can be thought of as
a way to interpolate spatial data much as an automatic contouring program would. In a more
precise manner, kriging can be defined as a best, linear unbiased estimator of a spatial
variable at a particular site or geographic area. Kriging assigns low weights to distant
samples and vice versa, but also takes into account the relative position of the samples to
each other and the site or area being estimated.
Figure 4-8 shows the average for the 1985 through 1987 period for the seasonal
(April to October) average of the daily maximum 7- and 12-h values across the United States.
The estimates made for the Rocky Mountain region had large uncertainties associated with
them because of a lack of monitoring sites.
Because of the importance of the higher hourly average concentrations in eliciting
injury and yield reduction for agricultural crops (U.S. Environmental Protection Agency,
1992b), kriging was used to predict O3 exposures in the eastern United States. The
sigmoidally weighted W126 exposure index was used as described earlier in this section.
Lefohn et al. (1992b) used the W126 index in its kriging to characterize the O3 exposures that
occurred during the period 1985 to 1989. Figure 4-9 illustrates the integrated O3 exposure for
the 1988 and 1989 periods (data derived from work described in Lefohn et al., 1992b). Using
the kriged data in the East, the 1988 exposures were the highest for the 5-year period,
whereas 1989 exhibited the lowest exposures. The O3 gradient pattern analyses described by
Lefohn et al. (1992b) identified contiguous areas of persistent,
4-43
-------
(a)
(b)
Figure 4-8. The kriged 1985 to 1986 maximum (a) 7-h and (b) 12-h average
concentrations of ozone across the United States.
Source: Lefohn et al. (1991).
4-44
-------
(a)
(b)
Figure 4-9. The kriged estimates of the W126 integrated ozone exposure index for the
eastern United States for (a) 1988 and (b) 1989.
Source: Lefohn et al. (1992b).
4-45
-------
relatively high seasonal O3 values. The largest area extended from New Jersey south to
northern Georgia and South Carolina. This area was roughly bounded on the west by the
Appalachian Mountains. A second area, which exhibited persistent relatively high seasonal
O3 exposures, was centered over the Ohio River Valley in the region near the Kentucky-
Indiana-Ohio borders. Relatively low O3 exposures were found in Minnesota, Iowa,
Wisconsin, Maine, Vermont, New Hampshire, and Florida. On a year-to-year basis, the
analysis by Lefohn et al. (1992b) showed that regions that tended to be high for a specific
year continued to experience O3 exposures that were higher when compared to other regions.
4.4 Diurnal Variations in Ozone Concentrations
4.4.1 Introduction
By definition, diurnal variations are those that occur during a 24-h period. Diurnal
patterns of O3 may be expected to vary with location, depending on the balance among the
many factors affecting O3 formation, transport, and destruction. Although they vary with
locality, diurnal patterns for O3 typically show a rise in concentration from low (or levels near
minimum detectable amounts) to an early afternoon peak. The 1978 criteria document (U.S.
Environmental Protection Agency, 1978) ascribed the diurnal pattern of concentrations to
three simultaneous processes: (1) downward transport of O3 from layers aloft; (2) destruction
of O3 through contact with surfaces and through reaction with nitric oxide (NO) at ground
level; and (3) in situ photochemical production of O3 (U.S. Environmental Protection Agency,
1978; Coffey et al., 1977; Mohnen et al., 1977; Reiter, 1977a).
The form of an average diurnal pattern may provide information on sources,
transport, and chemical formation and destruction effects at various sites (Lefohn, 1992b).
Nontransport conditions will produce early afternoon peaks. However, long-range transport
processes will influence the actual timing of a peak from afternoon to evening or early
morning hours. Investigators have utilized diagrams that illustrate composite diurnal patterns
as a means to describe qualitatively the differences in O3 exposures between sites (Lefohn and
Jones, 1986; Bohm et al., 1991). Although it might appear that composite diurnal pattern
diagrams could be used to quantify the differences in O3 exposures between sites, Lefohn
et al. (1991) cautioned their use for this purpose. The average diurnal patterns are derived
from long-term calculations of the hourly average concentrations, and the resulting diagram
cannot identify adequately, at most sites, the presence of high hourly average concentrations
and, thus, may not adequately distinguish O3 exposure differences among sites. Logan (1989)
noted that diurnal variation of O3 did not reflect the presence of high hourly average
concentrations.
Unique families of diurnal average profiles exist, and it is possible to distinguish
between two types of O3 monitoring sites. A seasonal diurnal diagram provides the
investigator with the opportunity to identify whether a specific O3 monitoring site has more
scavenging than any other site. Ozone is rapidly depleted near the surface below the
nocturnal inversion layer (Berry, 1964). Mountainous sites, which are above the nocturnal
inversion layer, do not necessarily experience this depletion (Stasiuk and Coffey, 1974).
Taylor and Hanson (1992) reported similar findings using data from the Integrated Forest
Study. For the low-elevation sites, the authors reported that intraday variability was most
significant due to the pronounced daily amplitude in O3 concentration between the predawn
minimum and midafternoon-to-early evening maximum. The authors reported that the
4-46
-------
interday variation was more significant in the high-elevation sites. Ozone trapped below the
inversion layer is depleted by dry deposition and chemical reactions if other reactants are
present in sufficient quantities (Kelly et al., 1984). Above the nocturnal inversion layer, dry
deposition generally does not occur, and the concentration of O3 scavengers generally is
lower, so O3 concentration remains fairly constant (Wolff et al., 1987). A flat diurnal pattern
is usually interpreted as indicating a lack of efficient scavenging of O3 or a lack of
photochemical precursors, whereas a varying diurnal pattern is taken to indicate the opposite.
With the composite diagrams alone, it is difficult to quantify the daily or long-term exposures
of O3. For example, the diurnal patterns for two such sites are illustrated in Figure 4-10. The
Jefferson County, KY, site is urban-influenced and experiences elevated levels of O3 and NOX.
The Oliver County, ND, site is fairly isolated from urban-influenced sources and hourly
average O3 concentrations are mostly below 0.09 ppm. The flat diurnal pattern observed for
the Oliver County site is usually interpreted as indicating a lack of efficient scavenging of
O3 or a lack of photochemical precursors, whereas the varying diurnal pattern observed at the
Jefferson County site may be interpreted to indicate the opposite. Logan (1989) has described
the diurnal pattern for several rural sites in the United States (Figure 4-11) and noted that
average daily profiles showed a broad maximum from about 1200 hours until about 1800
hours at all the eastern sites, except for the peak of Whiteface Mountain, NY. Logan (1989)
noted that the maximum concentrations were higher at the SURE sites than at the Western
National Air Pollution Background Network sites in the East, because the latter were situated
in more remote or coastal locations.
There is concern that the highest hourly average concentrations observed at rural
agricultural and forested sites occur outside the most biologically active period of the day.
To address this concern, a review of the hourly average data collected at all rural agricultural
and forested sites in EPA's AIRS database for 1990 to 1992 was undertaken to evaluate the
percentage of time hourly average concentrations >0.1 ppm occurred during the period 0900
to 1559 hours in comparison with the 24-h period. Each rural site for each of the 3 years that
experienced a 3-mo SUMO6 >26.4 ppm-h (see Chapter 5, Section 5.6 for more details
concerning the use of a 3-mo SUMO6 exposure index) was characterized. It was found that
70% of the rural agricultural and forested sites used in the analysis experienced at least 50%
of the occurrences >0.1 ppm during the period 0900 to 1559 hours when compared to the 24-
h period (Figure 4-12). When O3 monitoring sites in California were eliminated,
approximately 73% of the remaining sites experienced at least 50% of the occurrences >0.10
ppm during the daylight 7-h period when compared with the 24-h period (Figure 4-13).
Reviewing Figures 4-12 and 4-13, for most rural agricultural and forested sites that
experience 3-mo SUMO6 >26.4 ppm in the United States, most of the hourly average
concentrations >0.1 ppm occur during the 0900 to 1559 hours period.
4.4.2 Urban Area Diurnal Patterns
The U.S. Environmental Protection Agency (1986) has discussed diurnal patterns
for urban sites. Figure 4-14, reproduced from the previous document, shows the diurnal
pattern of O3 concentrations on July 13, 1979, in Philadelphia, PA. On this day a peak 1-h
average concentration of 0.20 ppm, the highest for the month, was reached at 1400 hours,
presumably as the result of meteorological factors, such as atmospheric mixing and local
4-47
-------
0.08-
0.06-
o-
Jefferson Co., KY
Oliver Co., ND
9
11 13
Hour
15 17 19 21
23
Figure 4-10. The comparison of the seasonal diurnal patterns using 1988 data for
Jefferson County, KY, and Oliver County, ND.
photochemical processes. The severe depression of concentrations to below detection limits
(less than 0.005 ppm) between 0300 and 0600 hours usually is explained as resulting from the
scavenging of O3 by local NO emissions. In this regard, this station is typical of most urban
locations.
Diurnal profiles of O3 concentrations can vary from day to day at a specific site,
however, because of changes in the various factors that influence concentrations. Composite
diurnal data (that is, concentrations for each hour of the day averaged over multiple days or
months) often differ markedly from the diurnal cycle shown by concentrations for a specific
day. In Figures 4-15 through 4-17 (reproduced from the previous document), diurnal data for
2 consecutive days are compared with composite diurnal data (1-mo averages of hour-by-hour
measurements) at three different kinds of sites: (1) center city-commercial (Washington, DC),
(2) rural-near urban (St. Louis, MO), and (3) suburban-residential (Alton, IL). Several
obvious points of interest present themselves in these figures: at some sites, at least, peaks
can occur at virtually any hour of the day or night, but these peaks may not show up strongly
in the longer term average data; some sites may be exposed to multiple peaks during a 24-h
period; and disparities, some of them large, can exist between peaks (the diurnal data) and the
1-mo average (the composite diurnal data) of hourly O3 concentrations.
4-48
-------
(D
c
o
N
O
50
40
30
60
50
40
30
20
10
0
80
70
60
50
40
30
20
10
0
(b) '
WFM, NY
(d)
IN(R)
PA
OH-
8 12 16
Time of Day (h)
20
24
Figure 4-11. Diurnal behavior of ozone at rural sites in the United States in July. Sites
are identified by the state in which they are located, (a) Western National
Air Pollution Background Network (NAPBN); (b) Whiteface Mountain
(WFM) located at 1.5 km above sea level; (c) eastern NAPBN sites; and
(d) sites selected from the Electric Power Research Institute's Sulfate
Regional Air Quality Study. IN(R) refers to Rockport.
Source: Logan (1989).
4-49
-------
10 20 30 40 50 60 70 80 90 100
Percent
Figure 4-12. Percent of time hourly average concentrations >0.1 ppm occurred between
0900 and 1559 hours in comparison to 24-h period for all rural agricultural
and forested sites with 3-mo SUM06 >26.4 ppm-h.
10 20 30 40 50 60 70 80 90 100
Figure 4-13. Percent of time hourly average concentrations >0.1 ppm occurred between
0900 and 1559 hours in comparison to 24-h period for all non-California
rural agricultural and forested sites with 3-mo SUM06 >26.4 ppm-h.
4-50
-------
Q.
a.
.o
2
8
8
O
O
N
O
2A 1 2 3 4
567
a.m.
8 9 10 11 12131415 16 17 18 1920 21 2223
^ p.m. s|
Hour
Figure 4-14. Diurnal pattern of 1-h ozone concentrations on July 13, 1979, Philadelphia,
PA.
When diurnal or short-term composite diurnal O3 concentrations are compared with
longer term composite diurnal O3 concentrations, the peaks are smoothed as the averaging
period is lengthened. Figure 4-1 demonstrates the effects of lengthening the period of time
over which values are averaged. This figure shows a composite diurnal pattern calculated on
the basis of 3 mo. Although seasonal differences are observed, the comparison of 3-mo
(Figure 4-18) with 1-mo composite diurnal concentrations (Figure 4-17) at the Alton, IL, site
readily demonstrates the smoothing out of peak concentrations as the averaging period is
lengthened. As indicated in the previous version of the document (U.S. Environmental
Protection Agency, 1986), although this is an obvious and familiar result in the statistical
treatment of monitoring data, it is highly pertinent to the protection of human health and
welfare from the effects of O3.
4.4.3 Nonurban Area Diurnal Patterns
Nonurban areas only marginally affected by transported O3 usually have a flatter
diurnal profile than sites located in urban areas. Nonurban O3 monitoring sites experience
differing types of diurnal patterns (Bohm et al., 1991; Lefohn, 1992b). As indicated earlier,
O3 concentrations at a specific location are influenced by local emissions and by long-range
transport from both natural and anthropogenic sources. Thus, considerable variation of
O3 exposures among sites characterized as agricultural or forested is found and there is no
preference for maximum diurnal patterns to occur in either the second or third quarter.
4-51
-------
Q.
a.
c
o
O
O
o
I
0.12
0.10
0.08
0.06
0.04
I I I I I I I I I
Center City - Commercial
\
I I I I I I I
July 8-9
1 mo
24 4 8 12 16 20 24 4 8 12 16 20 24
|* - a.m. — Noon — p.m.
a.m. — Noon - p.m. — \
Figure 4-15. Diurnal and 1-mo composite diurnal variations in ozone concentrations,
Washington, DC, July 1981.
Source: U.S. Environmental Protection Agency (1986).
The diurnal patterns for several agricultural sites have been characterized (U.S.
Environmental Protection Agency, 1986). Figures 4-19 and 4-20 show some typical patterns
of exposure. As discussed by U.S. Environmental Protection Agency (1986), the six sites,
whose diurnal patterns are illustrated in Figure 4-18, represent counties with high soybean,
wheat, or hay production. The figures show a distinct afternoon maximum with the lowest
concentrations occurring in the early morning and evening hours. Quarterly composite diurnal
patterns clearly show the division of the afternoon O3 concentrations into two seasonal
patterns, the low, "winter" levels in the first and fourth quarters and the high, "summer" levels
in the second and third quarters of the year.
Remote forested sites experience unique patterns of O3 concentrations (Evans et al.,
1983; Lefohn, 1984). These sites tend to experience a weak diurnal pattern, with
4-52
-------
0.12
0.10
I I I I I I I I I
Rural - Near Urban
I I I I I
Sept. 29-30
1 mo
TT
4 8 12 16 20 24
|c - a.m. — Noon — p.m.
a.m. —Noon
Figure 4-16. Diurnal and 1-mo composite diurnal variations in ozone concentrations,
St. Louis County, MO, September 1981.
Source: U.S. Environmental Protection Agency (1986).
hourly average O3 concentrations that occur frequently in the range of 0.04 to 0.05 ppm.
Figure 4-21 shows diurnal patterns for several sites in the NDDN network that are located in
forested areas. Several of the NDDN sites analyzed by Edgerton and Lavery (1992) exhibit
fairly flat average diurnal patterns. Such a pattern is based on average concentrations
calculated over an extended period. On a daily basis, some variation in O3 concentration does
occur from hour to hour, and, in some cases, high hourly average concentrations are
experienced either during daytime or nighttime periods (Lefohn and Mohnen, 1986; Lefohn
and Jones, 1986; Logan, 1989; Lefohn et al., 1990b; Taylor et al., 1992).
Lefohn et al. (1990b) characterized O3 concentrations at high-elevation monitoring
sites. The authors reported that a fairly flat diurnal pattern for the Whiteface Mountain
summit site (WF1) was observed (Figure 4-22a), with the maximum hourly average
concentrations occurring in the late evening or early morning hours. A similar pattern was
observed for the mid-elevation site at Whiteface Mountain (WF3). The site at the base of
Whiteface Mountain (WF4) showed the typical diurnal pattern expected from sites that
4-53
-------
0.12
0.10
& 0.08
c
.o
I 0.06
I
S 0.04
s
o
0.02
~~i r~ir~ir~ir
Suburban - Residential
i i
1 I T
1 T
Oct. 11-12
1 mo
\
ii
\
8 12 16 20 24 4 8 12 16 20 24
a.m. — Noon
p.m.
a.m.— Noon p.m.—*|
Hour of Day
Figure 4-17. Diurnal and 1-mo composite diurnal variations in ozone concentrations,
Alton, IL, October 1981 (fourth quarter).
Source: U.S. Environmental Protection Agency (1986).
experience some degree of O3 scavenging. More variation in the diurnal pattern for the
highest Shenandoah National Park sites occurred than for the higher elevation Whiteface
Mountain sites, with the typical variation for urban-influenced sites in diurnal pattern at the
lower elevation Shenandoah National Park site (Figure 4-22b). Aneja and Li (1992), in their
analysis of the five high-elevation Mountain Cloud Chemistry Program (MCCP) sites (see
Section 4.6.2 for site descriptions), noted the flat diurnal pattern typical of high-elevation sites
that has been described previously in the literature. Aneja and Li (1992) noted that the peak
of the diurnal patterns over the period May to October (1986 to 1988) for the five sites
occurred between 1800 and 2400 hours, whereas the minimum was observed between
0900 and 1200 hours. However, it is important to note that, as indicated by Lefohn et al.
(1990b), the flat diurnal pattern is not observed for all high-elevation sites.
4-54
-------
0.10
0.09
0.08
ET 0.07
a.
f 0.06
.o
I 0.05
§ 0.04
o
a> 0.03
c
<§ 0.02
0.01
Alton, IL
Quarter
- 4th Quarterx
24
6 8 10 12 14 16 18 20 22 24
a.m.
Noon
p.m.
Hour of Day
Figure 4-18. Composite diurnal patterns of ozone concentrations by quarter, Alton, IL,
1981.
4.5 Seasonal Patterns in Ozone Concentrations
4.5.1 Urban Area Seasonal Patterns
Seasonal variations in O3 concentrations in 1981 were described by the U.S.
Environmental Protection Agency (1986). The current form of the standard focuses on the
highest hourly average concentrations. The description that follows uses the highest hourly
average concentration as an indication of exposure. Figure 4-23 shows the 1-mo averages and
the single 1-h maximum concentrations within the month for eight sites across the nation.
The data from most of these sites exhibit the expected pattern of high O3 in late spring or in
summer and low levels in the winter. Data from Pomona, CA (Figure 4-23c), and Denver,
CO (Figure 4-23d), show summer maxima. Tampa, FL, shows a late spring maximum but
with concentrations in the fall (i.e., October) approaching those of spring (June)
(Figure 4-23f). Dallas data also tend to be skewed toward higher spring concentrations; but
note that November concentrations are also relatively high (Figure 4-23h). Because of
seasonal humidity and storm tracks from year to year, the general weather conditions in a
4-55
-------
0.10
•§- 0.09
& 0.08
O °'07
'JO 0.06
** 0.05
0.04
0.03
0.02
0.01
- N. Little Rock, AR
1 st Quarter
2nd Quarter
3rd Quarter
4th Quarter
8
&
I
I I I I I I I I I I I I I I I I I I I I I I I
I I I I I I I I I I I I I I I I I I I I I I I
'24 2 4 6 8
a.m.-
10 12 14 16 18 20 22 24
0.10
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
~ Bakersfield, CA /*N
I I I I I I I I I I I I I I I I I I I I I I I
24 2 4 6 8 10 12 14 16 18 20 22 24
l< a.m. 5\£ p.m. >|
••
j£
I
'jB
I
1
O
|
O
0.10
0.09
0.08
°-°7
0.06
0.05
°'04
0.03
0.02
I I I I I I I I I I I I I I I I I I I I I I I
• (c)
- Sacramento, CA
24
2 4 6 8 10 12 14 16 18 20 22 24
a.m. - ^K - p.m. - 5\
0.10
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
(d)
- Alton, IL
I I I I I I I I I I I I I I I I I I I I I I I
...--
'"'•*.
I I I I I I I I I I I I I I I I I I I I I
24 2 4 6 8 10 12 14 16 18 20 22 24
a.m. >K p.m. ^|
E
a.
&
8
0.10
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
- Clark Co., OH
I I I I I I I I I I I I I I I I I I l l l l l
" (e)
l l l l l l l l l l l l l l l l l l l l l l l
24 2 4 6 8 10 12 14 16 18 20 22 24
K a.m. >K p.m. >|
Hour
0.10
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
I I I I I I I I I I I I I I I I I I I I I I I
I I I I I I I I I I I I I I I I I I I I I I I
24 2 4 6 8 10 12 14 16 18 20 22 24
• a.m.-
p.m.-
Hour
Figure 4-19. Quarterly composite diurnal patterns of ozone concentrations at selected sites
representing potential for exposure of major crops, 1981.
Source: U.S. Environmental Protection Agency (1986).
4-56
-------
0.06
Q.
O.
O °'04
?
I
^0.02
o
N
O
24
a.m.
12
p.m.
24
Figure 4-20. Composite diurnal ozone pattern at a rural National Crop Loss Assessment
Network site in Argonne, IL, August 6 through September 30, 1980.
Source: U.S. Environmental Protection Agency (1986).
given year may be more favorable for the formation of O3 and other oxidants than during the
prior or following year. For example, 1988 was a hot and dry year during which some of the
highest O3 concentrations of the last decade occurred, whereas 1989 was a cold and wet year
in which some of the lowest concentrations occurred (U.S. Environmental Protection Agency,
1992a).
4.5.2 Nonurban Area Seasonal Patterns
In the literature, several investigators have reported on the tendency for average
O3 concentrations to be higher in the second versus the third quarter of the year for many
isolated rural sites (Evans et al., 1983; Singh et al., 1978). This observation has been
attributed either to stratospheric intrusions or to an increasing frequency of slow-moving,
high-pressure systems that promote the formation of O3. Lefohn et al. (1990a) reported that
for several clean sites, the highest values of exposure indices occurred in the third quarter
rather than in the second. The results of this analysis will be discussed in the Section 4.5.3.
Taylor et al. (1992) reported that for 10 forest sites in North America, the temporal patterns
of O3 on quarterly or annual periods exhibited less definitive patterns. Based on the exposure
index selected, different patterns were reported. The different patterns may be
4-57
-------
60
50 4
40
30
O 20-
o.
a.
••• 108-Prince Edward, VA
ooo 119-Cedar Creek, WV
9 o o
O
o
c r • * •
- •"..•
• o
o
o
10
o
o
o o
o o
o o
' 2 ' 4 ' 6 8 10 12 14 16 18 20 22 24
Hour
a.
a.
60
50
40
30-
<§ 20
10-
168- Glacier NP, MT
ooo 174 - Grand Canyon, AZ
O O O O O
oooooo ooooooooo oo
2 4 6 8 10 12 14 16 18 20 22 24
Hour
Figure 4-21. Composite diurnal ozone pattern at selected National Dry Deposition Network
sites.
Source: Edgerton and Lavery (1992).
4-58
-------
6 ' 2 '4 ' 6 '8 ' 10 ' 12 ' 14 ' 16 ' 18 ' ^0 ' & ' 214
0.06-
0.05
0.04-
o.
a.
c
o
N
0.03-
0.02-
0.01-
0.00
024
8 10 12 14 16 18 20 22 24
Hour
Figure 4-22. Composite diurnal pattern at (a) Whiteface Mountain, NY, and the
(b) Mountain Cloud Chemistry Program Shenandoah National Park site for
May to September 1987.
Source: Lefohn et al. (1990b).
4-59
-------
I
1
Ǥ
0.35
0.30
0.25
0.20
0.15
0.10
0.05
_ Tucson, AZ
' Q Maximum
H Average
JFMAMJJASOND
S
.1
o
o
1
8
o
I
i
u.oo
0.30
0.25
0.20
0.15
0.10
0.05
0
_ Pomona, C
-
n
J
^
In
I
-
_
-
-
A
-
"
-
-
A M J J A
-
-
-
-
l_l
S 6 N
"I-
-
D
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0
0.35
_ Washington, DC
d
JFMAMJJASOND
I
0.30-
0.25
0.20
0.15-
0.10-
0.05
Morris Co.,
NJ
JFMAMJJASOND
Month of Year
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0
(b)
_ Lennox, CA
JFMAMJJASOND
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0
(d)
_ Denver, CO
_ Tampa, FL
JF M A M J J ASOND
0.35
0.3C
0.2S-
0.2C-
0.1 E -
0.1C
0
0
h)
_ Dallas, TX
Month of Year
Figure 4-23. Seasonal variations in ozone concentrations as indicated by monthly averages
and the 1-h maximum in each month at selected sites, 1981.
Source: U.S. Environmental Protection Agency (1986).
4-60
-------
associated with the observations by Logan (1989) that rural O3 in the eastern United States in
the spring and summer is severely impacted by anthropogenic and possibly natural emissions
of NOX and hydrocarbons, and that O3 episodes occur when the weather is particularly
conducive to photochemical formation of O3. Meagher et al. (1987) reported for rural
O3 sites in the southeastern United States that the daily maximum 1-h average concentration
was found to peak during the summer months. Taylor and Norby (1985) reported that, in the
Shenandoah National Park, the probability of a day occurring in which a 1-h mean
O3 concentration was >0.08 ppm was the same during the months of May, June, and July,
whereas the probability was nearly 40% less in August. The probability of an episode during
each of the remaining months of the growing season was <5%. The month of July
experienced both the highest frequency of episodes and the highest mean duration of exposure
events.
Aneja and Li (1992) reported that the maximum monthly ozone levels occurred in
either the spring or the summer (May to August), and the minimum occurred in the fall
(September and October). The timing of the maximum monthly values differed across sites
and years. However, in 1988, an exceptionally high O3 concentration year, for almost all of
the five sites, June was the month in which the highest monthly average concentration
occurred. This was the month in which the greatest number of O3 episodes occurred in the
eastern United States.
4.5.3 Seasonal Pattern Comparisons with Sites Experiencing Low
Exposures
Lefohn et al. (1990a) have characterized the O3 concentrations that occurred at
several sites in the United States that experience low maximum hourly average
concentrations. The Theodore Roosevelt National Park site experienced its maximum in July
for 1984 and 1985 and in May for 1986. Of the three western national forest sites evaluated
by Lefohn et al. (1990a), only Apache National Forest experienced its maximum monthly
mean concentration in the spring. The Apache National Forest site was above mean nocturnal
inversion height, and no decrease of concentrations occurred during the evening hours. This
site also experienced the highest hourly maximum concentration, as well as the highest W126
O3 exposures. The Custer and Ochoco national forest sites experienced most of their
maximum monthly mean concentrations in the summer. The White River Oil Shale site in
Colorado experienced its maximum monthly mean during the spring and summer months.
The W126 sigmoidal weighting function index was also used to identify the month
of highest O3 exposure. A somewhat more variable pattern was observed than when the
maximum monthly average concentration was used. For some sites, the winter/spring pattern
was represented; for others, it was not. In some cases, the highest W126 exposures occurred
earlier in the year than was indicated by the maximum monthly concentration. For example,
in 1979, the Custer National Forest site experienced its highest W126 exposure in April,
although the maximum monthly mean occurred in August. In 1980, the reverse occurred.
There was no consistent pattern for those sites located in the continental United
States. The Theodore Roosevelt and Ochoco national park sites, the Custer National Forest
site, and the White River Oil Shale site experienced their maximum O3 exposures during the
spring and summer months. The sites experiencing their highest O3 exposures in the
fall-to-spring period did not necessarily experience the lowest O3 exposures.
4-61
-------
4.6 Spatial Variations in Ozone Concentrations
4.6.1 Urban-Nonurban Area Concentration Differences
Diurnal concentration data presented earlier indicate that peak O3 concentrations
can occur later in the day in rural areas than in urban, with the distances downwind from
urban centers generally determining how much later the peaks occur. Meagher et al. (1987)
reported that for five rural sites in the Tennessee Valley region of the southeastern United
States, O3 levels were found to equal or exceed urban values for the same region. Data
presented in the 1978 criteria document demonstrated that peak concentrations of O3 in rural
areas generally are lower than those in urban areas, but that average concentrations in rural
areas are comparable to or even higher than those in urban areas (U.S. Environmental
Protection Agency, 1978). Reagan (1984) noted that O3 concentrations measured near
population-oriented areas were depressed in comparison with data collected in more isolated
areas. As noted earlier, urban O3 values are often depressed because of titration by NO
(Stasiuk and Coffey, 1974). In reviewing the NCLAN's use of kriging to estimate the 7-h
seasonal average O3 levels, Lefohn et al. (1987a) found that the 7-h values derived from
kriging for sites located in rural areas tended to be lower than the actual values because of the
effect of using data from urban areas to estimate rural values. In addition to the occurrence
of higher average concentrations and occasionally higher peak concentrations of O3 in
nonurban than in urban areas, it is well documented that O3 persists longer in nonurban than
in urban areas (Coffey et al., 1977; Wolff et al., 1977; Isaksen et al., 1978). The absence of
chemical scavengers appears to be the main reason.
4.6.2 Concentrations Experienced at High-Elevation Sites
The distributions of hourly average concentrations experienced at high-elevation
cities are similar to those experienced in low-elevation cities. For example, the distribution of
hourly average concentrations for several O3 sites located in Denver were similar to
distributions observed at many low-elevation sites in the United States. However, as will be
discussed in Section 4.6.3, for assessing the possible impacts of O3 at high-elevation sites, the
use of absolute concentrations (e.g., in units of micrograms per cubic meter) instead of
mixing ratios (e.g., parts per million) may be an important consideration.
Lefohn et al. (1990b) summarized the characterization of gaseous exposures at
rural sites in 1986 and 1987 at several MCCP high-elevation sites. Aneja and Li (1992) have
summarized the ozone concentrations for 1986 to 1988. Table 4-14 summarizes the sites
characterized by Lefohn et al. (1990b). Table 4-15 summarizes the concentrations and
exposures that occurred at several of the sites for the period 1987 to 1988. In 1987, the
7- and 12-h seasonal means were similar at the Whiteface Mountain WF1 and WF3 sites
(Figure 4-24a). The 7-h mean values were 0.0449 and 0.0444 ppm, respectively, and the 12-h
mean values were 0.0454 and 0.0444 ppm, respectively. Note that, in some cases, the 12-h
mean was slightly higher than the 7-h mean value. This resulted when the 7-h mean period
(0900 to 1559 hours) did not capture the period of the day when the highest hourly mean
O3 concentrations were experienced. A similar observation was made, using the 1987
4-62
-------
Table 4-14. Description of Mountain Cloud Chemistry
Program Sites3
Site Elevation (m)
Rowland Forest (HF1), ME
Mt. Moosilauke (MSI), NH
Whiteface Mountain (WF1), NY
Shenandoah NP (SHI), VA
Shenandoah NP (SH2), VA
Shenandoah NP (SH3), VA
Whitetop Mountain (WT1), VA
Mt. Mitchell (MM1), NC
Mt. Mitchell (MM2), NC
65
1,000
1,483
1,015
716
524
1,689
2,006
1,760
Latitude
45°
43°
44°
38°
38°
38°
36°
35°
35°
11'
59'
23'
37'
37'
37'
38'
44'
45'
18"
26"
12"
30"
45"
20"
15"
68°
71°
73°
78°
78°
78°
81°
82°
82°
Longitude
46'
48'
51'
20'
21'
21'
36'
17'
15'
28"
34"
48"
13"
28"
21"
15"
aSee Appendix A for abbreviations and acronyms.
data, for the MCCP Shenandoah National Park sites. The 7- and 12-h seasonal means were
similar for the SHI and SH2 sites (Figure 4-24b). Based on cumulative indices, the
Whiteface Mountain summit (1,483-m) site (WF1) experienced a higher exposure than the
WF3 (1,026-m) site (Figure 4-24c). Both the sum of the concentrations >0.07 ppm (SUM07)
and the number of hourly concentrations >0.07 ppm were higher at the WF1 site than at the
WF3 site. The site at the base of the mountain (WF4) experienced the lowest exposure of
the three O3 sites. Among the MCCP Shenandoah National Park sites, the SH2 site
experienced marginally higher O3 exposures, based on the index that sums all of the hourly
average concentrations (i.e., referred to as "total dose" in the figure) and sigmoidal values,
than the high-elevation site (SHI; Figure 4-24d). The reverse was true for the sums of the
concentrations >0.07 ppm and the number of hourly concentrations >0.07 ppm.
When the Big Meadows, Dickey Ridge, and Sawmill Run, Shenandoah National
Park, data for 1983 to 1987 were compared, it again was found that the 7- and 12-h seasonal
means were insensitive to the different O3 exposure patterns. A better resolution of the
differences was observed when the cumulative indices were used (Figure 4-25). There was
no evidence that the highest elevation, Big Meadows, site consistently had experienced higher
O3 exposures than the other sites. In 2 of the 5 years, the highest elevation site experienced
lower exposures than the Dickey Ridge and Sawmill Run sites, based on the sum of all
concentration or sigmoidal indices. For 4 of the 5 years, the SUM07 index yielded the same
result.
Taylor et al. (1992) indicate that the forests they monitored experienced differences
in O3 exposure. The principal spatial factors underlying this variation were elevation,
proximity to anthropogenic sources of oxidant precursors, regional-scale meteorological
conditions, and airshed dynamics between the lower free troposphere and the surface
boundary layer. Table 4-16 summarizes the exposure values for the 10 EPRI Integrated
Forest Study sites located in North America.
4-63
-------
Table 4-15. Seasonal
for
(April to October) Percentiles, SUM06, SUM08, and W126 Values
the Mountain Cloud Chemistry Program Sites3
Site
Rowland Forest, ME
(HF1)
Mt. Moosilauke, NH
(MSI)
Whiteface Mountain, NY
(WF1) (36-03 1-0002)
Whiteface Mountain, NY
(WF3)
Whiteface Mountain, NY
(WF4)
Mt. Mitchell, NC
(MM1)
Mt. Mitchell, NC
(MM2)
Shenandoah Park, VA
(SHI)
Shenandoah Park, VA
(SH2)
Shenandoah Park, VA
(SH3)
Whitetop Mountain, VA
(WT1)
Elev.
(m)
65
1,000
1,483
1,026
604
2,006
1,760
1,015
716
524
1,689
Year
1987
1988
1987
1988
1987
1988
1987
1987
1987
1988
1989
1992
1987
1988
1987
1988
1987b
1987
1988
1987
1988
Min.
0.000
0.000
0.006
0.010
0.011
0.014
0.010
0.000
0.008
0.011
0.010
0.005
0.017
0.009
0.000
0.006
0.003
0.000
0.006
0.011
0.000
10
0.013
0.012
0.027
0.026
0.029
0.025
0.025
0.011
0.034
0.038
0.038
0.036
0.032
0.029
0.023
0.024
0.027
0.018
0.020
0.038
0.030
30
0.021
0.021
0.036
0.033
0.037
0.033
0.033
0.023
0.044
0.054
0.047
0.043
0.042
0.041
0.036
0.036
0.040
0.029
0.031
0.051
0.046
50
0.028
0.028
0.045
0.043
0.046
0.043
0.039
0.031
0.051
0.065
0.054
0.048
0.049
0.050
0.044
0.047
0.049
0.037
0.040
0.059
0.058
70
0.035
0.036
0.053
0.055
0.053
0.056
0.047
0.041
0.058
0.075
0.059
0.053
0.056
0.060
0.054
0.058
0.059
0.047
0.051
0.066
0.068
90
0.046
0.047
0.065
0.076
0.067
0.078
0.064
0.056
0.067
0.095
0.068
0.063
0.067
0.080
0.069
0.077
0.071
0.061
0.067
0.078
0.084
95
0.052
0.054
0.074
0.087
0.074
0.089
0.075
0.065
0.074
0.106
0.072
0.069
0.073
0.092
0.076
0.087
0.077
0.068
0.076
0.085
0.094
99
0.065
0.076
0.086
0.113
0.087
0.110
0.091
0.081
0.085
0.126
0.081
0.081
0.083
0.110
0.085
0.103
0.086
0.080
0.097
0.096
0.119
Max
0.076
0.106
0.102
0.127
0.104
0.135
0.117
0.117
0.105
0.145
0.147
0.096
0.096
0.162
0.135
0.140
0.145
0.108
0.135
0.111
0.163
No. Obs.
4,766
4,786
4,077
2,835
4,704
4,673
4,755
4,463
3,539
2,989
2,788
3,971
3,118
2,992
3,636
3,959
2,908
3,030
4,278
4,326
3,788
SUM06
5.9
10.9
45.0
51.9
62.0
65.8
45.4
23.8
59.4
145.1
54.8
37.8
47.0
68.7
54.2
80.9
55.7
23.1
52.3
147.7
133.8
SUM08
0.0
2.9
9.5
21.2
12.2
40.8
14.4
5.1
7.8
69.7
3.5
4.4
5.1
28.1
8.5
29.6
7.8
55.8
2.6
15.6
32.4
51.0
W126
7.7
11.6
40.1
43.4
49.5
56.5
40.3
21.3
46.5
116.6
40.7
36.7
37.4
57.7
42.0
67.2
41.8
19.2
44.2
105.7
102.8
"See Appendix A for abbreviations and acronyms.
""Calculations based on a May to September season.
-------
i 7h
I 12h
0.00
WF1
WF3
WF4
SH1
SH2
SH3
2001
• Total Dose
Sigmoidal
D Sum 2 0.07
200-1 (d)
100-
I
100-
• Total Dose
• Sigmoidal
D Sum ^ 0.07
WF1
WF3
WF4
SH1
SH2
SH3
Figure 4-24. Seven- and 12-h means at (a) Whiteface Mountain and (b) Shenandoah
National Park for May to September 1987 and integrated exposures at
(c) Whiteface Mountain and (d) Shenandoah National Park for May to
September 1987.
Source: Lefohn et al. (1990b).
4.6.3 Other Spatial Variations in Ozone Concentrations
Despite relative intraregional homogeneity, evidence exists for intracity variations
in concentrations that are pertinent to potential exposures of human populations and to the
assessment of actual exposures sustained in epidemiologic studies. Two illustrative pieces of
data are presented in this section: (1) a case of relative homogeneity in a city with a
population under 500,000 (New Haven, CT) and (2) a case of relative inhomogeneity of
concentrations in a city of greater than 9 million population (New York City).
As described in the previous version of the criteria document (U.S. Environmental
Protection Agency, 1986), the percentiles of the hourly average concentrations for a
New Haven site and two other monitoring stations that were operating at the time in the same
county, one in Derby, 9 mi west of New Haven, and one in Hamden, 6 mi north of
New Haven, generally are similar. Table 4-17 shows the monitoirng data and time of the
maximum hourly concentrations by quarter at these three sites.
4-65
-------
300-1
May - September 1983
(a) • Total Dose
Q Sigmoidal
Sum > 0.07
SOO-i
£ 200-
Q.
&
0>
o 100
O
Big Meadows Dickey Ridge Sawmill Run
June - September 1985
(c) • Total Dose
D Sigmoidal
n Sum > 0.07
LI.L
Big Meadows Dickey Ridge Sawmill Run
May - September 1984
300-1
200-
(b)
Total Dose
Sigmoidal
Sum ;> 0.07
300-i
May- September 1986
(d)
Total Dose
Sigmoidal
Sum i 0.07
Big Meadows Dickey Ridge Sawmill Run
Big Meadows Dickey Ridge Sawmill Run
300 n
May - September 1987
(e)
Total Dose
Sigmoidal
Sum a 0.07
Big Meadows Dickey Ridge Sawmill Run
Figure 4-25. Integrated exposures for three non-Mountain Cloud Chemistry Program
Shenandoah National Park sites, 1983 to 1987.
Source: Lefohn et al. (1990b).
4-66
-------
Table 4-16. Summary Statistics for 11 Integrated Forest Study Sites3
Site
HIGH ELEVATION SITES
Whiteface Mountain, NY
Great Smoky Mountain NP
Coweeta Hydrologic Lab, NC
LOW ELEVATION SITES
Huntington Forest, NY
Howland, MA
Oak Ridge, TN
Year
1987
1987
1988
1988
1987
1987
1988
1988
1987
1987
1988
1988
1987
1987
1988
1988
1987
1987
1988
1988
1987
1987
1988
1988
Quarter
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
24-h
(ppb)
42
45
49
44
54
53
71
59
50
47
61
57
36
24
40
37
34
26
36
24
42
29
40
32
12-h
(ppb)
43
44
50
43
52
51
70
57
48
44
59
54
42
32
46
46
39
32
41
30
53
44
57
47
7-h
(ppb)
42
43
49
43
49
49
68
55
47
42
59
51
42
33
46
48
39
31
41
30
50
41
58
51
1-h Max
(ppb)
104
114
131
119
99
95
119
120
85
95
104
100
88
76
106
91
69
76
90
71
112
105
104
122
SUM06
(ppm-h)
13.2
30.1
33.5
22.6
57.1
34.3
126.3
74.7
32.4
24.1
81.6
63.6
9.8
5.4
19.2
18.6
1.9
3.8
8.1
1.7
39.5
24.3
26.4
19.7
SUM08
(ppm-h)
2.5
11.8
13.9
10.4
10.9
8.8
61.2
22.2
2.6
2.4
18.5
19.8
0.9
0.2
6.1
2.7
0.0
0.0
2.9
0.0
13.5
9.0
9.8
7.7
-------
Table 4-16 (cont'd). Summary Statistics for 11 Integrated Forest Study Sites3
03
Site
LOW ELEVATION SITES (cont'd)
Thompson Forest, WA
B.F. Grant Forest, GA
Gainesville, FL
Duke Forest, NC
Nordmoen, Norway
Year
1987
1987
1988
1988
1987
1987
1988
1988
1987
1987
1988
1988
1987
1987
1988
1988
1987
1987
1988
1988
Quarter
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
24-h
(ppb)
36
30
32
32
32
33
47
32
42
29
35
20
38
52
54
38
32
14
22
11
12-h
(ppb)
43
36
39
39
46
52
63
47
53
44
48
29
48
59
69
51
40
18
28
15
7-h
(ppb)
41
34
37
36
48
54
64
48
50
41
51
30
52
50
75
54
41
20
29
16
1-hMax
(ppb)
103
94
103
140
99
102
127
116
b
b
84
70
100
124
115
141
75
32
53
30
SUM06
(ppm-h)
10.7
10.3
8.1
13.5
26.1
31.3
53.1
24.1
b
b
23.4
1.9
29.2
b
b
52.9
2.4
0.0
0.0
0.0
SUM08
(ppm-h)
3.6
2.1
2.3
6.7
5.1
10.3
21.9
7.4
b
b
0.5
0.1
7.8
b
b
23.4
0.0
0.0
0.0
0.0
aSee Appendix A for abbreviations and acronyms.
bData were insufficient to calculate statistic.
Source: Adapted from Taylor et al. (1992).
-------
Table 4-17. Quarterly Maximum One-Hour Ozone Values at Sites in
and Around New Haven, Connecticut, 1976
(Chemiluminescence Method, Hourly Values in ppm)
Quarter of Year
New Haven, CT
No. measurements
Max 1-h, ppm
Time of day
Date
Derby, CT
No. measurements
Max 1-h, ppm
Time of day
Date
Hamden, CT
No. measurements
Max 1-h, ppm
Time of day
Date
10
0.045
1100 hours
March 29
11
0.015
2300 hours
March 31
56
0.050
2400 hours
March 29
1,964
0.274
1400 hours
June 24
2,140
0.280
1400 hours
June 24
2,065
0.240
1500 hours
June 24
2,079
0.235
1400 hours
August 12
2,187
0.290
1400 hours
August 12
1,446
0.240
1300 hours
July 20
66
0.066
1000 hours
October 3
1,360
0.060
1900 hours
December 20
286
0.065
1500 hours
October 7
Source: U.S. Environmental Protection Agency (1986).
The source of much of the O3 experienced in the New Haven area is the greater
New York area (e.g., Wolff et al., 1975; Cleveland et al., 1976a,b). An urban plume
transported over the distance from New York City to New Haven would tend to be relatively
well-mixed and uniform, such that intracity variations in New Haven probably would be
minimal.
As indicated in the previous version of the criteria document (U.S. Environmental
Protection Agency, 1986), intracity differences in O3 concentrations also have been reported
by Kelly et al. (1986) for a 1981 study in Detroit, MI. Ozone concentrations were measured
for about 3 mo at 16 sites in the metropolitan Detroit area and in nearby Ontario, Canada.
Values at 15 sites were correlated with those at a site adjacent to the Detroit Science Center,
about 3 km north of the central business district in Detroit. In general, the correlation
decreased as distance from the Science Center site increased; and, in general, the actual
concentrations increased with distance from that site toward the north-northeast. The highest
O3 concentrations were recorded at sites about 10 to 70 km north-northeast of the urban core.
At greater distances or in other directions, O3 maxima decreased.
Chicago is an example where O3 concentrations increase as the distance from the
inner city increases. Figure 4-26 shows, for a number of O3 monitoring sites, the number of
days in 1991 that the maximum hourly average concentration was greater than 0.1 ppm. The
greatest number of exceedances of the daily maximum 1-h concentration of 0.1 ppm was to
the north of the city.
4-69
-------
Chicago, Illinois
Number of Days >0.10
1991
Days > 0.10
Figure 4-26. Number of days in 1991 for which the maximum hourly average
ozone concentration was greater than 0.1 ppm at Chicago, IL.
4-70
-------
Concentrations of O3 vary with altitude and with latitude. Although a number of
reports contain data on O3 concentrations at high altitudes (e.g., Coffey et al., 1977; Reiter,
1977b; Singh et al., 1977; Evans et al., 1985; Lefohn and Jones, 1986), fewer reports are
available that present data for different elevations in the same locality. There appears to be
no consistent conclusion concerning the relationship between O3 exposure and elevation.
Wolff et al. (1987) reported, for a short-term study at High Point Mountain in
northwestern New Jersey, that both the daily maximum and midday O3 concentrations were
similar at different altitudes, but that the O3 exposures increased with elevation. Wolff et al.
(1987) conducted a study of the effects of altitude on O3 concentrations at three sites located
at three separate elevations on High Point Mountain in northwestern New Jersey. Data for
several days indicate that in mid-July, when atmospheric mixing was good, vertical profiles
were nearly constant, with concentrations increasing only slightly with elevation. Likewise,
the daily O3 maxima were similar at different elevations. At night, however,
O3 concentrations were nearly zero in the valley (i.e., the lowest-elevation site) and increased
with elevation. Comparison of the O3 exposures at the three sites (number of hours
>0.08 ppm) showed that greater cumulative exposures were sustained at the higher elevations.
Comparable data from an urban area (Bayonne) about 80 km southeast of High Point
Mountain showed that the cumulative exposures were higher at all three of the mountain sites
than in the urban area (Wolff et al., 1987). The investigators concluded from their
concentration and meteorological data that elevated, mountainous sites in the eastern United
States may be expected to be exposed to higher O3 concentrations than valley sites throughout
the year.
Winner et al. (1989) reported that, for three Shenandoah National Park sites (i.e.,
Big Meadows, Dickey Ridge, and Sawmill Run), the 24-h monthly mean O3 concentrations
tended to increase with elevation, but that the number of elevated hourly occurrences equal to
or above selected thresholds did not. The authors reported that the highest elevation site (Big
Meadows) experienced a smaller number of concentrations at or below the minimum
detectable level than did the other two sites. The larger number of hourly average
concentrations that occurred at or below the minimum detectable level at both Dickey Ridge
and Sawmill Run resulted in lower 24-h averages at these sites.
Lefohn et al. (1990b), characterizing the O3 exposures at several high-elevation
sites, reported that, based on cumulative indices, the Whiteface Mountain summit site (WF1)
experienced a slightly higher exposure than the lower elevation Whiteface Mountain (WF3)
site. The site at the base of Whiteface Mountain (WF4) experienced the lowest exposure of
the three O3 sites. Among the MCCP Shenandoah National Park sites, the SH2 site
experienced higher O3 exposures than the high-elevation site (SHI). The "total dose"
(correctly referred to as the sum of all hourly average concentrations) and sigmoidal (W126)
indices were slightly higher at the SH2 than at the SHI site. The data capture at the two sites
for the 5-mo period was similar. However, the sum of the concentrations >0.07 ppm and the
number of hourly concentrations >0.07 ppm were slightly higher at the SHI than at the SH2
site. For the Whiteface Mountain sites, both the sum of the concentrations >0.07 ppm
(SUM07) and the number of hourly concentrations >0.07 ppm were higher at the WF1 site
than at the WF3 site.
When the Big Meadows, Dickey Ridge, and Sawmill Run, Shenandoah National
Park, data for 1983 to 1987 were compared, a higher resolution of the differences among the
regimes was observed when the cumulative indices were used. No specific trend could be
identified that showed the highest elevation site, Big Meadows, had consistently experienced
4-71
-------
higher O3 exposures than the lower elevation sites. In 2 of the 5 years, the highest elevation
site experienced lower exposures than the Dickey Ridge and Sawmill Run sites, based on the
sum of all concentrations or sigmoidal indices. For 4 of the 5 years, the SUM07 index
yielded the same result.
An important issue for assessing possible impacts of O3 at high-elevation sites that
requires further attention is the use of mixing ratios (e.g., parts per million) instead of
absolute concentration (e.g., in units of micrograms per cubic meter) to describe
O3 concentration. In most cases, mixing ratios or mole fractions are used to describe
O3 concentrations. Lefohn et al. (1990b) have pointed out that the manner in which
concentration is reported may be important when assessing the potential impacts of air
pollution on high-elevation forests. Concentration varies as a function of altitude. Although
the change in concentration is small when the elevational difference between sea level and the
monitoring site is small, it becomes substantial at high-elevation sites. Given the same part-
per-million value experienced at both a high- and low-elevation site, the absolute
concentrations (i.e., micrograms per cubic meter) at the two elevations will be different.
Because both O3 and ambient air are gases, changes in pressure directly affect their volume.
According to Boyle's law, if the temperature of a gas is held constant, the volume occupied
by the gas varies inversely with the pressure (i.e., as pressure decreases, volume increases).
This pressure effect must be considered when measuring absolute pollutant concentrations. At
any given sampling location, normal atmospheric pressure variations have very little effect on
air pollutant measurements. However, when mass/volume units of concentration are used and
pollutant concentrations measured at significantly different altitudes are compared, pressure
(and, hence, volume) adjustments are necessary. In practice, the summit site at Whiteface
Mountain had a slightly higher O3 exposure than the two low-elevation sites (Lefohn et al.,
1991). However, at Shenandoah National Park sites, the higher elevation site experienced
lower exposures than lower elevation sites in some years.
These exposure considerations are trivial at low-elevation sites. However, when
one compares exposure-effects results obtained at high-elevation sites with those from low-
elevation sites, the differences may become significant (Lefohn et al., 1990b). In particular,
assuming that the sensitivity of the biological target is identical at both low and high
elevations, some adjustment will be necessary when attempting to link experimental data
obtained at low-elevation sites with air quality data monitored at the high-elevation stations.
4.7 Indoor Ozone Concentrations
Most people in the United States spend a large proportion of their time indoors.
A knowledge of actual exposures of populations to indoor levels of O3 is essential for the
interpretation and use of results associated with epidemiological studies. However, essentially
all routine air pollution monitoring is done on outdoor air. Until the early 1970s, very little
was known about the O3 concentrations experienced inside buildings. The ratio of the
indoor/outdoor (I/O) O3 concentrations is a parameter that has been widely used for studying
the indoor and outdoor relationships, sources, and exposure patterns of O3. However, the
database on this subject is not large, and a wide range of I/O O3 concentration relationships
can be found in the literature. The only significant source of O3 in indoor residential air is
infiltration of outdoor O3, with ventilation rates affecting the flow of air between indoor and
outdoor (Zhang and Lioy, 1994).
4-72
-------
Reported I/O values for O3 are highly variable (U.S. Environmental Protection
Agency, 1986) and range from <0.1 to 0.8 for various indoor environments and ventilation
rates (Weschler et al., 1989). Unfortunately, the number of experiments and kinds of
structures examined to date provide only limited data for use in modeling indoor exposures.
Data were summarized by Yocom (1982) describing studies of indoor-outdoor gradients in
buildings and residences for either O3 or photochemical oxidant. The results were highly
variable. A relatively large number of factors can affect the difference in O3 concentrations
between the inside of a structure and the outside air. In general, outside air infiltration or
exchange rates, interior air circulation rates, and interior surface composition (e.g., rugs,
draperies, furniture, walls) affect the balance between replenishment and decomposition of O3
within buildings (U.S. Environmental Protection Agency, 1986). Although indoor
concentrations of O3 will almost invariably be less than outdoors, the fact that people spend
more time indoors than outdoors may result in greater overall indoor exposures.
Cass et al. (1991) have discussed the importance of protecting works of art from
damage due to O3. Experiments show that the fading of artists' pigments in the presence of
O3 is directly related to the product of concentration times duration of exposure. Druzik et al.
(1990) reported that, in a survey of 11 museums, galleries, historical houses, and libraries in
Southern California, facilities with a high air exchange with the outdoors and no pollutant
removal system have indoor O3 concentrations more than two-thirds those of outdoor
concentrations. The author reported that museums with conventional air-conditioning systems
showed indoor O3 concentrations about 30 to 40% of those outside, whereas museums with
no forced ventilation system, where slow air infiltration provides the only means of air
exchange, have indoor O3 levels typically 10 to 20% of those outdoors. Several other studies
have been reported in the literature and Table 4-18 lists the I/O ratios reported from these
efforts as well as those from earlier years.
Automobiles and other vehicles constitute another indoor environment in which
people may spend appreciable amounts of time. As with buildings, the mode of ventilation
and cooling helps determine the inside concentrations. The U.S. Environmental Protection
Agency (1986) describes studies for the I/O ratios. In one study reported by Contant et al.
(1985), the I/O ratios from 49 measurements inside vehicles were 0.44 for the mean, 0.33 for
the median, and 0.56 for maximum concentrations measured. Chan et al. (1991) reported an
I/O ratio of 0.20 for median in-vehicle concentrations (0.011 ppm) and time-matched fixed-
site measurements (0.051 ppm).
At present, there are no long-term monitoring data on indoor air pollutant
concentrations comparable to the concentration data available for outdoor locations. Thus, for
estimates of the exposure of building or vehicle occupants to O3 and other photochemical
oxidants, it is necessary to rely on extrapolations of very limited I/O data.
4.8 Estimating Exposure to Ozone
4.8.1 Introduction
Human exposure represents the joint occurrence of an individual being located at
point (x,y,z) during time t, with the simultaneous presence of an air pollutant at concentration
Cx,y,z (0 (U.S. Environmental Protection Agency, 1991). Consequently, an individual's
exposure to an air pollutant is a function of location as well as time. If a volume at a
location can be defined such that air pollutant concentrations within it are homogeneous yet
4-73
-------
Table 4-18. Summary of Reported Indoor-Outdoor Ozone Ratios
Structure
Indoor-Outdoor Ratio
Reference
Hospital
Residence (with evaporative cooler)
Office
(air-conditioned; 100% outside
air intake)
(air-conditioned; 70% outside air
intake)
Office
Office/Lab
Residence
Residence
Two offices
Residence
(gas stoves)
(all electric)
Office
School room
Residence
Residences (1 each)
(air-conditioned)
(100% outside air; no
air-conditioning)
Residences (12)
(air-conditioned)
Residences (41)
Residences (6)
(window open)
(window closed)
(air-conditioning)
Art gallery
Art gallery
(three modes of ventilation in
each 24-h period: recirculation,
mixture of recirculated and
outside air, and 100% outside air)
Museums
0.67a
0.60a
0.80 ± 0.10
0.65 ± 0.10
0.66
0.54
0.62
0.70
0.50-0.70
0.30
0.19
0.20
0.29
0.19 (max concentration)
0.10-0.25
0.00-0.09
1.00
0.21 (mean concentration)
0.12 (med. concentration)
0.59 (max concentration)
0.30
0.59 ±0.16
0.26 ± 0.12
0.28 ± 0.12
0.50
0.70 ± 0.10
(mean concentration)
0.10
Thompson (1971)
Thompson et al. (1973)
Sabersky et al. (1973)
Sabersky et al. (1973)
Shair and Heitner (1974)
Shair and Heitner (1974)
Hales et al. (1974)
Sabersky et al. (1973)
Moschandreas et al. (1978)
Moschandreas et al. (1978)
Moschandreas et al. (1981)
Moschandreas et al. (1978)
Berk et al. (1980)
Berk et al. (1981)
Stock et al. (1983)
Contant et al. (1985)
Lebowitz et al. (1984)
Zhang and Lioy (1994)
Shaver et al. (1983)
Davies et al. (1984)
Shaver et al. (1983)
4-74
-------
Table 4-18 (cont'd). Summary of Reported Indoor-Outdoor Ozone Ratios
Structure
Indoor-Outdoor Ratio
Reference
Museum
Museums
(with high air exchange, but no
air-conditioning)
(with no air-conditioning and
with low air exchange rate)
(with natural convection-induced
air-exchange system)
(with conventional
air-conditioning system but with
no activated carbon air filtration)
(with activated carbon
air-filtration system)
0.45
0.69-0.84 (1 h)
0.50-0.87 (8 h)
0.10-0.59 (1 h)
0.10-0.58 (8h)
0.33-0.49 (1 h)
0.28-0.40 (8 h)
0.24-0.40 (1 h)
0.25-0.41 (8 h)
0.03-0.37 (1 h)
0.03-0.31 (8 h)
Nazaroff and Cass (1986)
Druzik et al. (1990)
aMeasured as total oxidants.
potentially different from other locations, the volume may be considered a
"microenvironment" (Duan, 1982). Microenvironments may be aggregated by location
(i.e., indoor or outdoor) or activity performed at a location (i.e., residential, commercial) to
form microenvironment types. Also, activity has two major dimensions: location and
exertion. Various microenvironments can have different levels of ventilation that will
significantly influence the delivered dose.
Air Quality Criteria for Carbon Monoxide (U.S. Environmental Protection Agency,
1991) discusses the difference between individual and population exposures. The document
notes that Sexton and Ryan (1988) define the pollutant concentrations experienced by a
specific individual during normal daily activities as "personal" or "individual" exposures.
A personal exposure depends on the air pollutant concentrations that are present in the
location through which the person moves, as well as on the time spent at each location.
Because time-activity patterns can vary substantially from person to person, individual
exposures exhibit wide variability (U.S. Environmental Protection Agency, 1991). Thus,
although it is a relatively straightforward procedure to measure any one person's exposure,
many such measurements may be needed to quantify exposures for a defined group. The
daily activities of a person in time and space define the individual's activity pattern.
Accurate estimates of air pollution exposure generally require that an exposure model account
for the activity patterns of the population of interest.
From a public health perspective, it is important to determine the "population
exposure", which is the aggregate exposure for a specified group of people (e.g., a community
or an identified occupational cohort). Because exposures are likely to vary substantially
between individuals, specification of the distribution of personal exposures within a
4-75
-------
population, including the average value and the associated variance, is often the focus of
exposure assessment studies.
In many cases, the upper tail of the distribution, which represents those individuals
exposed to the highest concentrations, is frequently of special interest because the
determination of the number of individuals who experience elevated pollutant levels can be
critical for health risk assessments. This is especially true for pollutants for which the
relationship between dose and response is highly nonlinear. Runeckles and Bates (1991) have
pointed out the importance of peak concentrations in eliciting adverse human effects. As
indicated in Section 4.1, results using controlled human exposures have shown the possible
importance of concentration in relation to duration of exposure and inhalation rate. The
implication of the importance of concentration can be translated into the conclusion that the
simple definition of exposure (i.e., equal to concentration multiplied by time) may be too
simplified.
Several human exposure models have been developed for most cases, because it is
not possible to estimate population exposure solely from fixed-station data. Some of these
models include information on human activity patterns (i.e., the microenvironments people
visit and the time they spend there). These models also contain submodels depicting the
sources and concentrations likely to be found in each microenvironment, including indoor,
outdoor, and in-transit settings.
4.8.2 Fixed-Site Monitoring Information Used To Estimate Population
and
Vegetation Exposure
Based on the information provided in earlier sections in this chapter, fixed-site
monitors alone cannot accurately depict population exposures for most cases, because indoor
and in-transit concentrations of O3 may be significantly different from ambient
O3 concentrations, and ambient outdoor concentrations of O3 that people come in contact with
may vary significantly from O3 concentrations measured at fixed-site monitors. Fixed-site
monitors measure concentrations of pollutants in ambient air. Ambient air as noted by the
U.S. Environmental Protection Agency (1991) is defined in the U.S. Code of Federal
Regulations (1991) as air that is "external to buildings, to which the general public has
access." But the nature of modern urban lifestyles in many countries, including the United
States, is that people spend an average of over 20 h per day indoors (Meyer, 1983). Reviews
of studies summarized in Section 4.7 show that indoor O3 concentration measurements vary
significantly from simultaneous measurements in ambient air. The difference between indoor
and outdoor air quality and the amount of time people spend indoors reinforce the conclusion
that using ambient air quality measurements alone does not provide accurate estimates of
population exposure in most cases.
It is assumed that exposure for vegetation is the same as the concentration
information provided at fixed monitors in the field (see Sections 5.5 and 5.6). In some cases,
because of foliar scavenging and height differences between the vegetation canopy and the
pollutant monitor, the measured concentration is not equivalent to the vegetation exposure.
A subgroup, children attending summer camp, has been studied by several
investigators to evaluate the influence of ambient air pollution on respiratory health and
function. Because children are predominantly outdoors and relatively active while at camp,
they provide a unique opportunity to examine the relationships between respiratory health and
4-76
-------
function and concurrent air pollution levels. Children may be at potentially increased risk
from air pollution by virtue of their lifestyle patterns, which often involve several hours of
outdoor exercise, regardless of air quality, during daylight hours.
For campers, attempts have been made to estimate human exposure to O3 using
types of activity patterns (Mage et al., 1985; Paul et al., 1987). Mage et al. (1985) developed
an objective approach to estimate the dose delivered to the lung of a 12-year-old camper by
using pulmonary minute volume associated with a specific activity, the fractional penetration
beyond the trachea, and infiltration of ozone indoors. Lioy and Dyba (1989) have applied the
parameters used by Mage et al. (1985) to predict the delivered O3 dose over a 4-day episodic
period. The schedule of a hypothetical camper was matched to the actual O3 concentrations,
and the predicted doses were estimated.
Several studies involving children attending summer camp have been summarized
in Chapter 7. In one study, Avol et al. (1990) reported that O3 levels at a Southern California
summer camp, located 190 km southeast of Los Angeles, CA, rose gradually throughout each
day, displaying a "broad peak" between 1000 and 2000 hours each day. Daily maxima
typically occurred in late afternoon (1500 to 1700 hours); subsequently, concentrations
gradually declined to overnight O3 levels of 0.025 to 0.050 ppm. Spektor et al. (1991)
investigated the pulmonary function of 46 healthy children on at least 7 days for each child
during a 4-week period at a northwestern New Jersey residential summer camp in 1988. The
daily levels of 1-h peak O3 and the 12-h average hygroden ion (H+) concentrations are shown
in Figure 4-27. On 5 of these days, the current NAAQS of 0.12 ppm was exceeded. The
maximum hourly concentration attained during the study was 0.15 ppm. The year 1984 was a
milder O3 exposure year and Figure 4-28 summarizes the maximal 1-h O3 concentrations at
Fairview Lake during a 1984 study period (Spektor et al., 1988).
4.8.3 Personal Monitors
A personal exposure profile can be identified by using a personal exposure
monitor. McCurdy (1994) has described the development of personal exposure monitors by
several companies. However, few data are available describing personal exposures for
individuals using these monitors. An example of a pilot study using a personal exposure
monitor was described for assessing O3 exposure in 23 children by Liu et al. (1993). The
accuracy of the monitor was within 20% of the actual value. The authors collected indoor,
outdoor, and personal O3 concentration data as well as time-activity data in State College, PA.
Results from the pilot study demonstrated that fixed-site ambient measurements may not
adequately represent individual exposures. Outdoor O3 concentrations showed substantial
spatial variation between rural and residential regions. In addition, Liu et al. (1993) reported
that models based on time-weighted indoor and outdoor concentrations explained only 40% of
the variability in personal exposures. When the model used included observations for only
those participants who spent the majority of their day in or near their homes, an R2 of 0.76
resulted when estimates were regressed on measured personal exposures. The authors
concluded that contributions from diverse indoor and outdoor microenvironments should be
considered to estimate personal O3 exposure accurately. From these results, it is clear that
additional data are needed to better quantify the O3 exposures to which populations are
exposed.
Figure 4-27. Maximum 1-h ozone (O3) concentrations (in parts per billion) and average
0800 to 2000 hours strong acid concentrations (expressed as micrograms per
4-77
-------
160
I140
.0 120
| 100
O 80
cf
s= 60
I 40
E
| 20
0
National Ambient
Air Quality Standard
H
Ozone
20 .0
I
•15 <§,
o w
10 :c<8_
^D^*^
^D c
o>
I
10 15 20
Day Number
25
30
cubic meter of sulfuric acid [H^SOJ) for each day that pulmonary function
data were collected at Fairview Lake camp in 1988. The correlation
coefficient between O3 and the hydrogen ion (H+) was 0.56.
Source: Spektor et al. (1991).
4.8.4 Population Exposure Models
McCurdy (1994) has reviewed the current status of human exposure modeling.
The author describes two distinct types of O3 exposure models: (1) those that focus narrowly
on predicting indoor O3 levels and (2) those that focus on predicting O3 exposures on a
community-wide basis. The models that predict indoor O3 levels have been described by
Sabersky et al. (1973), Shair and Heitner (1974), Nazaroff and Cass (1986), and Hayes (1989,
1991). McCurdy (1994) discusses four distinct models that predict O3 exposure on a
community-wide basis. These models and their distinguishing features are:
1. pNEM/O3 based on the National Air Quality Standards Exposure Model (NEM)
series of models (Paul et al., 1987; Johnson et al., 1990; McCurdy et al., 1991).
• Uses mass-balance approach and seasonal considerations for I/O ratio
estimation.
• Variables affecting indoor exposure obtained by Monte Carlo sampling from
empirical distributions of measured data.
4-78
-------
National Ambient Air Quality Standard
9 11 13 15 17 10 21 23 25 27 29 31 2 4
July August
1984
Figure 4-28. Maximal 1-h ozone concentrations at Fairview Lake during the study period.
Source: Spektor et al. (1988).
2. Systems Applications International (SAI)/NEM (Hayes et al., 1984; Hayes and
Lundberg, 1985; Austin et al., 1986; Hayes et al., 1988; Hayes and Rosenbaum,
1988).
• More districts and microenvironments and more detailed mass-balance
model than pNEM/O3.
• Human activity data outdated and inflexible.
3. Regional Human Exposure Model (REHEX) (Lurmann and Colome, 1991;
Winer et al., 1989; Lurmann et al., 1989; Lurmann et al., 1990).
• More detailed geographic resolution than NEM.
• Uses California-specific activity data and emphasizes in-transit and outdoor
microenvironments.
4. Event probability exposure model (EPEM) (Johnson et al., 1992).
• Estimates probability that a randomly selected person will experience a
particular exposure regime.
• Lacks multiday continuity.
McCurdy (1994) points out that all four models are related to the NEM. The
NEM is an EPA exposure model developed in the 1980s (Biller et al., 1981). Outdoor air
quality data are obtained from monitoring or modeling data. In most applications of NEM,
fixed-site monitoring data are used. The hourly average values are transformed by a suitable
relationship so that they better represent air quality outside of the various microenvironments
of interest. McCurdy (1994) points out that the important point of the NEM spatial
dimension is that people can be assigned to a monitor using U.S. census data. In addition,
4-79
-------
community trips can be assigned among the districts, grid cells, or neighborhood types using
census data. Thus, the NEM model simulates the movement of people through space for
work-trip purposes. Interested readers are referred to McCurdy (1994) for further discussion
of the pNEM model.
4.8.5 Concentration and Exposures Used in Research Experiments
It is important to adequately characterize the exposure patterns that result in
vegetation and human health effects. Hourly average concentrations used in many of the high
treatment experimental studies did not necessarily mimic those concentrations observed under
ambient conditions. Although the ramifications of this observation on the effects observed
are not clear, it was pointed out that the highest treatments used in many of the open-top
chamber experiments were bimodal in the distribution of the hourly average concentrations.
In other experiments designed to assess the effects of O3 on vegetation, constant concentration
(i.e., square wave) exposures were implemented. As has been discussed in earlier sections of
this chapter, hourly average concentrations change by the hour and square wave exposure
regimes do not normally occur under ambient conditions. In addition to the exposures used at
the highest treatment levels, there is concern that the hourly average concentrations used in
the control treatments may be lower than those experienced at isolated sites in the United
States or in other parts of the world. Although the ramifications of using such exposure
regimes are unclear, there is some concern that the use of atypically low control levels may
result in an overestimation of vegetation yield losses when used as the baseline for evaluating
the effects of treatments at higher concentrations (Lefohn and Foley, 1992).
For assessing the human health effects of O3 exposure, a series of studies has
explored prolonged 6.6-h O3 exposures at low levels (i.e., 0.08 to 0.12 ppm) (Horstman et al.,
1990). McDonnell et al. (1991), using similar hourly average concentration regimes, have
confirmed the findings reported by Horstman et al. (1990). All the research investigations
using 6.6-h durations have applied constant concentrations during the exposure period. If, as
indicated in the introduction of this chapter, concentration is more important than duration
and ventilation rate, different human health effects may occur as a result of different exposure
regimes that have identical 6.6-h average concentrations. Because of this, it is important to
explore the different types of exposure regimes that occur under ambient conditions during an
8-h episode.
Lefohn and Foley (1993) reported on an analysis of hourly average data for
O3 monitoring sites that never experienced an exceedance of an hourly average concentration
>0.12 ppm and that experienced 8-h daily maximum average concentrations >0.08 ppm. For
those monitoring sites that met the above two criteria, they identified the number of times the
8-h daily maximum average concentration exceeded 0.08 ppm during the monitoring year.
For the period 1987 to 1989, there were 925 exposure regimes identified from 166 site-years
of data that met the above criteria. The data were then organized into the following seven
categories:
I. The occurrence of 8-h daily maximum averages >0.08 ppm and <0.09 ppm;
II. The occurrence of 8-h daily maximum averages >0.08 ppm but <0.082 ppm,
which contained only hourly average concentrations >0.08 ppm but
>0.082 ppm;
III. 8-h daily maximum averages >0.08 ppm, which contained hourly average
concentrations <0.09 ppm;
4-80
-------
IV. 8-h daily maximum averages >0.08 ppm and <0.09 ppm, which contained at
least one hourly average concentration >0.09 ppm but <0.10 ppm;
V. 8-h daily maximum averages >0.08 ppm and <0.09 ppm, which contained at
least one hourly average concentration >0.10 ppm;
VI. 8-h daily maximum averages <0.08 ppm, which contained at least one hourly
average concentration >0.09 ppm but <0.10 ppm; and
VII. 8-h daily maximum averages <0.08 ppm, which contained at least one hourly
average concentration >0.10 ppm.
Figure 4-29 summarizes the results of the analysis. The results indicated that there was a
poor relationship between the value of the 8-h daily maximum average concentration and the
frequency of occurrence of hourly average concentrations within specific ranges (e.g., between
0.09 and 0.10 ppm). In no case could the authors identify a monitoring site that experienced
the square-wave type of exposure that was described in Category II (i.e., the occurrence of 8-
h daily maximum averages >0.08 ppm but <0.082 ppm). Lefohn and Foley (1993) concluded
that the square wave exposures used in the 6.6-h human health effects experiments were not
found under ambient conditions. The authors identified 453 additional exposure regimes,
where the 8-h daily maximum average was <0.08 ppm but experienced maximum hourly
average concentrations >0.09 ppm. Thus, if hourly average concentrations >0.08 ppm are of
concern for affecting human health, there will be instances where occurrences above this
threshold are evident, but the 8-h average value is below 0.08 ppm.
4.9 Concentrations of Peroxyacetyl Nitrates in
Ambient
Atmospheres
4.9.1 Introduction
The biological effects of PAN in human exposures, toxicological studies of
animals, and plant response and yield have been considered previously (U.S. Environmental
Protection Agency, 1986). Controlled human exposure studies involving O3 and O3 + PAN
are discussed elsewhere in this document (Chapter 7, Section 7.2.6.3). Some effects on
respiratory parameters have been reported in one study, but not in others. However, the PAN
concentrations used in these studies have been well above the maximum ambient
concentrations usually experienced many years ago within the Los Angeles Basin (U.S.
Environmental Protection Agency, 1986) and, more importantly, above the maximum ambient
concentrations in the more recent measurements considered in this section.
The PANs are of importance as reservoirs for NO2 as NOX is depleted relative to
VOCs in plumes moving downwind into less polluted areas (Chapter 3, Section 3.2.4).
In performance evaluation of ozone air quality models, measured concentrations of PANs are
useful in model evaluation (Chapter 3, Section 3.6.4.2).
4-81
-------
,000n
800-
600-
W
§
2
§
O
2 400
200-
IV V
Categories
VI
VII
Figure 4-29. The number of occurrences for each of the seven categories described in text.
In the previous air quality criteria for O3 and other photochemical oxidants (U.S.
Environmental Protection Agency, 1986), extensive tabulations of PAN and peroxypropionyl
nitrate (PPN, CH3CH2C(O)OONO2) concentrations were given based on measurements made
between 1965 and 1981 from references up to 1983. In the present work, references from
1983 to the present are used for measurements of PANs in urban and rural locations. The
urban area measurements are from the United States, Canada, France, Greece, and Brazil.
The use of measurements from aboard serve to illustrate or support certain U.S. results as
well as to demonstrate the widespread presence of PANs in the atmosphere. These PAN
measurements usually were of limited duration, and the results should not be assumed to be
comparable to those obtained at the O3 monitoring sites discussed earlier in this chapter (see
also Section 4.10).
4.9.2 Urban Area Peroxyacetyl Nitrate Concentrations
For urban sites, the prior criteria document for ozone and other photochemical
oxidants contains a number of tables tabulating measurements of PAN, PPN, and PPN to
PAN and PAN to O3 ratios (Altshuller, 1983; U.S. Environmental Protection Agency, 1986).
Based on comparisons of PAN measurements in Los Angeles in 1980 with those made in the
1960s, it was uncertain whether PAN concentrations had decreased. In the Los Angeles area,
the average and maximum PAN concentrations reported ranged from 1.6 to 31 ppb and from
6 to 214 ppb, respectively. The wide variations, at least in part, were associated with the
range of years, different seasons, and differing average times among studies. On average, the
4-82
-------
PPN to PAN ratios among studies in Los Angeles ranged from 0.15 to 0.20, whereas the PAN
to O3 ratios among studies ranged from 0.04 to 0.20. In the earlier PAN measurement results,
studies conducted in the South Coast Air Basin predominated.
The average PAN concentrations measured in other cities usually were lower than
in the Los Angeles area, whereas the maximum PAN concentrations overlapped with the
lower end of range in Los Angeles. The PPN and PAN ratios in other cities ranged from 0.1
to 0.4, whereas the PAN to O3 ratios were in the 0.01 to 0.05 range.
Seasonally, PAN to O3 ratios tended to be somewhat higher in the winter. The
diurnal characteristics of O3 and of PAN were similar, but not identical.
The urban area measurement results are tabulated in Table 4-19. The earlier
maximum PAN concentrations reported usually were substantially higher than those given in
Table 4-19. A possible exception occurs for the Claremont, CA, results. Measurements of
PAN and PPN were made in 1989 and 1990 at sites downwind of Los Angeles: Perrin,
90 km to the east-southeast, and Palm Springs, 120 km to the east (Grosjean and Williams,
1992). The concentrations of PAN and PPN were high, and the concentration maxima
occurred during the evening hours, which was consistent with downwind transport from the
Los Angeles area rather than from local sources.
In Southern California, the maximum PAN concentrations appear to be more
evenly distributed spatially during the fall than during the summer (Williams and Grosjean,
1990). At coastal and central locations, the PAN maxima during the fall were comparable to
those observed at inland locations during the summer.
As observed previously, PAN concentrations in other U.S. cities and in cities in
other countries tend to be substantially lower than in Los Angeles and its surrounding urban
areas (Table 4-19). An exception occurs for the measurements from Paris (Tsalkani et al.,
1991). Maximum PAN concentrations in the 20 to 35 ppb range were observed.
In measurements made in 1992 in Atlanta, GA, at the Georgia Institute of
Technology campus site, not only were PAN and PPN measured, but very occasionally
peroxymethacryloyl nitrate (MPAN, CH2=C(CH3) C(O)OONO2) (a product of the atmospheric
photooxidation of local biogenic sources of isoprene) was observed (Williams et al., 1993).
Maximum diurnal concentrations of peroxyacyl nitrates and O3 occur in late afternoon and
early evening. The average MPAN concentration was 0.3 ppb, and the maximum value was
0.5 ppb and constituted about 15% of the concurrent PAN concentrations.
In a study in Rio de Janeiro performed to investigate the effects of the use of
ethanol or ethanol-containing fuel on PAN concentrations, the maximum PAN concentration
reached 5.4 ppb (Tanner et al., 1988). However, this maximum concentration is well below
the maximum concentrations reported in and around Los Angeles, and it falls within the
maximum PAN values reported for a number of other cities (Table 4-19).
4.9.3 Concentration of Peroxyacetyl Nitrate and Peroxypropionyl
Nitrate
in Rural Areas
Prior measurements of nonurban PAN and PPN concentrations and PAN to
O3 ratios are available (Altshuller, 1983; U.S. Environmental Protection Agency, 1986).
At nonurban sites that are not impacted by urban plumes, PAN and PPN concentrations are
4-83
-------
Table 4-19. Summary of Measurements
and Peroxypropionyl Nitrate in
of Peroxyacetyl Nitrate
Urban Areas3
Site
Long Beach, CA
Anaheim, CA
Los Angeles, CA
Burbank, CA
Azusa, CA
Claremont, CA
Perrin, CA
Palm Springs, CA
Downey, CA
Boulder, CO
-^
2 Denver, CO
Houston, TX
Philadelphia, PA
Staten Island, NY
Atlanta, GA
Edmonton, Alberta,
Canada
Calgary, Alberta,
Canada
University of
Calgary, Alberta,
Canada
Simcoe, Ontario,
Canada
Month/
Year
6-12/1987
6-12/1987
6-12/1987
6-12/1987
6-9/1987
6-9/1987
6/1989 to 6/1990
6/1989 to 6/1990
2/1984
5, 6 and
8, 9/1987
3/1984
3/1984
4/1983
4/1983
7, 8/1992
12/1983 to 4/1984
7/1981 to 2/1982
10/1980 to 8/1981
6/1980 to 3/1981
Number
of Days
Sampled
16
14
16
16
11
10
NA
NA
10
12
45
9
9
19
7
36
66
213
175
191
PAN
Concentration
(ppb)
Average/Mean
NA
NA
NA
NA
NA
NA
1.6
1.6
1.2
0.63
0.59
0.64
0.75
1.1
1.6
0.71
0.14
0.22
1.3
Max
16
19
13
19
13
30
9.1
7.6
6.7
2.0
3.8
2.0
7.9
3.7
5.5
2.9
7.5
6.6
2.4
5.6
PPN
Concentration
(ppb)
Average/Mean
NA
NA
NA
NA
NA
NA
NA
NA
0.06
0.08
0.07
0.02
0.045
0.14
0.21
0.14
NA
NA
NA
NA
Max
NA
NA
NA
NA
NA
NA
0.73
0.42
0.40
0.3
0.6
0.09
0.54
0.50
0.90
0.37
NA
NA
NA
NA
Reference
Williams and Grosjean (1990)
Williams and Grosjean (1990)
Williams and Grosjean (1990)
Williams and Grosjean (1990)
Williams and Grosjean (1990)
Williams and Grosjean (1990)
Grosjean and Williams (1992)
Grosjean and Williams (1992)
Singh and Salas (1989)
Ridley et al. (1990)
Singh and Salas (1989)
Singh and Salas (1989)
Singh and Salas (1989)
Singh and Salas (1989)
Williams et al. (1993)
Peake et al. (1988)
Peake and Sandhu (1983)
Peake and Sandhu (1983)
Corkum et al. (1986)
-------
Table 4-19 (cont'd). Summary of Measurements of Peroxyacetyl Nitrate
and Peroxypropionyl Nitrate in Urban Areas3
Site
Rio de Janeiro
Vila Isabel
PUC/RJ
Athens, Greece
Paris, France
aSee Appendix A
£
Oi
Number
Month/ of Days
Year Sampled
7/1985 8
7/1985 4
2-11/1985 113
11/1985 to 11/1986 NA
for abbreviations and acronyms.
PAN
Concentration
(ppb)
Average/Mean Max
NA 5.4
NA 3.3
NA 3.7
1.1 20.5
PPN
Concentration
(ppb)
Average/Mean
NA
NA
NA
NA
Max Reference
1.0 Tanner et al. (1988)
0.6 Tanner et al. (1988)
NA Tsani-Bazaca et al. (1988)
NA Tsalkani et al. (1991)
-------
much lower than those in urban areas. Average PAN concentrations ranged between 0.1 and
1.0 ppb, whereas the PAN to O3 ratios were at or below 1.
Concentrations of PAN, PPN, and other peroxyacyl nitrates have been reported
(Table 4-20) at Tanbark Flat, CA, 35 km northeast of Los Angeles, during 1989, 1990, and
1991 and at Franklin Canyon, CA, 25 km west of Los Angeles, during 1991 (Grosjean and
Williams, 1992; Grosjean et al., 1993). As indicated by the results tabulated in Table 4-20,
the concentrations were high at these mountain sites, the PPN to PAN ratios were relatively
high, and the concentration maxima occurred during the afternoon hours. These concentration
levels of PAN and PPN are attributed to downwind transport from the Los Angeles urban
area. The MPAN was occasionally detected with average concentrations of 1.2 ppb at
Tanbark Flat and 1.0 ppb at Franklin Canyon in 1991.
At Tanbark Flat, the O3 and PAN diurnal concentration patterns were similar to
those in upwind urban areas. The PAN to O3 ratios at the O3 maximum were as follows:
1989, 0.05; 1990, 0.08; 1991, 0.05; all the ratios are within the same range as at sites in
urban areas in and around Los Angeles.
Additional measurements of PAN and PPN or other peroxyacyl nitrates are
available over a period of years at Niwot Ridge, CO, just west of the Denver-Boulder area; at
Point Arena, CA; and at a forest site, Scotia, PA (Ridley et al., 1990). The concentrations
reported at all of these sites are much lower than the mountain sites in California. The Niwot
Ridge site concentrations, which show the effects of easterly upslope flow of air parcels from
Denver-Boulder, are still low compared to the sites downwind of the urban Los Angeles area
(Table 4-20).
The PAN concentrations at the Scotia rural site in the eastern United States tend to
be somewhat higher than the Niwot Ridge or Point Arena sites (Table 4-20). This difference
may relate to higher regional precursor concentration levels.
4.10 Concentration and Patterns of Hydrogen
Peroxide in
the Ambient Atmosphere
Efforts to measure H2O2 began in the 1970s, but the early reports of H2O2
concentrations above 10 ppb and even 100 ppb appear to be in error because of the artifact
H2O2 generated within the presence of O3 (Chapter 3, Section 3.5.1.3). Subsequent
measurements of H2O2 in the 1980s resulted in maximum H2O2 concentrations at or below
5 ppb and mean concentrations at or below 1 ppb (Sakugawa et al., 1990).
Studies comparing more recent methods for measuring H2O2, which were
conducted in North Carolina, indicated differences among measurement methods in synthetic
mixtures of H2O2, including possible interferences, and in the ambient atmosphere of up to
about ±25% (Kleindienst et al., 1988). However, results from the same study from mixtures
irradiated in a smog chamber produced larger differences among methods, especially for the
luminol technique compared to the fluorescence technique and with tunable-diode laser
absorption spectroscopy. Another comparison study conducted in California resulted in
differences between methods for measuring H2O2 that varyied by a factor or two (Lawson et
al., 1988). In the measurements of H2O2 discussed below, the cryogenic fluorescence method
or the scrubber-coil fluorescence method generally was used.
4-86
-------
Table 4-20. Summary of Measurements of Peroxyacetyl Nitrate
and Peroxypropionyl Nitrate in Rural Areas3
Site
Tanbark Flat, CA
Franklin Canyon, CA
Niwot Ridge, CO
Point Arena, CA
&
Scotia, PA
Kananaskis Valley, Alberta,
Canada
Frijoles Mesa, NM
Month/
Year
8-10/1989
8,9/1990
8/1991
9/1991
7/1984
6, 7/1984
8, 9/1984
6, 7/1987
1/1984
Spring 1985
Summer 1986
6-8/1988
9/1979,
4/1982,
6-8/1982
10/1987 to
1/1989
Number
of Days
Sampled
69
34
22
9
16
23
21
46
14
NA
NA
47
NA
NA
PAN
Concentration
(ppb)
Average/Mean
2.9
4.8
2.8
1.6
0.28
=0.25
=0.25
0.81 (b)
0.21 O
0.12
0.05
=0.6
1.0
=0.5
0.26
Max
>16.1
22.0
12.8
7.0
2.3
NA
NA
3.2
1.1
NA
NA
NA
2.3
1.9
PPN
Concentration
(ppb)
Average/Mean
0.75
0.76
0.43
0.18
0.016
NA
NA
0.08 (b)
0.01 O
0.005
NA
NA
NA
NA
NA
Max
5.1
4.3
2.66
1.15
0.17
NA
NA
0.45
0.07
NA
NA
NA
NA
NA
Reference
Williams and Grosjean (1991)
Grosjean et al. (1993)
Grosjean et al. (1993)
Grosjean et al. (1993)
Singh and Salas (1989)
Fahey et al. (1986)
Fahey et al. (1986)
Ridley et al. (1990)
Singh and Salas (1989)
Ridley (1991)
Ridley (1991)
Buhr et al. (1990)
Peake et al. (1983)
Gaffney et al. (1993)
"See Appendix A for abbreviations and acronyms.
bFlow from Boulder-Denver area.
°Flow across the Rockies.
-------
Based on interpretation of a compilation of H2O2 measurements made between
1984 and 1988 at a number of urban locations, at rural/remote locations, and on aircraft
flights, it was concluded that the higher H2O2 concentrations were associated with the
following measurement conditions: in afternoon hours, during summer months, at rural
locations, and at lower latitudes (Sakugawa et al., 1990; Van Valin et al., 1987). The H2O2
concentrations increase from the surface to the top of the boundary layer (Daum et al., 1990).
Available values for mean H2O2 concentrations at three U.S. locations were (1) at the summit
of Whitetop Mountain, VA: summer, 0.80 ppb; winter, 0.15 ppb (Olszyna et al., 1988); (2) at
the summit of Whiteface Mountain: 1986, 0.6 ppb; 1987; 0.8 ppb (Mohnen and Kadlecek,
1989); and (3) at Westwood, CA: summer, =1.0 ppb; winter, 0.2 ppb (Sakugawa and Kaplan,
1989). At Westwood, the highest correlation with various parameters was found for solar
radiation consistent with the higher H2O2 concentrations being observed in the afternoon
during the late spring and early summer months (Sakugawa and Kaplan, 1989). In the same
study, the average H2O2 concentrations were observed to increase from Westwood, near the
coast in the Los Angeles Basin, to Duarte, inland; at Daggett in the Mohave Desert; and at
Sky Mountain and Lake Gregory in the San Bernadino Mountains. The ratios of O3 to H2O2
concentrations at the these sites were >100. In subsequent measurements, the same
relationship in H2O2 concentrations between Westwood and the other California sites listed
above was observed (Sakugawa and Kaplan, 1993). Unlike the results at several urban sites
and other mountain sites, it was reported that the highest diurnal H2O2 concentrations at Lake
Gregory in the San Bernardino Mountains were observed during the nighttime hours
(Sakugawa and Kaplan, 1993).
4.11 Co-occurrence of Ozone
4.11.1 Introduction
There have been several attempts to characterize air pollutant mixtures (Lefohn and
Tingey, 1984; Lefohn et al., 1987b). Pollutant combinations can occur at or above a
threshold concentration either together or temporally separated from one another. For
example, for characterizing the different types of co-occurrence patterns, Lefohn et al.
(1987b) grouped air quality data within a 24-h period starting at 0000 hours and ending at
2359 hours. Patterns that showed air pollutant pairs appearing at the same hour of the day at
concentrations equal to or greater than a minimum hourly mean value were defined as
"simultaneous-only" daily co-occurrences. When pollutant pairs occurred at or above a
minimum concentration during the 24-h period, without occurring during the same hour, a
"sequential-only" co-occurrence was defined. During a 24-h period, if the pollutant pair
occurred at or above the minimum level at the same hour of the day and at different hours
during the period, the co-occurrence pattern was defined as "complex-sequential".
A co-occurrence was not indicated if one pollutant exceeded the minimum concentration just
before midnight and the other pollutant exceeded the minimum concentration just after
midnight. As will be discussed below, studies of the joint occurrence of gaseous NO2/O3 and
SO2/O3 reached two conclusions: (1) the co-occurrence of two-pollutant mixtures lasted only
a few hours per episode, where an episode was defined by the threshold concentration used,
and (2) the time between episodes is generally long (i.e., weeks, sometimes months) (Lefohn
and Tingey, 1984; Lefohn et al., 1987b).
4-88
-------
For exploring the co-occurrence of O3 and other pollutants (e.g., acid precipitation,
acidic cloudwater, and acidic sulfate aerosols), there are limited data available. In most cases,
routine monitoring data are not available from which to draw general conclusions. However,
published results are reviewed and summarized for the purpose of assessing an estimate of the
possible importance of co-occurrence patterns of exposure.
4.11.2 Nitrogen Oxides
Ozone occurs frequently at concentrations >0.03 ppm at many rural and remote
monitoring sites in the United States (Evans et al., 1983; Lefohn, 1984; Lefohn and Jones,
1986). Therefore, for many rural locations in the United States, the co-occurrence patterns
observed by Lefohn and Tingey (1984) for O3 and NO2 were defined by the presence or
absence of NO2. As anticipated, Lefohn and Tingey (1984) reported that most of the sites
analyzed experienced fewer than 10 co-occurrences (when both pollutants were present at an
hourly average concentration >0.05 ppm). However, the authors did note that several urban
monitoring sites in the South Coast Air Basin experienced more than 450 co-occurrences.
The rural sites of Riverside, Fontana, and Rubidoux, CA, had more than 100 co-occurrences.
Denver and San Jose, CA, also experienced more than 100 co-occurrences of O3/NO2.
Lefohn and Tingey (1984) reported that for Rubidoux, because NO2 concentration maxima
tended to peak in the evenings or early morning, the co-occurrences were present at these
times. For more moderate areas of the country, Lefohn et al. (1987b) reported that even with
a threshold of 0.03 ppm O3, the number of co-occurrences with NO2 was small.
4.11.3 Sulfur Dioxide
Because elevated SO2 concentrations are mostly associated with industrial activities
(U.S. Environmental Protection Agency, 1992a), co-occurrence observations are usually
associated with monitors located near these types of sources. Lefohn and Tingey (1984)
reported that, for the rural and nonrural monitoring sites investigated, most sites experienced
fewer than 10 co-occurrences of SO2 and O3. Only Rockport, IN, and Paradise No. 21 (KY)
had more than 40 co-occurrences during the monitoring period (48 and 45, respectively). The
monitors at these two sites were influenced by the local sources. The authors noted that at
Fontana there were numerous O3 episodes above 0.05 ppm, and there was a high probability
that when the SO2 hourly average concentrations rose above 0.05 ppm, both pollutants would
be present at levels equal to or greater than 0.05 ppm.
Meagher et al. (1987) reported that several documented O3 episodes at specific
rural locations appeared to be associated with elevated SO2 levels. The investigators defined
the co-occurrence of O3 and SO2 to be when hourly mean concentrations were equal to or
greater than 0.10 and 0.01 ppm, respectively. On reviewing the hourly mean O3 and SO2 data
used by Lefohn et al. (1987b) in 1980 (using a threshold of 0.05 ppm for both pollutants), the
Paradise No. 23 (KY); Giles County, TN; Murphy Hill (reported as Marshall County by
Meagher et al., 1987), AL; and Saltillo (reported as Hardin Co. by Meagher et al., 1987), TN,
sites experienced fewer than 7 days over a 153-day period for a co-occurrence of any form
(i.e., simultaneous only, sequential, and complex co-occurrence). Thus, as reported by Lefohn
et al. (1987b), the co-occurrence pattern of O3 and SO2 was infrequent.
The above discussion was based on the co-occurrence patterns associated with the
presence or absence of hourly average concentrations of pollutant pairs. Taylor et al. (1992)
have discussed the joint occurrence of O3, nitrogen, and sulfur in forested areas using
4-89
-------
cumulative exposures of O3 with data on dry deposition of sulfur and nitrogen. The authors
concluded in their study that the forest landscapes with the highest loadings of sulfur and
nitrogen via dry deposition tended to be the same forests with the highest average
O3 concentrations and largest cumulative exposure. Although the authors concluded that the
joint occurrences of multiple pollutants in forest landscapes were important, nothing was
mentioned about the hourly co-occurrences of O3 and SO2 or O3 and NO2.
4.11.4 Acidic Sulfate Aerosols
Acid sulfates, which are usually composed of sulfuric acid (H2SO4), ammonium
bisulfate, and ammonium sulfate, have been measured at a number of locations in North
America. Acidic sulfate and neutralized species can accumulate and range in concentration
from 0 to 50 |ig/m3 at a specific location or a number of locations simultaneously (Lioy,
1989). For many summertime studies, peaks of H2SO4 or FT appear to be associated with the
presence of a slow-moving high pressure system (Lioy and Waldman, 1989). Acid sulfates
are found primarily in the fine particle size range (<2.5 jim in diameter). Lioy (1989) reports
that the acidic sulfate concentrations measured in the summertime can be found at 20 |ig/m3
for over an hour and can be found at high concentrations of 10 to 20 |ig/m3 for 6 to 24 h at
one or more sites (Lioy, 1989). Acidic sulfate aerosol concentrations can occur at
concentrations in the summertime above 10 |ig/m3 for periods longer than 5 h (Lioy, 1989).
As has been discussed earlier in this chapter, the highest O3 exposures for sites affected by
anthropogenically derived photooxidant precursors are expected to occur during the late spring
and summer months. Thus, the potential for O3 and acidic sulfate aerosols to co-occur at
some locations in some form (i.e., simultaneously, sequentially, or complex-sequent ally) is
real. Our knowledge of the potential exposure of the co-occurrence of acidic sulfate aerosols
and O3 is limited because routine monitoring data for acidic aerosols are not available.
Information on the co-occurrence patterns is limited to research studies and some of the
results of these studies are provided in this section.
Spektor et al. (1991) investigated the effects of single- and multiday O3 exposures
on respiratory function in active normal children aged 8 to 14 years at a northwestern New
Jersey residential summer camp in 1988. During the investigation, the authors measured daily
levels of 1-h peak O3 and the 12-h average H+ concentrations. On 7 days, the acid aerosol
concentrations (reported as H2SO4) were higher than 10 |ig/m3, reaching a 12-h maximum of
18.6 |ig/m3. Figure 4-27 shows the relationship between daily maximum O3 and daily 12-h
average H+ concentrations. Thurston et al. (1992) reported occurrences in 1988 of maximum
24-h average concentrations of H+ as high as 18.7 |ig/m3 (Buffalo, NY) and a maximum daily
hourly average concentration of 0.164 ppm. Although lower than Buffalo, high O3 or FT
values were reported by the investigators for Albany and White Plains, NY. It is unclear
whether the O3 or H+ maximum concentrations occurred simultaneously; however, it is clear
that high concentrations could occur either sequentially, complex-sequentially, or
simultaneously. Evidence exists in the literature indicating that hourly co-occurrences are
experienced. Raizenne and Spengler (1989) described an episodic co-occurrence pattern in
1986 of high hourly averaged concentrations of O3 and H2SO4 that occurred at a residential
summer camp located on the north shore of Lake Erie, Ontario, (Figure 4-30). Thurston et al.
(1994) conducted a study of ambient acidic aerosols in the Toronto, Ontario, metropolitan
area in July and August of 1986, 1987, and 1988, and reported on the fine particle
(aerodynamic equivalent diameter <2.5 jim) samples collected twice per day. The authors
4-90
-------
reported that their results indicated that acidic aerosol episodes (i.e., H+ > 100 nmol/m3)
occurred routinely during the summer months and that H+ peaks were correlated with sulfate
(SO4) episodes. Figure 4-31 illustrates the relationship among SO4, H+, and O3.
CO
C
O
1
*j
I
O
O
CD
O
N
O
300
290
280
270
260
250
240
230
220
210
200
190
180
170
..A-
Ozone
H2SO4
50
40
30 '-g
*->
I
20°
Cf
W
103?
8
10 11
15 16 17 18 19
12 13 14
Time (h)
Figure 4-30. The co-occurrence pattern of ozone and sulfuric acid (H^OJ for July 25,
1986, at a summer camp on the north shore of Lake Erie, Ontario, Canada,
Source: Raizenne and Spengler (1989).
4.11.5 Acid Precipitation
Concern has been expressed about the possible effects on vegetation from
co-occurring exposures of O3 and acid precipitation (Prinz et al., 1985; National Acid
Precipitation Assessment Program, 1987; Prinz and Krause, 1988). Little information has
been published concerning the co-occurrence patterns associated with the joint distribution of
O3 and acidic deposition (i.e., FT). Lefohn and Benedict (1983) reviewed EPA's SAROAD
monitoring data for 1977 through 1980 and, using National Atmospheric Deposition Program
(NADP) and EPRI wet deposition data, evaluated the frequency distribution of pH events for
34 NADP and 8 EPRI chemistry monitoring sites located across the United States.
Unfortunately, there were few sites where O3 and acidic deposition were comonitored.
Figure 4-31. Sulfate (SO'), hydrogen ion (H+), and ozone (OJ measured at Breadalbane
Street (Site 3) in Toronto during July and August 1986, 1987, and 1988.
Source: Thurston et al. (1994).
4-91
-------
600
1986
1987
a 13
July
August
July
August
As a result, Lefohn and Benedict (1983) focused their attention on O3 and
acidic deposition monitoring sites that were closest to one another. In some cases, the sites
were as far apart as 144 km. Using hourly O3 monitoring data and weekly and event acidic
deposition data from the NADP and EPRI databases, the authors identified specific locations
where the hourly mean O3 concentrations were >0.1 ppm and 20% of the wetfall daily or
weekly samples were below pH 4.0. Elevated levels of O3 were defined as hourly mean
concentrations equal to or greater than 0.1 ppm. Although for many cases, experimental
research results of acidic deposition on agricultural crops show few effects at pH levels above
3.5 (National Acid Precipitation Assessment Program, 1987), it was decided to use a pH
threshold of 4.0 to take into consideration the possibility of synergistic effects of O3 and
acidic deposition.
Based on their analysis, Lefohn and Benedict (1983) reported five sites where
there may be the potential for agricultural crops to experience additive, less than additive, or
synergistic (i.e., greater than additive) effects from elevated O3 and H+ concentrations. The
authors stated that they believed, based on the available data, the greatest potential for
interaction between acid rain and O3 concentrations in the United States, with possible effects
on crop yields, may be in the most industrial areas (e.g., Ohio and Pennsylvania).
However,they cautioned that, because no documented evidence existed to show that pollutant
interaction had occurred under field growth conditions and ambient exposures, their
conclusions should only be used as a guide for further research.
In their analysis, Lefohn and Benedict (1983) found no colocated sites. The
authors rationalized that data from non-co-monitoring sites (i.e., O3 and acidic deposition)
could be used because O3 exposures are regional in nature. However, work by Lefohn et al.
(1988a) has shown that hourly mean O3 concentrations vary from location to location within a
region, and that cumulative indices, such as the percent of hourly mean concentrations >0.07
ppm, do not form a uniform pattern over a region. Thus, extrapolating hourly mean
O3 concentrations from known locations to other areas within a region may provide only
qualitative indications of actual O3 exposure patterns.
4-92
-------
In the late 1970s and the 1980s, both the private sector and the government
funded research efforts to better characterize gaseous air pollutant concentrations and wet
deposition. The event-oriented wet deposition network, EPRI/Utility Acid Precipitation Study
Program, and the weekly oriented sampling network, NADP, provided information that can be
compared with hourly mean concentrations of O3 collected at several comonitored locations.
No attempt was made to include H+ cloud deposition information. In some cases, for
mountaintop locations (e.g., Clingman's Peak, Shenandoah, Whiteface Mountain, and
Whitetop Mountain), the H+ cloud water deposition is greater than the H+ deposition in
precipitation (Mohnen, 1989), and the co-occurrence patterns associated with O3 and cloud
deposition will be different than those patterns associated with O3 and deposition in
precipitation.
Smith and Lefohn (1991) explored the relationship between O3 and H+ in
precipitation, using data from sites that monitored both O3 and wet deposition simultaneously
and within one minute latitude and longitude of each other. The authors reported that
individual sites experienced years in which both H+ deposition and total O3 exposure were at
least moderately high (i.e., annual H+ deposition >0.5 kg ha"1 and an annual O3 cumulative,
sigmoidally weighted exposure (W126) value >50 ppm-h). With data compiled from all sites,
it was found that relatively acidic precipitation (pH < 4.31 on a weekly basis or pH < 4.23 on
a daily basis) occurred together with relatively high O3 levels (i.e., W126 values > 0.66 ppm-
h for the same week or W126 values > 0.18 ppm-h immediately before or after a rainfall
event) approximately 20% of the time, and highly acidic precipitation (i.e, pH < 4.10 on a
weekly basis or pH < 4.01 on a daily basis) occurred together with a high O3 level (i.e.,
W126 values > 1.46 ppm-h for the same week or W126 values > 0.90 ppm-h immediately
before or after a rainfall event) approximately 6% of the time. Whether during the same
week or before, during, or after a precipitation event, correlations between O3 level and pH
(or H+ deposition) were weak to nonexistent. Sites most subject to relatively high levels of
both hydrogen ion and O3 were located in the eastern portion of the United States, often in
mountainous areas.
4.11.6 Acid Cloudwater
In addition to the co-occurrence of O3 and acid precipitation, results have been
reported on the co-occurrence of O3 and acidic cloudwater in high-elevation forests. Vong
and Guttorp (1991) characterized the frequent O3-only and pH-only, single-pollutant episodes,
as well as the simultaneous and sequential co-occurrences of O3 and acidic cloudwater. The
authors reported that both simultaneous and sequential co-occurrences were observed a few
times each month above the cloud base. Episodes were classified by considering hourly
O3 average concentrations >0.07 ppm and cloudwater events with pH < 3.2. The authors
reported that simultaneous occurrences of O3 and pH episodes occurred two to three times per
month at two southern sites (Mitchell, NC, and Whitetop, VA) and the two northern sites
(Whiteface Mountain, NY, and Moosilauke, NH) averaged one episode per month.
No co-occurrences were observed at the central Appalachian site (Shenandoah, VA), due to a
much lower cloud frequency. Vong and Guttorp (1991) reported that the simultaneous
occurrences were usually of short duration (mean =1.5 h/episode) and were followed by an
O3-only episode. As would be expected, O3-only episodes were longer than co-occurrences
and pH episodes, averaging an 8-h duration.
4-93
-------
4.12 Summary
Ozone is a pervasive compound that is detected at all monitoring locations
throughout the world. To obtain a better understanding of the potential for ambient
O3 exposures affecting human health and vegetation, hourly average concentration information
is summarized for urban, rural forested, and rural agricultural areas in the United States.
The distribution of O3 or its precursors at a rural site near an urban source is
affected by wind direction (i.e., whether the rural site is located up- or downwind from the
source). It is difficult to apply land-use designations to the generalization of exposure
regimes that may be experienced in urban versus rural areas, because the land use
characterization of "rural" does not imply that a specific location is isolated from
anthropogenic influences. Rather, the characterization implies only the current use of the
land. Because it is possible for urban emissions, as well as O3 produced from urban area
emissions, to be transported to more rural downwind locations, elevated O3 concentrations can
occur at considerable distances from urban centers. Urban O3 concentration values often are
depressed because of titration by NO. Because of the absence of chemical scavenging, O3
tends to persist longer in nonurban than in urban areas, and exposures may be higher in
nonurban than in urban locations.
For vegetation, as indicated in Chapter 5 (Section 5.5), extensive research has
focused on identifying exposure indices with a firm foundation on biological principles.
Many of these exposure indices have been based on research results indicating that the
magnitude of vegetation responses to air pollution is more an effect of the magnitude of the
concentration than the length of the exposure. For O3, the short-term (1- to 8-h), high
concentration exposures (>0.1 ppm) have been identified by many researchers as being more
important than long-term, low concentration exposures in producing visible injury to plants
(see Chapter 5 for further discussion). Similarly, for human health considerations, results
using controlled human exposures have shown the possible importance of concentration in
relation to duration of exposure and inhalation rate.
In summarizing the hourly average concentrations in this chapter, specific attention
is given to the relevance of the exposure indices used. For example, for human health
considerations, concentration (or exposure) indices such as the daily maximum 1-h average
concentrations, as well as the number of daily maximum 4- or 8-h average concentrations
above a specified threshold, are used to characterize information in the population-oriented
locations. For vegetation, several different types of exposure indices are used. Because much
of the NCLAN exposure information is summarized in terms of 7-h average concentrations,
this exposure index is used. However, because peak-weighted, cumulative indices (i.e.,
exposure parameters that sum the products of hourly average concentrations multiplied by
time over an exposure period have shown considerable promise in relating exposure and
vegetation response (see Chapter 5, Section 5.5), several exposure indices that use either a
threshold or a sigmoidally weighted scheme are used in this chapter to provide insight
concerning the O3 exposures that are experienced at a select number of rural monitoring sites
in the United States. The peak-weighted cumulative exposure indices such as SUM06,
SUM08, and W126 are used.
Ozone hourly average concentrations have been recorded for many years by the
State and local air pollution agencies who report their data to EPA. The 10-year (1983 to
1992) composite average trend for the second highest daily maximum hourly average
concentration during the O3 season for 509 trend sites and a subset of 196 NAMS sites,
4-94
-------
shows that the 1992 composite average for the trend sites is 21% lower than the 1983 average
and 20% lower for the subset of NAMS sites. The 1992 value is the lowest composite
average of the past 10 years. The 1992 composite average is significantly less than all the
previous 9 years, 1983 to 1991. The relatively high O3 concentrations in 1983 and 1988 were
attributable, in part, to hot, dry, stagnant conditions in some areas of the country that were
especially conducive to O3 formation.
From 1991 to 1992, the composite mean of the second highest daily maximum 1-h
O3 concentrations decreased 7% at the 672 sites and 6% at the subset of 222 NAMS sites.
Also, from 1991 to 1992, the composite average of the number of estimated exceedances of
the O3 standard decreased by 23% at the 672 sites, and 19% at the 222 NAMS sites.
Nationwide VOC emissions decreased 3% from 1991 to 1992 (U.S. Environmental Protection
Agency, 1993). The composite average of the second daily maximum concentrations
decreased in 8 of the 10 EPA regions from 1991 to 1992, and remained unchanged in Region
VII. Except for Region VII, the 1992 regional composite means are lower than the
corresponding 1990 levels. Although meteorological conditions in the east during 1993 were
more conducive to O3 formation than those in 1992, the composite mean level for 1993 was
the second lowest composite average for the decade (1984 to 1993).
Information is provided in this chapter on methods used for investigating
techniques for adjusting O3 trends for meteorological influences. Historically, the long-term
O3 trends in the United States characterized by EPA have emphasized air quality statistics that
are closely related to the NAAQS. Information is provided on the use of alternative indices.
Besides EPA, additional investigators have assessed trends at several locations in the United
States, and information is provided for both urban and rural areas.
Interest has been expressed in characterizing O3 exposure regimes for sites
experiencing daily maximum 8-h concentrations above specific thresholds (e.g., 0.08 or
0.10 ppm). Documented evidence has been published showing the occurrence, at some sites,
of multihour periods within a day of O3 at levels of potential health effects. Although most
of these analyses were made using monitoring data collected from sites in or near
nonattainment areas, one analysis showed that at five sites, two in New York state, two in
rural California, and one in rural Oklahoma, an alternative O3 standard of an 8-h average of
0.10 ppm would be exceeded even though the existing 1-h standard would not be. The study
indicated the occurrence at these five sites, none of which was in or near a nonattainment
area, of O3 concentrations showing only moderate peaks but showing multihour levels above
0.10 ppm.
An important question is whether an improvement in O3 levels would produce
distributions of 1-h O3 concentrations that result in a broader diurnal profile than those seen
in high-oxidant urban areas where O3 regimes contain hourly average concentrations with
sharper peaks. The result would be an increase in the number of exceedances of daily
maximum 8-h average concentrations >0.08 ppm, when compared to those sites experiencing
sharper peaks. One research effort observed, using aerometric data at specific sites, how O3
concentrations change when the sites change compliance status. One of the parameters
examined was 4-h daily maxima. The number of exceedances for a specific daily maximum
average concentration tended to decrease as fewer exceedances of the current 1-h standard
were observed at a given site. The number of occurrences of the daily maximum 4-h average
concentration >0.08 ppm and the number of exceedances of the current form of the standard
had a positive, weak correlation (r = 0.51). The investigators reported few changes in the
shape of the average diurnal patterns as sites changed attainment status. The lack of a change
4-95
-------
in shape may have explained why the investigators could not find evidence that the number of
occurrences of the daily maximum 4-h average concentration >0.08 ppm increased when the
sites experienced few high hourly average concentrations.
There has been considerable interest in possibly substituting one index for another
when attempting to relate O3 exposure with an effect. For example, using O3 ambient air
quality data, the number of exceedances of 0.125 ppm and the number of occurrences of the
daily maximum 8-h average concentrations >0.08 ppm have been compared with the result
that a positive correlation (r = 0.79) existed between the second-highest 1-h daily maximum
in a year and the expected number of days with an 8-h daily maximum average concentration
>0.08 ppm O3. However, there was not much predictive strength in using one O3 exposure
index to predict another. Similarly, the maximum 3-mo SUM06, second highest daily
maximum hourly average concentration, and second highest daily maximum 8-h average
concentration exposure indices were compared. For the rural agricultural and forest sites, the
correlations among the indices were not strong.
One of the difficulties in attempting to use correlation analysis between indices for
rationalizing the substitution of one exposure index for another for predicting an effect (e.g.,
SUM06 versus the second highest daily maximum hourly average concentration) is the
introduction of the error associated with estimating levels of one index from those of another.
Evidence has been presented in the literature for recommending that, if a different exposure
index (e.g., second highest daily maximum hourly average concentration) is to be compared
to, for example, the SUM06 for adequacy in predicting crop loss, then the focus should be on
how well the two exposure indices predict crop loss using the effects model that is a function
of the most relevant index and not on how well the indices predict one another. Less error
would be introduced if either of the two indices were used directly in the development of an
exposure-response model.
The EPA has indicated that a reasonable estimate, as an annual average, of
O3 background concentration near sea level in the United States today is from 0.020 to
0.035 ppm; this estimate included a 0.005 to 0.015 ppm contribution from the stratosphere.
The EPA concluded that a reasonable estimate of natural O3 background concentration for a
1-h daily maximum at sea level in the United States during the summer is on the order of
0.03 to 0.05 ppm. Reviewing data from sites that appear to be isolated from anthropogenic
sources, it has been reported that, in almost all cases, none of the sites experienced hourly
average concentrations >0.08 ppm and that the maximum hourly average concentrations were
in the range of 0.060 to 0.075 ppm. Using data from these sites, in the continental United
States, the 7-mo (April to October) average of the 7-h daily average concentrations range
from approximately 0.025 to 0.045 ppm. At an O3 monitoring site at the Theodore Roosevelt
National Park, 7-mo (April to October) averages of the 7-h daily average concentrations of
0.038, 0.039, and 0.039 ppm, respectively, were experienced in 1984, 1985, and 1986. These
7-h seasonal averages appear to be representative of the 8-h daily average O3 concentrations
that may occur at other fairly clean sites in the United States and other locations in the
northern hemisphere.
Diurnal variations are those that occur during a 24-h period. Diurnal patterns of
O3 may be expected to vary with location, depending on the balance among the many factors
affecting O3 formation, transport, and destruction. Although they vary with locality, diurnal
patterns for O3 typically show a rise in concentration from low levels or levels near minimum
detectable amounts to an early afternoon peak. The diurnal pattern of concentrations can be
ascribed to four simultaneous processes: (1) downward transport of O3 from layers aloft,
4-96
-------
(2) destruction of O3 through contact with surfaces and through reaction with NO at ground
level, (3) in situ photochemical production of O3, and (4) horizontal transport of O3 and its
precursors.
Although it might appear that composite diurnal pattern diagrams could be used to
quantify the differences in O3 exposures among sites, caution has been expressed in their use
for this purpose. The average diurnal patterns are derived from long-term calculations of the
hourly average concentrations, and the resulting diagram cannot adequately identify, at most
sites, the presence of high hourly average concentrations; thus, they may not be adequate to
distinguish O3 exposure differences among sites. Unique families of diurnal average profiles
exist, and it is possible to distinguish between two types of O3 monitoring sites. A seasonal
diurnal diagram provides the investigator with the opportunity to identify whether a specific
O3 monitoring site has more scavenging than any other site. For low-elevation sites, intraday
variability is most significant due to the pronounced daily amplitude in O3 concentration
between the predawn minimum and midafternoon to early-evening maximum, whereas
interday variation is more significant in the high-elevation sites.
Seasonal variations in O3 concentrations in urban areas usually show the pattern of
high O3 in late spring or in summer and low levels in the winter. Because of temperature,
relative humidity, and seasonal changes in storm tracks from year to year, the general weather
conditions in a given year may be more favorable for the formation of O3 and other oxidants
than during the prior or following year. For example, 1988 was a hot and dry year in which
some of the highest O3 concentrations of the last decade occurred, whereas 1989 was a cold
and wet year in which some of the lowest concentrations occurred.
Several investigators have reported on the tendency for average O3 concentrations
to be higher in the second than in the third quarter of the year for many isolated rural sites.
This observation has been attributed to either stratospheric intrusions or an increasing
frequency of slow-moving, high-pressure systems that promote the formation of O3.
However, for several clean rural sites, the highest exposures have occurred in the third quarter
rather than in the second. For rural O3 sites in the southeastern United States, the daily
maximum 1-h average concentration was found to peak during the summer months. For sites
located in rural areas, but not isolated from anthropogenic sources of pollution, the different
patterns may be associated with anthropogenic emissions of NOX and hydrocarbons.
Concentrations of O3 vary with altitude and with latitude. There appears to be no
generalizable conclusion concerning the relationship between O3 exposure and elevation. The
differences in exposure occur when one site is above the natural inversion and the other is
not. An important issue for assessing possible impacts of O3 at high-elevation sites that
requires further attention is the use of mixing ratios (e.g., parts per million) instead of
absolute concentration (e.g., in units of micrograms per cubic meter) to describe
O3 concentration. In most cases, mixing ratios, or mole fractions, are used to describe
O3 concentrations. The manner in which concentration is reported may be important when
assessing the potential impacts of air pollution on high-elevation forests. Concentration
varies as a function of altitude. Although the change in concentration is small when the
elevational difference between sea level and the monitoring site is small, it becomes
substantial at high-elevation sites. Given the same part-per-million value registering at both a
high- and low-elevation site, the absolute concentrations (i.e., micrograms per cubic meter) at
the two elevations will be different. Because both pollutants and ambient air are gases,
changes in pressure directly affect their volume. This pressure effect must be considered
when measuring absolute pollutant concentrations. Although these exposure considerations
4-97
-------
are trivial at low-elevation sites, when one compares exposure-effects results obtained at high-
elevation sites with those from low-elevation sites, the differences may become significant.
Most people in the United States spend a large proportion of their time indoors.
Until the early 1970s, very little was known about the O3 concentrations experienced inside
buildings. Even to date, the database on this subject is not extensive, and a wide range of I/O
O3 concentration relationships can be found in the literature. Reported I/O values for O3 are
highly variable. A relatively large number of factors can affect the difference in
O3 concentrations between the inside of a structure and the outside air. In general, outside air
infiltration or exchange rates, interior air circulation rates, and interior surface composition
(e.g., rugs, draperies, furniture, walls) affect the balance between replenishment and
decomposition of O3 within buildings. The I/O O3 concentration ratios generally fall in the
range of 0.1 to 0.7, and indoor concentrations of O3 will almost invariably be less than those
outdoors.
It is important that accurate estimates of both human and vegetation exposure to O3
be available for assessing the risks posed by the pollutant. Examples are provided on how
both fixed-site monitoring information and human exposure models are used to estimate risks
associated with O3 exposure.
In many cases, the upper tail of the distribution, which represents those individuals
exposed to the highest concentrations, often generates special interest because the
determination of the number of individuals who experience elevated pollutant levels can be
critical for health risk assessments. This is especially true for pollutants for which the
relationship between dose and response is highly nonlinear.
Because it is not possible to estimate population exposure, in most cases, solely
from fixed-station data, several human exposure models have been developed. Some of these
models include information on human activity patterns (i.e., the microenvironments people
visit and the times they spend there). These models also contain submodels depicting the
sources and concentrations likely to be found in each microenvironment, including indoor,
outdoor, and in-transit settings.
A subgroup that has been studied by several investigators to assess the influence of
ambient air pollution on their respiratory health and function is children attending summer
camp. Because children are predominantly outdoors and relatively active while at camp, they
provide a unique opportunity to assess the relationships between respiratory health and
function and concurrent air pollution levels. Examples are provided on the type of exposure
patterns that children experience.
A personal exposure profile can be identified by using a personal exposure
monitor. Few data are available for individuals using personal exposure monitors. Results
from a pilot study demonstrated that fixed-site ambient measurements may not adequately
represent individual exposures. Outdoor O3 concentrations showed substantial spatial
variation between rural and residential regions. The study showed that the use of fixed-site
measurements could result in an error as high as 127%. In addition, the study showed that
models based on time-weighted I/O concentrations explained only 40% of the variability in
personal exposures. The investigators concluded that contributions from diverse indoor and
outdoor microenvironments could estimate personal O3 exposure accurately.
The field of human exposure modeling is relatively young, with the first rigorous
exposure modeling analyses appearing in the mid-1970s and the theoretical constructs
regarding human exposure to environmental pollution being published in the early 1980s.
Two distinct types of O3 exposure models exist: (1) those that focus narrowly on predicting
4-98
-------
indoor O3 levels and (2) those that focus on predicting O3 exposures on a community-wide
basis. The following four distinct models address the prediction of O3 exposures on a
community-wide basis: (1) pNEM/O3, (2) SAI/NEM, (3) REHEX, and (4) EPEM. All four
O3 exposure models are derived from the NEM (NAAQS exposure model), which was first
developed in the early 1980s. Most applications of NEM use fixed-site monitoring data, U.S.
census data, and human-activity data. The calculations result in an estimate of the
O3 concentration experienced by an individual in each microenvironment that the person
inhabits.
The hourly average concentrations used in many of the high-treatment
experimental vegetation and human health effects studies did not necessarily simulate those
concentrations observed under ambient conditions. Although the ramifications of this
observation on the effects observed are not clear, it has been pointed out that the highest
treatments used in many of the vegetation open-top chamber experiments were bimodal in the
distribution of the hourly average concentrations. In other experiments designed to assess the
effects of O3 on vegetation, constant concentration (i.e., square wave) exposures were
implemented. As discussed in earlier sections of this Chapter, square wave exposure regimes
do not normally occur under ambient conditions. Similar square wave exposures have been
used in human health effects studies. In addition to the exposures used at the highest
treatment levels for vegetation experiments, there is concern that the hourly average
concentrations used in the charcoal-filtered control treatments may be lower than those
experienced at isolated sites in the United States and in other parts of the world. Although
the ramifications of using such exposure regimes is unclear, there is some concern that the
use of atypically low control levels may result in an overestimation of vegetation yield losses
when used as the baseline for evaluating the effects of treatments at higher concentrations.
Published data on the concentrations of photochemical oxidants other than O3 in
ambient air are neither comprehensive nor abundant. A review of the data shows that PAN
and PPN are the most abundant of the non-O3 oxidants in ambient air in the United States,
other than the inorganic nitrogenous oxidants such as NO2 and possibly nitric acid. At least
one study has reported that a higher homologue of the series, peroxybenzoyl nitrate (PBzN),
like PAN, is a lachrymator. No unambiguous identification of PBzN in the ambient air of the
United States has been made.
Given the information available on PAN, the concentrations of PAN that are of
most concern are those to which vegetation could potentially be exposed, especially during
daylight hours in agricultural areas. These are followed in importance by concentrations both
indoors and outdoors, in urban and nonurban areas, to which human populations potentially
could be exposed. Most of the available data on concentrations of PAN and PPN in ambient
air are from urban areas. The levels to be found in nonurban areas will be highly dependent
on the transport of PAN and PPN or their precursors from urban areas, because the
concentrations of the NOX precursors to these compounds are considerably lower in nonurban
than in urban areas.
There have been several attempts to characterize air pollutant mixtures. Pollutant
combinations can occur at or above a threshold concentration either together or temporally
separated from one another. Studies of the joint occurrence of gaseous NO2/O3 and SO2/O3
have concluded that the co-occurrence of two-pollutant mixtures lasted only a few hours per
episode, and that the time between episodes is generally long (i.e, weeks, sometimes months).
Using hourly averaged data collected at rural sites for vegetation considerations, the periods
of co-occurrence represent a small portion of the potential plant growing period. For human
4-99
-------
ambient exposure considerations, the simultaneous co-occurrence of NO2/O3 was infrequent in
most cases. However, for several sites located in the South Coast Air Basin, more than 450
simultaneous co-occurrences of each pollutant, at hourly average concentrations >0.05 ppm,
were present. Although the focus of co-occurrence research has been on patterns associated
with the presence or absence of hourly average concentrations of pollutant pairs, some
researchers have discussed the joint occurrence of O3, nitrogen, and sulfur in forested areas,
combining cumulative exposures of O3 with data on dry deposition of sulfur and nitrogen.
One study reported that several forest landscapes with the highest dry deposition loadings of
sulfur and nitrogen tended to experience the highest average O3 concentrations and the largest
cumulative exposure. Although the investigators concluded that the joint occurrences of
multiple pollutants in forest landscapes were important, nothing was mentioned about hourly
co-occurrences of O3 and SO2 or O3 and NO2.
Knowledge of the potential exposure of the co-occurrence of acidic sulfate aerosols
and O3 is limited because routine monitoring data for acidic aerosols are not available.
Information on the co-occurrence patterns is limited to research studies; some of the results
are provided in this chapter. Acid sulfates, which are composed of H2SO4, ammonium
bisulfate, and ammonium sulfate, have been measured at a number of locations in North
America. Acidic sulfate and neutralized species can accumulate and range in concentration
from 0 to 50 |ig/m3 at a specific location or a number of locations simultaneously. For many
summertime studies, peaks of H2SO4 or H+ appear to be associated with the presence of a
slow-moving high pressure system. Acid sulfates are found primarily in the fine particle size
range (<2.5 jam in diameter). The acidic sulfate concentrations measured in the summertime
can be found at 20 |ig/m3 for over an hour and at high concentrations of 10 to 20 |ig/m3 for 6
to 24 h at one or more sites. Acidic sulfate aerosol concentrations can occur at
concentrations in the summertime above 10 |ig/m3 for periods longer than 5 h. The highest
O3 exposures for sites affected by anthropogenically derived photooxidant precursors are
expected to occur during the late spring and summer months. Thus, the potential for O3 and
acidic sulfate aerosols to co-occur at some locations in some form (i.e., simultaneously,
sequentially, or complex-sequentially) is real and requires further characterization.
Concern has been expressed about the possible effects on vegetation from
co-occurring exposures of O3 and acid precipitation. One study explored the relationship
between O3 and H+ in precipitation, using data from sites that monitored both O3 and wet
deposition simultaneously and within one minute latitude and longitude of each other. The
investigators reported that individual sites experienced years in which both FT deposition and
total O3 exposure were at least moderately high (i.e., annual H+ deposition >0.5 kg ha"1 and an
annual O3 cumulative sigmoidally weighted exposure (W126) value >50 ppm-h). Based on
data compiled from all sites, relatively acidic precipitation (pH < 4.31 on a weekly basis or
pH < 4.23 on a daily basis) occurred together with relatively high O3 levels (i.e., W126 values
> 0.66 ppm-h for the same week or W126 values > 0.18 ppm-h immediately before or after a
rainfall event) approximately 20% of the time, and highly acidic precipitation (i.e, pH < 4.10
on a weekly basis or pH < 4.01 on a daily basis) occurred together with a high O3 level (i.e.,
W126 values > 1.46 ppm-h for the same week or W126 values >0.90 ppm-h immediately
before or after a rainfall event) approximately 6% of the time. Whether during the same
week or before, during, or after a precipitation event, correlations between O3 level and pH
(or FT deposition) were weak to nonexistent. Sites most subject to relatively high levels of
both FT and O3 were located in the eastern portion of the United States, often in mountainous
areas.
4-100
-------
The co-occurrence of O3 and acidic cloudwater in high-elevation forests has been
characterized. The frequent O3-only and pH-only, single-pollutant episodes, as well as the
simultaneous and sequential co-occurrences of O3 and acidic cloudwater, have been reported.
Both simultaneous and sequential co-occurrences were observed a few times each month
above the cloud base.
4-101
-------
References
Adams, R. M.; Hamilton, S. A.; McCarl, B. A. (1985) An assessment of the economic effects of ozone on U.S.
agriculture. J. Air Pollut. Control Assoc. 35: 938-943.
Adams, R. M.; Glyer, J. D.; Johnson, S. L.; McCarl, B. A. (1989) A reassessment of the economic effects of
ozone on U.S. agriculture. JAPCA 39: 960-968.
Altshuller, A. P. (1983) Measurements of the products of atmospheric photochemical reactions in laboratory
studies and in ambient air-relationships between ozone and other products. Atmos. Environ.
17: 2383-2427.
Altshuller, A. P. (1986) The role of nitrogen oxides in nonurban ozone formation in the planetary boundary layer
over N America, W Europe and adjacent areas of ocean. Atmos. Environ. 20: 245-268.
Altshuller, A. P. (1989) Sources and levels of background ozone and its precursors and impact at ground level.
In: Schneider, T.; Lee, S. D.; Wolters, G. J. R.; Grant, L. D., eds. Atmospheric ozone research and
its policy implications: proceedings of the 3rd US-Dutch international symposium; May 1988;
Nijmegen, The Netherlands. Amsterdam, The Netherlands: Elsevier Science Publishers; pp. 127-157.
(Studies in environmental science: 35).
Altshuller, A. P.; Lefohn, A. S. (1996) Background ozone in the planetary boundary layer over the United States.
J. Air Waste Manage. Assoc. 46: 134-141.
Aneja, V. P.; Li, Z. (1992) Characterization of ozone at high elevation in the eastern United States: trends,
seasonal variations, and exposure. J. Geophys. Res. [Atmos.] 97: 9873-9888.
Austin, B. S.; Andersen, G. E.; Weir, B. R.; Seigneur, C. (1986) Further development and application of an
improved population exposure model. San Rafael, CA: Systems Applications, Inc.
Avol, E. L.; Trim, S. C.; Little, D. E.; Spier, C. E.; Smith, M. N.; Peng, R.-C.; Linn, W. S.; Hackney, J. D.;
Gross, K. B.; D'Arcy, J. B.; Gibbons, D.; Higgins, I. T. T. (1990) Ozone exposure and lung
function in children attending a southern California summer camp. Presented at: 83rd annual
meeting and exhibition of the Air & Waste Management Association; June; Pittsburgh, PA.
Pittsburgh, PA: Air & Waste Management Association; paper no. 90-150.3.
Berglund, R. L.; Dittenhoefer, A. C.; Ellis, H. M.; Watts, B. J.; Hansen, J. L. (1988) Evaluation of the stringency
of alternative forms of a national ambient air quality standard for ozone. In: Wolff, G. T.; Hanisch,
J. L.; Schere, K., eds. Transactions of an APCA international specialty conference on the scientific
and technical issues facing post-1987 ozone control strategies; November, 1987; Hartford, CT.
Pittsburgh, PA: Air & Waste Management Association; pp. 343-369.
Berk, J. V.; Young, R.; Hollowell, C. D.; Turiel, I.; Pepper, J. (1980) The effects of energy-efficient ventilation
rates on indoor air quality at an Ohio elementary school. Berkeley, CA: University of California,
Lawrence Berkeley Laboratory; report no. EEB-VENT 80-9. Available from: NTIS, Springfield,
VA; LBL-10223.
Berk, J. V.; Young, R. A.; Brown, S. R.; Hollowell, C. D. (1981) Impact of energy-conserving retrofits on
indoor air quality in residential housing. Presented at: 74th annual meeting of the Air Pollution
Control Association; June; Philadelphia, PA. Pittsburgh, PA: Air Pollution Control Association;
paper no. 81-22.1.
Berry, C. R. (1964) Differences in concentrations of surface oxidant between valley and mountaintop conditions
in the southern Appalachians. J. Air Pollut. Control Assoc. 14: 238-239.
4-102
-------
Biller, W. F.; Feagans, T. B.; Johnson, T. R.; Duggan, G. M; Paul, R. A.; McCurdy, T.; Thomas, H. C. (1981)
A general model for estimating exposures associated with alternative NAAQS. Presented at: 74th
annual meeting of the Air Pollution Control Association; June; Philadelphia, PA. Pittsburgh, PA:
Air Pollution Control Association.
Bohm, M.; McCune, B.; Vandetta, T. (1991) Diurnal curves of troposheric ozone in the western United States.
Atmos. Environ. Part A 25: 1577-1590.
Bojkov, R. D. (1986) Surface ozone during the second half of the nineteenth century. J. Clim. Appl. Meteorol.
25: 343-352.
Bower, J. S.; Broughton, G. F. J.; Dando, M. T.; Stevenson, K. J.; Lampert, J. E.; Sweeney, B. P.; Parker, V. J.;
Driver, G. S.; Clark, A. G.; Waddon, C. J.; Wood, A. J.; Williams, M. L. (1989) Surface ozone
concentrations in the U. K. in 1987-1988. Atmos. Environ. 23: 2003-2016.
Buhr, M. P.; Parrish, D. D.; Norton, R. B.; Fehsenfeld, F. C.; Sievers, R. E.; Roberts, J. M. (1990) Contribution
of organic nitrates to the total reactive nitrogen budget at a rural eastern U.S. site. J. Geophys. Res.
[Atmos.] 95: 9809-9816.
California Air Resources Board. (1992) Technical support document for proposed amendments to the criteria for
designating areas of California as nonattainment, attainment, or unclassified for state ambient air
quality standards. Sacramento, CA: California Environmental Protection Agency.
Cartalis, C.; Varotsos, C. (1994) Surface ozone in Athens, Greece, at the beginning and at the end of the
twentieth century. Atmos. Environ. 28: 3-8.
Cass, G. R.; Nazaroff, W. W.; Tiller, C.; Whitmore, P. M. (1991) Protection of works of art from damage due to
atmospheric ozone. Atmos. Environ. Part A 25: 441-451.
Chameides, W. L.; Lindsay, R. W.; Richardson, J.; Kiang, C. S. (1988) The role of biogenic hydrocarbons in
urban photochemical smog: Atlanta as a case study. Science (Washington, DC) 241: 1473-1475.
Chan, C.-C.; Ozkaynak, H.; Spengler, J. D.; Sheldon, L. (1991) Driver exposure to volatile organic compounds,
CO, ozone and NO2 under different driving conditions. Environ. Sci. Technol. 25: 964-972.
Cleveland, W. S.; Guarino, R.; Kleiner, B.; McRae, J. E.; Warner, J. L. (1976b) The analysis of the ozone
problem in the northeast United States. In: Specialty conference on: ozone/oxidants-interactions with
the total environment; March; Dallas, TX. Pittsburgh, PA: Air Pollution Control Association; pp.
109-119.
Cleveland, W. S.; Kleiner, B.; McRae, J. E.; Warner, J. L. (1976a) Photochemical air pollution: transport from
New York City area into Connecticut and Massachusetts. Science (Washington, DC) 191: 179-181.
Code of Federal Regulations. (1991) National primary and secondary ambient air quality standards. C. F. R. 40:
§50.
4-103
-------
Coffey, P.; Stasiuk, W.; Mohnen, V. (1977) Ozone in rural and urban areas of New York State: part I.
In: Dimitriades, B., ed. International conference on photochemical oxidant pollution and its control -
proceedings: volume I; September 1976; Raleigh, NC. Research Triangle Park, NC:
U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory; pp. 89-96;
report no. EPA-600/3-77-001a. Available from: NTIS, Springfield, VA; PB-264232.
Contant, C. F., Jr.; Gehan, B. M; Stock, T. H.; Holguin, A. H.; Buffler, P. A. (1985) Estimation of individual
ozone exposures using microenvironment measurements. In: Lee, S. D., ed. Transactions: evaluation
of the scientific basis for ozone/oxidants standards. An APCA international specialty conference;
November, 1984; Houston, TX. Pittsburgh, PA: Air Pollution Control Association; pp. 250-261.
Corkum, R.; Giesbrecht, W. W.; Bardsley, T.; Cherniak, E. A. (1986) Peroxyacetyl nitrate (PAN) in the
atmosphere at Simcoe, Canada. Atmos. Environ. 20: 1241-1248.
Cox, W. M.; Chu, S.-H. (1993) Meteorologically adjusted ozone trends in urban areas: a probabilistic approach.
Atmos. Environ. Part B 27: 425-434.
Curran, T. C.; Frank, N. H. (1991) Ambient ozone trends using alternative indicators. In: Berglund, R. L.;
Lawson, D. R.; McKee, D. I, eds. Tropospheric ozone and the environment: papers from an
international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
Association; pp. 749-759. (A&WMA transactions series no. TR-19).
Daum, P. H.; Kleinman, L. L; Hills, A. I; Lazrus, A. L.; Leslie, A. C. D.; Busness, K.; Boatman, J. (1990)
Measurement and interpretation of concentrations of H2O2 and related species in the upper midwest
during summer. J. Geophys. Res. [Atmos.] 95: 9857-9871.
Davidson, A. (1993) Update of ozone trends in California's South Coast Air Basin. Air Waste 43: 226-227.
Davies, T. D.; Ramer, B.; Kaspyzok, G.; Delany, A. C. (1984) Indoor/outdoor ozone concentrations at a
contemporary art gallery. J. Air Pollut. Control Assoc. 31: 135-137.
Druzik, J. R.; Adams, M. S.; Tiller, C.; Cass, G. R. (1990) The measurement and model predictions of indoor
ozone concentrations in museums. Atmos. Environ. Part A 24: 1813-1823.
Duan, N. (1982) Models for human exposure to air pollution. Environ. Int. 8: 305-309.
Edgerton, E. S.; Lavery, T. F. (1992) National dry deposition network fourth annual progress report, 1990.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and
Exposure Assessment Laboratory; contract no. 68-02-4451.
Evans, G.; Finkelstein, P.; Martin, B.; Possiel, N.; Graves, M. (1983) Ozone measurements from a network of
remote sites. J. Air Pollut. Control Assoc. 33: 291-296.
Evans, E. G.; Rhodes, R. C.; Mitchell, W. J.; Puzak, J. C. (1985) Summary of precision and accuracy
assessments for the state and local air monitoring networks: 1982. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
EPA-600/4-85-031. Available from: NTIS, Springfield, VA; PB85-208171/HSU.
Fahey, D. W.; Hubler, G.; Parrish, D. D.; Williams, E. J.; Norton, R. B.; Ridley, B. A.; Singh, H. B.; Liu, S. C.;
Fehsenfeld, F. C. (1986) Reactive nitrogen species in the troposphere: measurements of NO, NO2,
HNO3, paniculate nitrate, peroxyacetyl nitrate (PAN), O3, and total reactive odd nitrogen (NOy) at
Niwot Ridge, Colorado. J. Geophys. Res. [Atmos.] 91: 9781-9793.
4-104
-------
Faith, R.; Sheshinski, R. (1979) Misspecification of trend in spatial random-function interpolation with
application to oxidant mapping. Palo Alto, CA: SIAM Institute for Mathematics and Society,
Stanford University; technical report no. 28.
Finlayson-Pitts, B. I; Pitts, J. N., Jr. (1986) Atmospheric chemistry: fundamentals and experimental techniques.
New York, NY: John Wiley & Sons.
Fowler, D.; Cape, J. N. (1982) Air pollutants in agriculture and horticulture. In: Unsworth, M. H.; Ormrod, D.
P., eds. Effects of gaseous air pollution in agriculture and horticulture. London, United Kingdom:
Butterworth Scientific; pp. 3-26.
Gaffney, J. S.; Marley, N. A.; Prestbo, E. W. (1993) Measurements of peroxyacetyl nitrate at a remote site in the
southwestern United States: tropospheric implications. Environ. Sci. Technol. 27: 1905-1910.
Gallopoulos, N. E.; Heuss, J. M.; Chock, D. P.; Dunker, A. M.; Williams, R. L.; Wolff, G. T. (1988) Progress
and prospects in air pollution control and the need to strengthen its scientific base. In: Wolff, G. T.;
Hanisch, J. L.; Schere, K., eds. The scientific and technical issues facing post-1987 ozone control
stategies: transactions of an APCA international specialty conference; November 1987; Hartford,
CT. Pittsburgh, PA: Air & Waste Management Association; pp. 30-45.
Grivet, C. D. (1980) Modeling and analysis of air quality data. Palo Alto, CA: SIAM Institute for Mathematics
and Society, Stanford University; technical report no. 43.
Grosjean, D.; Williams, E. L., II (1992) Photochemical pollution at two southern California smog receptor sites.
J. Air Waste Manage. Assoc. 42: 805-809.
Grosjean, D.; Williams, E. L., II; Grosjean, E. (1993) Peroxyacyl nitrates at southern California mountain forest
locations. Environ. Sci. Technol. 27: 110-121.
Hales, C. H.; Rollinson, A. M.; Shair, F. H. (1974) Experimental verification of linear combination model for
relating indoor-outdoor pollutant concentrations. Environ. Sci. Technol. 8: 452-453.
Hayes, S. R. (1989) Estimating the effect of being indoors on total personal exposure to outdoor air pollution.
JAPCA 39: 1453-1461.
Hayes, S. R. (1991) Use of an indoor air quality model (IAQM) to estimate indoor ozone levels. J. Air Waste
Manage. Assoc. 41: 161-170.
Hayes, S. R.; Lundberg, G. W. (1985) Further improvement and sensitivity analysis of an ozone population
exposure model. San Rafael, CA: Systems Applications, Inc.
Hayes, S. R.; Rosenbaum, A. S. (1988) An examination of acute ozone health risk and exposure effects on
proposed SCAQMD rule 1109 using the SOCAB 6-7 June 1985 ozone episode. San Rafael, CA:
Systems Applications, Inc.
Hayes, S. R.; Seigneur, C.; Lundberg, G. W. (1984) Numerical modeling of ozone population exposure:
application to a comparison of alternative ozone standards. San Rafael, CA: Systems Applications,
Inc.; final report no. SYSAPP-84/140.
Hayes, S. R.; Austin, B. S.; Rosenbaum, A. S. (1988) A technique for assessing the effects of ROG and NOX
reductions on acute ozone exposure and health risk in the south coast air basin. San Rafael, CA:
Systems Applications, Inc.
4-105
-------
Heck, W. W.; Taylor, O. C.; Adams, R.; Bingham, G.; Miller, I; Preston, E.; Weinstein, L. (1982) Assessment
of crop loss from ozone. J. Air Pollut. Control Assoc. 32: 353-361.
Heuss, J. M. (1982) Comment on "Assessment of crop loss from ozone." J. Air Pollut. Control Assoc.
32: 1152-1153.
Hong, S.; Candelone, J.-P.; Patterson, C. C.; Boutron, C. F. (1994) Greenland ice evidence of hemispheric lead
pollution two millennia ago by Greek and Roman civilizations. Science (Washington, DC) 265:
1841-1843.
Horstman, D. H.; Folinsbee, L. J.; Ives, P. J.; Abdul-Salaam, S.; McDonnell, W. F. (1990) Ozone concentration
and pulmonary response relationships for 6.6-hour exposures with five hours of moderate exercise
to 0.08, 0.10, and 0.12 ppm. Am. Rev. Respir. Dis. 142: 1158-1163.
Husar, R. B.; Patterson, D. E.; Paley, C. C.; Gillani, N. V. (1977) Ozone in hazy air masses. In: Dimitriades, B.,
ed. International conference on photochemical oxidant pollution and its control. Proceedings:
volume I; September 1976; Raleigh, NC. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Environmental Sciences Research Laboratory; pp. 275-282; report no.
EPA-600/3-77-001a. Available from: NTIS, Springfield, VA; PB-264232.
Isaksen, I. S. A.; Hov, O.; Hesstvedt, E. (1978) Ozone generation over rural areas. Environ. Sci. Technol.
12: 1279-1284.
Johnson, T. R.; Paul, R. A.; Capel, J. E.; McCurdy, T. (1990) Estimation of ozone exposure in Houston using a
probabilistic version of NEM. Presented at: 83rd annual meeting and exhibition of the Air & Waste
Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management
Association; paper no. 90-150.1.
Johnson, T. R.; Wijnberg, L.; Capel, J. E.; Vostal, J. J. (1992) The use of activity diary data to estimate the
probability of exposure to air pollution. In: Berglund, R. L., ed. Tropospheric ozone and the
environment II: effects, modeling and control [papers from an international specialty conference];
November; Atlanta, GA. Pittsburgh, PA: Air & Waste Management Association; pp. 713-724.
(A&WMA transactions series no. 20).
Jones, K.; Militana, L.; Martini, J. (1989) Ozone trend analysis for selected urban areas in the continental U.S.
Presented at: 82nd annual meeting and exhibition of the Air & Waste Management Association;
June; Anaheim, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 89-3.6.
Kelly, N. A.; Wolff, G. T.; Ferman, M. A. (1982) Background pollutant measurements in air masses affecting
the eastern half of the United States -1. air masses arriving from the northwest. Atmos. Environ.
16: 1077-1088.
Kelly, N. A.; Wolff, G. T.; Ferman, M. A. (1984) Sources and sinks of ozone in rural areas. Atmos. Environ. 18:
1251-1266.
Kelly, N. A.; Ferman, M. A.; Wolff, G. T. (1986) The chemical and meteorological conditions associated with
high and low ozone concentrations in southeastern Michigan and nearby areas of Ontario. J. Air
Pollut. Control Assoc. 36: 150-158.
Kleindienst, T. E.; Shepson, P. B.; Hodges, D. N.; Nero, C. M.; Arnts, R. R.; Dasgupta, P. K.; Hwang, H.; Kok,
G. L.; Lind, J. A.; Lazrus, A. L.; Mackay, G. L; Mayne, L. K.; Schiff, H. I. (1988) Comparison of
techniques for measurement of ambient levels of hydrogen peroxide. Environ. Sci. Technol. 22:
53-61.
4-106
-------
Knudsen, H. P.; Lefohn, A. S. (1988) The use of geostatistics to characterize regional ozone exposures. In: Heck,
W. W.; Taylor, O. C.; Tingey, D. T., eds. Assessment of crop loss from air pollutants: proceedings
of an international conference; October 1987; Raleigh, NC. Essex, United Kingdom: Elsevier
Science Publishers, Ltd.; pp. 91-105.
Kolaz, D. J.; Swinford, R. L. (1990) How to remove the influence of meteorology from the Chicago area ozone
trend. Presented at: 83rd annual meeting and exhibition of the Air & Waste Management
Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; paper no.
90-97.5.
Korsog, P. E.; Wolff, G. T. (1991) An examination of urban ozone trends in the northeastern U.S. (1973-1983)
using a robust statistical method. Atmos. Environ. Part B 25: 47-57.
Kuntasai, G.; Chang, T. Y. (1987) Trends and relationships of O3, NOX and HC in the south coast air basin of
California. JAPCA 37: 1158-1163.
Lawson, D. R.; Dasgupta, P. K.; Kok, G. L.; Kaplan, I. R.; Sakugawa, H.; Mackay, G. I. (1988) Interlaboratory
comparison of ambient hydrogen peroxide measurements in Los Angeles. In: Proceedings of the
196th American Chemical Society national meeting; September; Los Angeles, CA. Washington, DC:
American Chemical Society, Division of Environmental Chemistry.
Lebowitz, M. D.; Gorman, G.; O'Rourke, M. K.; Holberg, C. J. (1984) Indoor-outdoor air pollution, allergen and
meteorological monitoring in an arid southwest area. J. Air Pollut. Control Assoc. 34: 1035-1038.
Lefohn, A. S. (1984) A comparison of ambient ozone exposures for selected nonurban sites. Presented at: 77th
annual meeting of the Air Pollution Control Association; June; San Francisco, CA. Pittsburgh, PA:
Air Pollution Control Association; paper no. 84-104.1.
Lefohn, A. S. (1992a) Ozone standards and their relevance for protecting vegetation. In: Lefohn, A. S., ed.
Surface level ozone exposures and their effects on vegetation. Chelsea, MI: Lewis Publishers, Inc.;
pp. 325-359.
Lefohn, A. S. (1992b) The characterization of ambient ozone exposures. In: Lefohn, A. S., ed. Surface level
ozone exposures and their effects on vegetation. Chelsea, MI: Lewis Publishers, Inc.; pp. 31-92.
Lefohn, A. S.; Benedict, H. M. (1983) The potential for the interaction of acidic precipitation and ozone pollutant
doses affecting agricultural crops. Presented at: 76th annual meeting of the Air Pollution Control
Association; June; Atlanta, GA. Pittsburgh, PA: Air Pollution Control Association; report no. 83-2.2.
Lefohn, A. S.; Foley, J. K. (1992) NCLAN results and their application to the standard-setting process:
protecting vegetation from surface ozone exposures. J. Air Waste Manage. Assoc. 42: 1046-1052.
Lefohn, A. S.; Foley, J. K. (1993) Establishing relevant ozone standards to protect vegetation and human health:
exposure/dose-response considerations. Air Waste 43: 106-112.
Lefohn, A. S.; Jones, C. K. (1986) The characterization of ozone and sulfur dioxide air quality data for assessing
possible vegetation effects. J. Air Pollut. Control Assoc. 36: 1123-1129.
Lefohn, A. S.; Lucier, A. A. (1991) Spatial and temporal variability of ozone exposure in forested areas of the
United States and Canada: 1978-1988. J. Air Waste Manage. Assoc. 41: 694-701.
Lefohn, A. S.; Mohnen, V. A. (1986) The characterization of ozone, sulfur dioxide, and nitrogen dioxide for
selected monitoring sites in the Federal Republic of Germany. J. Air Pollut. Control Assoc.
36: 1329-1337.
4-107
-------
Lefohn, A. S.; Runeckles, V. C. (1987) Establishing standards to protect vegetation - ozone exposure/dose
considerations. Atmos. Environ. 21: 561-568.
Lefohn, A. S.; Shadwick, D. S. (1991) Ozone, sulfur dioxide, and nitrogen dioxide trends at rural sites located in
the United States. Atmos. Environ. Part A 25: 491-501.
Lefohn, A. S.; Tingey, D. T. (1984) The co-occurrence of potentially phytotoxic concentrations of various
gaseous air pollutants. Atmos. Environ. 18: 2521-2526.
Lefohn, A. S.; Knudsen, H. P.; Logan, J. A.; Simpson, J.; Bhumralkar, C. (1987a) An evaluation of the kriging
method to predict 7-h seasonal mean ozone concentrations for estimating crop losses. JAPCA 37:
595-602.
Lefohn, A. S.; Davis, C. E.; Jones, C. K.; Tingey, D. T.; Hogsett, W. E. (1987b) Co-occurrence patterns of
gaseous air pollutant pairs at different minimum concentrations in the United States.
Atmos. Environ. 21: 2435-2444.
Lefohn, A. S.; Knudsen, H. P.; McEvoy, L. R., Jr. (1988a) The use of kriging to estimate monthly ozone
exposure parameters for the southeastern United States. Environ. Pollut. 53: 27-42.
Lefohn, A. S.; Laurence, J. A.; Kohut, R. J. (1988b) A comparison of indices that describe the relationship
between exposure to ozone and reduction in the yield of agricultural crops. Atmos. Environ.
22: 1229-1240.
Lefohn, A. S.; Runeckles, V. C.; Krupa, S. V.; Shadwick, D. S. (1989) Important considerations for establishing
a secondary ozone standard to protect vegetation. JAPCA 39: 1039-1045.
Lefohn, A. S.; Krupa, S. V.; Winstanley, D. (1990a) Surface ozone exposures measured at clean locations around
the world. Environ. Pollut. 63: 189-224.
Lefohn, A. S.; Shadwick, D. S.; Mohnen, V. A. (1990b) The characterization of ozone concentrations at a select
set of high-elevation sites in the eastern United States. Environ. Pollut. 67: 147-178.
Lefohn, A. S.; Benkovitz, C. M; Tanner, R. L.; Smith, L. A.; Shadwick, D. S. (1991) Air quality measurements
and characterizations for terrestrial effects research. In: Irving, P. M., ed. Acidic deposition: state of
science and technology, volume I, emissions, atmospheric processes, and deposition. Washington,
DC: The U.S. National Acid Precipitation Assessment Program. (State of science and technology
report no. 7).
Lefohn, A. S.; Shadwick, D. S.; Somerville, M. C.; Chappelka, A. H.; Lockaby, B. G.; Meldahl, R. S. (1992a)
The characterization and comparision of ozone exposure indices used in assessing the response of
loblolly pine to ozone. Atmos. Environ. Part A 26: 287-298.
Lefohn, A. S.; Knudsen, H. P.; Shadwick, D. S.; Hermann, K. A. (1992b) Surface ozone exposures in the eastern
United States (1985-1989). In: Flagler, R. B., ed. The response of southern commercial forests to
air pollution: papers from an international specialty conference; November 1991; Atlanta, GA.
Pittsburgh, PA: Air & Waste Management Association; pp. 81-93. (A&WMA transactions series:
no. 21).
Lefohn, A. S.; Shadwick, D. S.; Feister, U.; Mohnen, V. A. (1992c) Surface-level ozone: climate change and
evidence for trends. J. Air Waste Manage. Assoc. 42: 136-144.
Lefohn, A. S.; Foley, J. K.; Shadwick, D. S.; Tilton, B. E. (1993a) Changes in diurnal patterns related to changes
in ozone levels. Air Waste 43: 1472-1478.
4-108
-------
Lefohn, A. S.; Foley, J. K.; Tilton, B. E. (1993b) Changes in ozone concentration regimes as a function of
change in site attainment status. In: Vostal, J. J., ed. Tropospheric ozone: nonattainment and design
value issues [proceedings of a U.S. EPA/A&WMA international specialty conference]; October
1992; Boston, MA. Pittsburgh, PA: Air & Waste Management Association; pp. 477-488. (A&WMA
transactions series no. 23).
Lindsay, R. W.; Chameides, W. L. (1988) High-ozone events in Atlanta, Georgia, in 1983 and 1984. Environ.
Sci. Technol. 22: 426-431.
Lindsay, R. W.; Richardson, J. L.; Chameides, W. L. (1989) Ozone trends in Atlanta, Georgia: have emission
controls been effective? JAPCA 39: 40-43.
Lioy, P. J. (1989) Exposure assessment of oxidant gases and acidic aerosols. Annu. Rev. Public Health
10: 69-84.
Lioy, P. J.; Dyba, R. V. (1989) The dynamics of human exposure to tropospheric ozone. In: Schneider, T.; Lee,
S. D.; Wolters, G. J. R.; Grant, L. D., eds. Atmospheric ozone research and its policy implications:
proceedings of the 3rd US-Dutch international symposium; May 1988; Nijmegen, The Netherlands.
Amsterdam, The Netherlands: Elsevier Science Publishers; pp. 711-721. (Studies in environmental
science 35).
Lioy, P. J.; Waldman, J. M. (1989) Acidic sulfate aerosols: characterization and exposure. In: Symposium on the
health effects of acid aerosols; October 1987; Research Triangle Park, NC. Environ. Health
Perspect. 79: 15-34.
Lisac, I.; Grubisic, V. (1991) An analysis of surface ozone data measured at the end of the 19th century in
Zagreb, Yugoslavia. Atmos. Environ. Part A 25: 481-486.
Liu, L.-J. S.; Koutrakis, P.; Suh, H. H.; Mulik, J. D.; Burton, R. M. (1993) Use of personal measurements for
ozone exposure assessment: a pilot study. Environ. Health Perspect. 101: 318-324.
Lloyd, A. C.; Lents, J. M.; Green, C.; Nemeth, P. (1989) Air quality management in Los Angeles: perspectives
on past and future emission control strategies. JAPCA 39: 696-703.
Logan, J. A. (1985) Tropospheric ozone: seasonal behavior, trends, and anthropogenic influence. J. Geophys.
Res. [Atmos.] 90: 10,463-10,482.
Logan, J. A. (1989) Ozone in rural areas of the United States. J. Geophys. Res. [Atmos.] 94: 8511-8532.
Logan, J. A. (1994) Trends in the vertical distribution of ozone: an analysis of ozonesonde data. J. Geophys. Res.
[Atmos.] 99: 25,553-25,585.
Lurmann, F.; Colome, S. (1991) An assessment of current and future human exposure to ozone in the South
Coast Air Basin. Santa Rosa, CA: Sonoma Technology, Inc.; report no. STI-90070-1022-FR;
1991 AQMP revision, technical report V-H.
4-109
-------
Lurmann, F. W.; Coyner, L.; Winer, A. M; Colome, S. (1989) Development of a new regional human exposure
(REHEX) model and its application to the California South Coast Air Basin. Presented at: 82nd
annual meeting and exhibition of the Air & Waste Management Association; June; Anaheim, CA.
Pittsburgh, PA: Air & Waste Management Association; paper no. 89-27.5.
Lurmann, F. W.; Winer, A. M.; Colome, S. D. (1990) Development and application of a new regional human
exposure (REHEX) model. In: Total exposure assessment methodology: proceedings of the
EPA/A&WMA specialty conference; November 1989; Las Vegas, NV. Pittsburgh, PA: Air & Waste
Management Association; pp. 478-498. (A&WMA publication VTP-16).
Mage, D. T.; Raizenne, M.; Spengler, J. (1985) The assessment of individual human exposures to ozone in a
health study. In: Lee, S. D., ed. Evaluation of the scientific basis for ozone/oxidants standards:
transactions of an APCA international specialty conference; November 1984; Houston, TX.
Pittsburgh, PA: Air Pollution Control Association; pp. 238-249; report no. TR-4.
Matheron, G. (1963) Principles of geostatistics. Econ. Geol. 58: 1246-1266.
McCurdy, T. (1988) Relationships among ozone air quality indicators in urban areas. In: Wolff, G. T.; Hanisch,
J. L.; Schere, K., eds. The scientific and technical issues facing post-1987 ozone control strategies;
transactions of an APCA specialty conference; November 1987; Hartford, CT. Pittsburgh, PA: Air
& Waste Management Association; pp. 331-342.
McCurdy, T. R. (1994) Human exposure to ambient ozone. In: McKee, D. J., ed. Tropospheric ozone: human
health and agricultural impacts. Boca Raton, FL: Lewis Publishers; pp. 85-127.
McCurdy, T.; Capel, J. E.; Paul, R. A.; Johnson, T. (1991) Preliminary analysis of ozone exposures in Houston
using pNEM/O3. Presented at: 84th annual meeting and exhibition of the Air & Waste Management
Association; June; Vancouver, BC, Canada. Pittsburgh, PA: Air & Waste Management Association;
paper no. 91-141.1.
McDonnell, W. F.; Kehrl, H. R.; Abdul-Salaam, S.; Ives, P. J.; Folinsbee, L. J.; Devlin, R. B.; O'Neil, J. J.;
Horstman, D. H. (1991) Respiratory response of humans exposed to low levels of ozone for
6.6 hours. Arch. Environ. Health 46: 145-150.
Meagher, J. F.; Lee, N. T.; Valente, R. J.; Parkhurst, W. J. (1987) Rural ozone in the southeastern United States.
Atmos. Environ. 21: 605-615.
Meyer, B. (1983) Indoor air quality. Reading, MA: Addison-Wesley Publishing Company, Inc.
Mohnen, V. A. (1989) Mountain cloud chemistry project: wet, dry and cloud water deposition. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment
Laboratory; report no. EPA/600/3-89/009. Available from: NTIS, Springfield, VA; PB89-148597.
Mohnen, V. A.; Kadlecek, J. A. (1989) Cloud chemistry research at Whiteface Mountain. Tellus Ser.
B 41B: 79-91.
Mohnen, V. A.; Hogan, A.; Coffey, P. (1977) Ozone measurements in rural areas. J. Geophys. Res.
82: 5889-5895.
Moschandreas, D. J.; Stark, J. W. C.; McFadden, J. E.; Morse, S. S. (1978) Indoor air pollution in the residential
environment: volume I. data collection, analysis, and interpretation. Research Triangle Park, NC:
U.S. Environmental Protection Agency, Environmental Monitoring and Support Laboratory; report
no. EPA-600/7-78-229a. Available from: NTIS, Springfield, VA; PB-290999.
4-110
-------
Moschandreas, D. I; Zabransky, J.; Pelton, D. J. (1981) Comparison of indoor and outdoor air quality.
Palo Alto, CA: Electric Power Research Institute; EPRI report no. EA-1733.
National Acid Precipitation Assessment Program. (1987) Interim assessment: the causes and effects of acidic
deposition, volume I, executive summary. Washington, DC: Office of the Director of Research.
National Research Council. (1991) Rethinking the ozone problem in urban and regional air pollution.
Washington, DC: National Academy Press.
Nazaroff, W. W.; Cass, G. R. (1986) Mathematical modeling of chemically reactive pollutants in indoor air.
Environ. Sci. Technol. 20: 924-934.
O'Gara, P. J. (1922) Sulfur dioxide and fume problems and their solution, (presented at the 14th semiannual
meeting of the American Institute of Chemical Engineers, Niagara Falls, Canada, June 19-22, 1922).
Presented at: 14th semiannual meeting of the American Institute of Chemical Engineers; June;
Niagara Falls, Canada. Proceedings and papers of the meeting summarized in J. Ind. Eng. Chem.
14: 744-745 by J. C. Olsen.
Olszyna, K. J.; Meagher, J. F.; Bailey, E. M. (1988) Gas-phase, cloud and rain-water measurements of hydrogen
peroxide at a high-elevation site. Atmos. Environ. 22: 1699-1706.
Olszyk, D. M.; Bytnerowicz, A.; Takemoto, B. K. (1989) Photochemical oxidant pollution and vegetation: effects
of gases, fog, and particles. Environ. Pollut. 61: 11-29.
Paul, R. A.; Biller, W. F.; McCurdy, T. (1987) National estimates of population exposure to ozone. Presented at:
80th annual meeting of the Air Pollution Control Association; June; New York, NY. Pittsburgh, PA:
Air Pollution Control Association; paper no. 87-42.7.
Peake, E.; Sandhu, H. S. (1983) The formation of ozone and peroxyacetyl nitrate (PAN) in the urban
atmospheres of Alberta. Can. J. Chem. 61: 927-935.
Peake, E.; MacLean, M. A.; Sandhu, H. S. (1983) Surface ozone and peroxyacetyl nitrate (PAN) observations at
rural locations in Alberta, Canada. J. Air Pollut. Control Assoc. 33: 881-883.
Peake, E.; MacLean, M. A.; Lester, P. F.; Sandhu, H. S. (1988) Peroxyacetylnitrate (PAN) in the atmosphere of
Edmonton, Alberta, Canada. Atmos. Environ. 22: 973-981.
Pedersen, U.; Lefohn, A. S. (1994) Characterizing surface ozone concentrations in Norway. Atmos. Environ. 28:
89-101.
Pinkerton, J. E.; Lefohn, A. S. (1986) Characterization of ambient ozone concentrations in commercial
timberlands using available monitoring data. Tappi J. 69: 58-62.
Pratt, G. C.; Hendrickson, R. C.; Chevone, B. I.; Christopherson, D. A.; O'Brien, M. V.; Krupa, S. V. (1983)
Ozone and oxides of nitrogen in the rural upper-midwestern U.S.A. Atmos. Environ. 17: 2013-2023.
Prinz, B.; Krause, G. H. M. (1988) State of scientific discussion about the causes of the novel forest decline in
the Federal Republic of Germany and surrounding countries. Presented at: 15th international
meeting for specialists in air pollution effects on forest ecosystems; October; Interlaken,
Switzerland. IUFRO.
4-111
-------
Prinz, B.; Krause, G. H. M; Jung, K.-D. (1985) Untersuchungen der LIS Essen zur Problematik der
Waldschaden [Investigations of the LIS Essen into the problem of forest damage].
In: Kortzfleisch, G., ed. Waldschaden: Theorie und Praxis auf der Suche nach Antworten
[Forest damage: theory and practice in the search for answers]. Mtinchen, Germany: R. Oldenbourg
Verlag; pp. 143-194.
Raizenne, M. E.; Spengler, J. D. (1989) Dosimetric model of acute health effects of ozone and acid aerosols in
children. In: Schneider, T.; Lee, S. D.; Wolters, G. J. R.; Grant, L. D., eds. Atmospheric ozone
research and its policy implications: proceedings of the 3rd US-Dutch international symposium;
May 1988; Nijmegen, The Netherlands. Amsterdam, The Netherlands: Elsevier Science Publishers;
pp. 319-329. (Studies in environmental science 35).
Rao, S. T.; Sistla, G.; Henry, R. (1992) Statistical analysis of trends in urban ozone air quality. J. Air Waste
Manage. Assoc. 42: 1204-1211.
Reagan, J. (1984) Air quality data interpretation. In: Heck, W. W.; Taylor, O. C.; Adams, R. M.; Bingham, G.
E.; Miller, J. E.; Preston, E. M.; Weinstein, L. H., eds. National Crop Loss Assessment Network
(NCLAN) 1982 annual report. Corvallis, OR: U.S. Environmental Protection Agency, Corvallis
Environmental Research Laboratory; pp. 198-219; report no. EPA-600/4-84-049. Available from:
NTIS, Springfield, VA; PB84-169358/HSU.
Reiter, E. R. (1977a) Review and analysis. In: Mohnen, V. A.; Reiter, E. R., eds. International conference on
oxidants, 1976 - analysis of evidence and viewpoints; part III. the issue of stratospheric ozone
intrusion. Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental
Sciences Research Laboratory; pp. 67-117; report no. EPA-600/3-77-115. Available from: NTIS,
Springfield, VA; PB-279010.
Reiter, E. R. (1977b) The role of stratospheric import on tropospheric ozone concentrations. In: Dimitriades, B.,
ed. International conference on photochemical oxidant pollution and its control - proceedings:
volume I; September 1976; Raleigh, NC. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Environmental Sciences Research Laboratory; pp. 393-410; report no.
EPA-600/3-77-001a. Available from: NTIS, Springfield, VA; PB-264232.
Ridley, B. A. (1991) Recent measurements of oxidized nitrogen compounds in the troposphere. Atmos. Environ.
Part A 25: 1905-1926.
Ridley, B. A.; Shelter, J. D.; Walega, J. G.; Madronich, S.; Elsworth, C. M.; Grahek, F. E.; Fehsenfeld, F. C.;
Norton, R. B.; Parrish, D. D.; Htibler, G.; Buhr, M.; Williams, E. J.; Allwine, E. J.; Westberg, H.
H. (1990) The behavior of some organic nitrates at Boulder and Niwot Ridge, Colorado. J.
Geophys. Res. [Atmos.] 95: 13,949-13,961.
Runeckles, V C. (1974) Dosage of air pollutants and damage to vegetation. Environ. Conserv. 1: 305-308.
Runeckles, V. C.; Bates, D. V. (1991) The form of the ozone standard in relation to vegetation and health
effects. Presented at: 84th annual meeting and exhibition of the Air & Waste Management
Association; June; Vancouver, BC, Canada. Pittsburgh, PA: Air & Waste Management Association;
paper no. 91-74.3.
Sabersky, R. H.; Sinema, D. A.; Shair, F. H. (1973) Concentrations, decay rates, and removal of ozone and their
relation to establishing clean indoor air. Environ. Sci. Technol. 7: 347-353.
Sakugawa, H.; Kaplan, I. R. (1989) H2O2 and O3 in the atmosphere of Los Angeles and its vicinity: factors
controlling their formation and their role as oxidants of SO2. J. Geophys. Res. [Atmos.]
94: 12,957-12,973.
4-112
-------
Sakugawa, H.; Kaplan, I. R. (1993) Comparison of H2O2 and O3 content in atmospheric samples in the
San Bernardino mountains, southern California. Atmos. Environ. Part A 27: 1509-1515.
Sakugawa, H.; Kaplan, I. R.; Tsai, W.; Cohen, Y. (1990) Atmospheric hydrogen peroxide: does it share a role
with ozone in degrading air quality? Environ. Sci. Technol. 24: 1452-1462.
Sexton, K.; Ryan, P. B. (1988) Assessment of human exposure to air pollution: methods, measurements, and
models. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile, and
public health. Washington, DC: National Academy Press; pp. 207-238.
Shair, F. H.; Heitner, K. L. (1974) Theoretical model for relating indoor pollutant concentrations to those outside.
Environ. Sci. Technol. 8: 444-451.
Shaver, C. L.; Cass, G. R.; Druzik, J. R. (1983) Ozone and the deterioration of works of art. Environ. Sci.
Technol. 17: 748-752.
Shively, T. S. (1991) An analysis of the trend in ground-level ozone using non-homogeneous Poisson processes.
Atmos. Environ. Part B 25: 387-395.
Showman, R. E. (1991) A comparison of ozone injury to vegetation during moist and drought years. J. Air
Waste Manage. Assoc. 41: 63-64.
Singh, H. B.; Salas, L. J. (1989) Measurements of peroxyacetyl nitrate (PAN) and peroxypropionyl nitrate (PPN)
at selected urban, rural and remote sites. Atmos. Environ. 23: 231-238.
Singh, H. B.; Ludwig, F. L.; Johnson, W. B. (1977) Ozone in clean remote atmospheres: concentrations and
variabilities. New York, NY: Coordinating Research Council, Inc.; report no. CRC-APRAC
CAPA-15-76. Available from: NTIS, Springfield, VA; PB-272290.
Singh, H. B.; Ludwig, F. L.; Johnson, W. B. (1978) Tropospheric ozone: concentrations and variabilities in clean
remote atmospheres. Atmos. Environ. 12: 2185-2196.
Smith, L. A.; Lefohn, A. S. (1991) Co-occurence of ozone and wet deposited hydrogen ion in the United States.
Atmos. Environ. Part A 25: 2707-2716.
Spektor, D. M.; Lippmann, M.; Lioy, P. J.; Thurston, G. D.; Citak, K.; James, D. J.; Bock, N.; Speizer, F. E.;
Hayes, C. (1988) Effects of ambient ozone on respiratory function in active, normal children. Am.
Rev. Respir. Dis. 137: 313-320.
Spektor, D. M.; Thurston, G. D.; Mao, J.; He, D.; Hayes, C.; Lippmann, M. (1991) Effects of single- and
multiday ozone exposures on respiratory function in active normal children. Environ. Res.
55: 107-122.
Stasiuk, W. N., Jr.; Coffey, P. E. (1974) Rural and urban ozone relationships in New York State. J. Air Pollut.
Control Assoc. 24: 564-568.
Stock, T. H.; Holguin, A. H.; Selwyn, B. J.; Hsi, B. P.; Contant, C. F.; Buffler, P. A.; Kotchmar, D. J. (1983)
Exposure estimates for the Houston area asthma and runners studies. In: Lee, S. D.; Mustafa, M.
G.; Mehlman, M. A., eds. International symposium on the biomedical effects of ozone and related
photochemical oxidants; March 1982; Pinehurst, NC. Princeton, NJ: Princeton Scientific Publishers,
Inc.; pp. 539-548. (Advances in modern environmental toxicology: v. 5).
4-113
-------
Stoeckenius, T. (1991) Adjustment of ozone trends for meteorological variations. In: Berglund, R. L.;
Lawson, D. R.; McKee, D. I, eds. Tropospheric ozone and the environment: papers from an
international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
Association; pp. 781-800. (A&WMA transactions series no. TR-19).
Stoeckenius, T. E.; Hudischewskyj, A. B. (1990) Adjustment for ozone trends for meteorological variation.
San Rafael, CA: Systems Applications, Inc.; paper no. SYSAPP-90/008.
Tanner, R. L.; Miguel, A. H.; de Andrade, J. B.; Gaffney, J. S.; Streit, G. E. (1988) Atmospheric chemistry of
aldehydes: enhanced peroxyacetyl nitrate formation from ethanol-fueled vehicular emissions.
Environ. Sci. Technol. 22: 1026-1034.
Taylor, G. E., Jr.; Hanson, P. J. (1992) Forest trees and tropospheric ozone: role of canopy deposition and leaf
uptake in developing exposure-response relationships. Agric. Ecosyst. Environ. 42: 255-273.
Taylor, G. E., Jr.; Norby, R. J. (1985) The significance of elevated levels of ozone on natural ecosystems of
North America. In: Lee, S. D., ed. Evaluation of the scientific basis for ozone/oxidants standards:
proceedings of an APCA international specialty conference; November 1984; Houston, TX.
Pittsburgh, PA: Air Pollution Control Association; pp. 152-175. (APCA transactions: TR-4).
Taylor, G. E., Jr.; McLaughlin, S. B.; Shriner, D. S. (1982) Effective pollutant dose. In: Unsworth, M. H.;
Ormrod, D. P., eds. Effects of gaseous air pollution in agriculture and horticulture. London, United
Kingdom: Butterworth Scientific; pp. 458-460.
Taylor, G. E., Jr.; Ross-Todd, B. M.; Allen, E.; Conklin, P.; Edmonds, R.; Joranger, E.; Miller, E.; Ragsdale, L.;
Shepard, J.; Silsbee, D.; Swank, W. (1992) Patterns of tropospheric ozone in forested landscapes of
the Integrated Forest Study. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric deposition and
forest nutrient cycling: a synthesis of the integrated forest study. New York, NY: Springer-Verlag;
pp. 50-71. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological
studies: analysis and synthesis: v. 91).
Thompson, C. R. (1971) Measurement of total oxidant levels at Riverside Community Hospital. Arch. Environ.
Health 22: 514-515.
Thompson, A. M. (1992) The oxidizing capacity of the Earth's atmosphere: probable past and future changes.
Science (Washington, DC) 256: 1157-1165.
Thompson, C. R.; Hensel, E. G.; Kats, G. (1973) Outdoor-indoor levels of six air pollutants. J. Air Pollut.
Control Assoc. 23: 881-886.
Thurston, G. D.; Ito, K.; Kinney, P. L.; Lippmann, M. (1992) A multi-year study of air pollution and respiratory
hospital admissions in three New York State metropolitan areas: results for 1988 and 1989
summers. J. Exposure Anal. Environ. Epidemiol. 2: 429-450.
Thurston, G. D.; Gorczynski, J. E., Jr.; Currie, J. H.; He, D.; Ito, K.; Hipfner, J.; Waldman, J.; Lioy, P. J.;
Lippmann, M. (1994) The nature and origins of acid summer haze air pollution in metropolitan
Toronto, Ontario. Environ. Res. 65: 254-270.
Trainer, M.; Williams, E. J.; Parrish, D. D.; Buhr, M. P.; Allwine, E. J.; Westberg, H. H.; Fehsenfeld, F. C.; Liu,
S. C. (1987) Models and observations of the impact of natural hydrocarbons on rural ozone. Nature
(London) 329: 705-707.
Tsalkani, N.; Perros, P.; Dutot, A. L.; Toupance, G. (1991) One-year measurements of PAN in the Paris basin:
effect of meteorological parameters. Atmos. Environ. Part A 25: 1941-1949.
4-114
-------
Tsani-Bazaca, E.; Glavas, S.; Glisten, H. (1988) Peroxyacetyl nitrate (PAN) concentrations in Athens, Greece.
Atmos. Environ. 22: 2283-2286.
U.S. Environmental Protection Agency. (1978) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental
Criteria and Assessment Office; report no. EPA-600/8-78-004. Available from: NTIS, Springfield,
VA; PB80-124753.
U.S. Environmental Protection Agency. (1986) Air quality criteria for ozone and other photochemical oxidants.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental
Criteria and Assessment Office; report nos. EPA-600/8-84-020aF-eF. 5v. Available from: NTIS,
Springfield, VA; PB87-142949.
U.S. Environmental Protection Agency. (1989) Review of the national ambient air quality standards for ozone:
assessment of scientific and technical information. OAQPS staff paper. Research Triangle Park, NC:
Office of Air Quality Planning and Standards; report no. EPA-450/2-92/001. Available from: NTIS,
Springfield, VA; PB92-190446.
U.S. Environmental Protection Agency. (1991) Air quality criteria for carbon monoxide. Research Triangle Park,
NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; report no. EPA/600/8-90/045F. Available from: NTIS, Springfield, VA; PB93-167492.
U.S. Environmental Protection Agency. (1992a) National air quality and emissions trends report, 1991. Research
Triangle Park, NC: U.S. Environmental Protection Agency,Office of Air Quality Planning and
Standards; report no. EPA/450-R-92-001. Available from: NTIS, Springfield, VA;
PB93-143998/XAB.
U.S. Environmental Protection Agency. (1992b) Summary of selected new information on effects of ozone on
health and vegetation: supplement to 1986 air quality criteria for ozone and other photochemical
oxidants. Research Triangle Park, NC: Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office; report no. EPA/600/8-88/105F. Available from:
NTIS, Springfield, VA; PB92-235670.
U.S. Environmental Protection Agency. (1993) National air quality and emissions trends report, 1992. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/454/R-93/031.
Available from: NTIS, Springfield, VA; PB94-146669/XAB.
U.S. Environmental Protection Agency. (1994) National air quality and emissions trends report, 1993. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA 454/R-94-026.
Van Valin, C. C.; Ray, J. D.; Boatman, J. F.; Gunter, R. L. (1987) Hydrogen peroxide in air during winter over
the south-central United States. Geophys. Res. Lett. 14: 1146-1149.
Volz, A.; Kley, D. (1988) Evaluation of the Montsouris series of ozone measurements made in the nineteenth
century. Nature (London) 332: 240-242.
Vong, R. J.; Guttorp, P. (1991) Co-occurence of ozone and acidic cloudwater in high-elevation forests. Environ.
Sci. Technol. 25: 1325-1329.
Vukovich, F. M; Bach, W. D., Jr.; Crissman, B. W.; King, W. J. (1977) On the relationship between high ozone
in the rural surface layer and high pressure systems. Atmos. Environ. 11: 967-983.
4-115
-------
Wakim, P. G. (1990) 1981 to 1988 ozone trends adjusted to meteorological conditions for 13 metropolitan areas.
Presented at: 83rd annual meeting and exhibition of the Air & Waste Management Association;
June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; paper no. 90-97.9.
Weschler, C. I; Shields, H. C; Naik, D. V. (1989) Indoor ozone exposures. JAPCA 39: 1562-1568.
Wight, G. D.; Wolff, G. T.; Lioy, P. I; Meyers, R. E.; Cederwall, R. T. (1978) Formation and transport of ozone
in the northeast quadrant of the United States. In: Morris, A. L.; Barras, R. C., eds. Air quality
meteorology and atmospheric ozone: a symposium sponsored by ASTM Committee D-22 on
Methods of Sampling and Analysis of Atmospheres; July-August 1977; Boulder, CO. Philadelphia,
PA: American Society for Testing and Materials; pp. 445-457; ASTM special technical publication
653.
Williams, E. L., II; Grosjean, D. (1990) Southern California air quality study: peroxyacetyl nitrate.
Atmos. Environ. Part A 24: 2369-2377.
Williams, E. L., II; Grosjean, D. (1991) Peroxypropionyl nitrate at a Southern California mountain forest site.
Environ. Sci. Technol. 25: 653-659.
Williams, E. L., II; Grosjean, E.; Grosjean, D. (1993) Ambient levels of the peroxyacyl nitrates PAN, PPN, and
MPAN in Atlanta, Georgia. Air Waste 43: 873-879.
Winer, A. M.; Lurmann, F. W.; Coyner, L. A.; Colome, S. D.; Poe, M. P. (1989) Characterization of air
pollution exposures in the California south coast air basin: application of a new regional human
exposure (REHEX) model. Riverside, CA: University of California at Riverside.
Winner, W. E.; Lefohn, A. S.; Cotter, I. S.; Greitner, C. S.; Nellessen, J.; McEvoy, L. R., Jr.; Olson, R. L.;
Atkinson, C. J.; Moore, L. D. (1989) Plant responses to elevational gradients of O3 exposures in
Virginia. Proc. Natl. Acad. Sci. U. S. A. 86: 8828-8832.
Wolff, G. T.; Lioy, P. J. (1980) Development of an ozone river associated with synoptic scale episodes in the
eastern United States. Environ. Sci. Technol. 14: 1257-1260.
Wolff, G. T.; Stasiuk, W. N., Jr.; Coffey, P. E.; Pasceri, R. E. (1975) Aerial ozone measurements over
New Jersey, New York, and Connecticut. Presented at: 68th annual meeting of the Air Pollution
Control Association; June; Boston, MA. Pittsburgh, PA: Air Pollution Control Association;
paper no. 75-58.6.
Wolff, G. T.; Lioy, P. J.; Wight, G. D.; Meyers, R. E.; Cederwall, R. T. (1977) An investigation of long-range
transport of ozone across the midwestern and eastern United States. In: Dimitriades, B., ed.
International conference on photochemical oxidant pollution and its control - proceedings: volume I;
September 1976; Raleigh, NC. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Environmental Sciences Research Laboratory; pp. 307-317; report no. EPA-600/3-77-001a.
Available from: NTIS, Springfield, VA; PB-264232.
Wolff, G. T.; Lioy, P. J.; Taylor, R. S. (1987) The diurnal variations of ozone at different altitudes on a rural
mountain in the eastern United States. JAPCA 37: 45-48.
World Meteorological Organization. (1994) Scientific assessment of ozone depletion: 1994. Executive summary.
Geneva, Switzerland: World Meteorological Organization, United Nations Environment Programme;
report no. 37.
Yocom, J. E. (1982) Indoor-outdoor air quality relationships: a critical review. J. Air Pollut. Control Assoc. 32:
500-520.
4-116
-------
Zeldin, M. D.; Cassmassi, J. C.; Hoggan, M. (1991) Ozone trends in the South Coast Air Basin: an update.
In: Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment:
papers from an international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air &
Waste Management Association; pp. 760-771. (A&WMA transactions series no. TR-19).
Zhang, I; Lioy, P. J. (1994) Ozone in residential air: concentrations, I/O ratios, indoor chemistry, and exposures.
Indoor Air 4: 95-105.
Zimmerman, P. R. (1979) Determination of emission rates of hydrocarbons from indigenous species of
vegetation in the Tampa/St. Petersburg, Florida area. Atlanta, GA: U.S. Environmental Protection
Agency, Air Programs Branch; report no. EPA-904/9-77-028. Available from: NTIS, Springfield,
VA; PB-297 057.
4-117
-------
Appendix A
Abbreviations and Acronyms
ADOM
AGL
AIRS
AM
AQCD
AQCR
AUSPEX
C
C
CA
CAA
CAAA
CAL-RAMS
CAR
CASAC
CBM
CCM
CFC
CH3OH
CH4
CI
CIT
CL
CMB
CNG
CO
C02
CTWM
DIAL
Acid Deposition and Oxidant Model
Above ground level
Aerometric Information Retrieval System
Alveolar macrophage
Air Quality Criteria Document
Air Quality Control Region
Atmospheric Utility Signatures, Predictions, and Experiments
Carbon
Concentration
Chromotropic acid
Clean Air Act
Clean Air Act Amendments of 1990
Coast and Lake Regional Atmospheric Modeling System
Centriacinar region
Clean Air Scientific Advisory Committee
Carbon-bond mechanism
Community Climate Model
Chlorofluorocarbon
Methanol
Methane
Chemical ionization
California Institute of Technology
Chemiluminescence
Chemical mass balance
Compressed natural gas
Carbon monoxide
Carbon dioxide
Complex Terrain Wind Model
Differential absorption lidar
A-1
-------
DNPH
DOAS
DWM
BCD
EKMA
EMS
EPA
EPEM
EPRI
EPS
ERAQS
ETBE
EtOH
FDDA
FeSO4
FEV,
FVC
FID
FTIR
GC
GMEP
GPT
H+
HC
HCFC
HCHO
HNO2
HNO3
HO2
H202
HPLC
H2S04
1C
ID
I/O
2,4-Dinitrophenylhydrazine
Differential optical absorption spectrometry
Diagnostic Wind Model
Electron capture detection
Empirical Kinetic Modeling Approach
Emissions Modeling System
U.S. Environmental Protection Agency
Event Probability Exposure Model
Electric Power Research Institute
Emissions Preprocessor System
Eastern Regional Air Quality Study
Ethyl-tertiary-butyl ether
Ethanol
Four-dimensional data assimilation
Ferrous sulfate
Forced expiratory volume in 1 s
Forced vital capacity
Flame ionization detection
Fourier transform infrared absorption spectroscopy
Gas chromatography
Geocoded Model of Emissions and Projections
Gas-phase titration
Hydrogen ion
Hydrocarbon
Hydrochlorofluorocarbon
Formaldehyde
Nitrous acid
Nitric acid
Hydroperoxyl
Hydrogen peroxide
High-performance liquid chromatography
Sulfuric acid
Ion chromatography
Identification (number)
Indoor/outdoor
A-2
-------
IR
Infrared radiation
IR
LMOS
LPG
MBTH
MCCP
MM4/MM5
MOBILE
MODELS 3
MPAN
MSA
MSCET
MTBE
NA
NAAQS
NADP
NAMS
NAPAP
NAPBN
NAS
NBKI
NBS
NCAR
NCLAN
NDDN
NEM
NF
NH3
NH4HSO4
NH4OH
(NH4)2S04
Incremental reactivity
Lake Michigan Oxidant Study
Liquified petroleum gas
3 -Methyl-2-benzothiazolone hydr azone
Mountain Cloud Chemistry Program
Mesoscale Model, versions 4 and 5
U.S. Environmental Protection Agency emissions model for mobile
sources
Modeling framework that consolidates all of the U.S. Environmental
Protection Agency's three-dimensional photochemical air
quality models
Peroxymethacryloyl nitrate
Metropolitan Statistical Area
Month and state current emissions trends
Methyl-tertiary-butyl ether
Not available
National Ambient Air Quality Standards
National Atmospheric Deposition Program
National Air Monitoring Station
National Acid Precipitation Assessment Program
Western National Air Pollution Background Network
National Academy of Sciences
Neutral buffered potassium iodide
National Bureau of Standards; now National Institute of
Standards and Technology
National Center for Atmospheric Research
National Crop Loss Assessment Network
National Dry Deposition Network
National Air Quality Standards Exposure Model
National forest
Ammonia
Ammonium bisulfate
Ammonium hydroxide
Ammonium sulfate
A-3
-------
NIST
NM
NMHC
NMOC
NO
NO2
N2O
NO3
NOX
NP
NPN
NTP
03
OAQPS
Obs.
OH
OHBA
PAMS
PAN
PANs
PAR
PEL
PBzN
PDFID
PF/TPLIF
pH
PL
PLANR
PMN
ppmC
PPN
PSD
PVOC
QE
QH
r
National Institute of Standards and Technology
National monument
Nonmethane hydrocarbon
Nonmethane organic compound
Nitric oxide
Nitrogen dioxide
Nitrous oxide
Nitrate
Nitrogen oxides
National park
n-propyl nitrate
National Toxicology Program
Ozone
Office of Air Quality Planning and Standards
Observations
Hydroxyl
Hydroxybenzoic acid
Photochemical Aerometric Monitoring System
Peroxyacetyl nitrate
Peroxyacyl nitrates
Proximal alveolar region
Planetary boundary layer
Peroxybenzoyl nitrate
Cryogenic preconcentration-direct flame ionization detection
Photofragmentation two-photon laser-induced fluorescence
Hydrogen ion concentration
Liquid-phase vapor pressure
Practice for Low-cost Application in Nonattainment Regions
Polymorphonuclear leukocyte (also called neutrophil)
Parts per million carbon
Peroxypropionyl nitrate
Passive sampling device
Polar volatile organic compound
Latent heat flux
Heat flux
Linear regression correlation coefficient
A-4
-------
R2
RADM
RAPS
REHEX
RMSD
ROG
ROM
ROMNET
RT
SAB
SAI
SAPRC
SARMAP
SAROAD
SCAQS
SIP
SLAMS
SJVAQS
SO2
S042
SOS
SRM
SRP
STEM-II
SUM06
SUM07
SUM08
SURE
T
TAMS
TDLAS
Multiple correlation coefficient
Regional Acid Deposition Model
Regional Air Pollution Study
Regional Human Exposure Model
Root-mean-square difference
Reactive organic gas
Regional Oxidant Model
Regional Ozone Modeling for Northeast Transport program
Respiratory tract
Science Advisory Board
Systems Applications International
Statewide Air Pollution Research Center, University
of California, Riverside
San Joaquin Valley Air Quality Study (SJVAQS)/Atmospheric Utility
Signatures, Predictions, and Experiments (AUSPEX) Regional
Model Adaptation Project
Storage and Retrieval of Aerometric Data (U.S. Environmental
Protection Agency centralized database; superseded by
Aerometric Information Retrieval System [AIRS])
South Coast Air Quality Study (California)
State Implementation Plan
State and Local Air Monitoring Station
San Joaquin Valley Air Quality Study
Sulfur dioxide
Sulfate
Southern Oxidant Study
Standard reference material
Standard reference photometer
Sulfur Transport Eulerian Model (version II)
Seasonal sum of all hourly average concentrations DO.06 ppm
Seasonal sum of all hourly average concentrations DO.07 ppm
Seasonal sum of all hourly average concentrations DO.08 ppm
Sulfate Regional Experiment Program
Temperature
Toxic Air Monitoring Study (U.S. Environmental Protection Agency)
Tunable-diode laser absorption spectroscopy
A-5
-------
TEA
Tg
TGTP
TNMHC
TPLIF
TTFMS
UAM
UV
UV-B
VMT
VOC
VE
WFM
WMO/UNEP
W126
Triethanolamine
Teragram
The Global Thinking Project
Total nonmethane hydrocarbons
Two-photon laser-induced fluorescence
Two-tone frequency-modulated spectroscopy
Urban Airshed Model
Ultraviolet
Ultraviolet radiation of wavelengths 280 to 320 nm
Vehicle miles traveled
Volatile organic compound
Minute ventilation; expired volume per minute
White Face Mountain
World Meteorological Organization/United Nations Environment
Program
Cumulative integrated exposure index with a sigmoidal weighting
function
A-6
-------
Appendix B
Colloquial and Latin Names
Alder
Alder, red
Alder, speckled
Alfalfa
Almond
Apple
Apricot
Ash, green
Ash, white
Aspen, trembling
Avocado
Azalea
Barley, spring
Basswood (linden)
Bean, broad
Bean, bush
Bean, kidney, pinto, snap, white
Beech, European
Beet, sugar
Begonia
Begonia, bedding
Bentgrass
Birch, European white
Birch, downy
Birch, paper
Blackberry, common
Black-gram
Bluegrass, Kentucky
Alnus serrulata (Alton) Willdenow
Alnus rubra Bong.
Alnus incana (L.) Moench.
Medicago sativa L.
Prunus amygdalus Batsch cv. Nonpariel
Mains spp.
Prunus armeniaca L.
Fraxinus pennsylvanica Marsh.
Fraxinus americana L.
Populus tremuloides L.
Per sea americana Mill.
Rhododendron spp.
Hordeum vulgare L.
Tilia americana L.
Viciafaba L.
Phaseolus vulgaris L. var. humulis Alef.
Phaseolus vulgaris L.
Fagus sylvatica L.
Beta vulgaris L.
Begonia sp.
Begonia semperflorens Link & Otto
Agrostis capillaris L.
Betula pendula Roth.
Betula pubescens Ehrb.
Betula papyri/era Marsh.
Rubus allegheniensis Porter
Vigna mungo L.
B-l
-------
Buckhorn
Cabbage
Campion, bladder
Campion, moss
Carnation
Cedar, incense
Cedar, western red
Celery
Chestnut, American
Cherry, black
Chickpea
Clover, ladino, white
Clover, red
Corn
Cotton
Cottonwood (poplar)
Cress, garden
Cucumber
Dock
Fenugreek
Fescue, tall
Fir, balsam
Fir, Douglas
Fir, Douglas, big-cone
Fir, Fraser
Fir, silver
Fir, white
Geranium
Golden-rain
Poa praetensis L.
Plantago lanceolata L.
Brassica oleracea capitata L.
Silene cucabalus Wibel.
Silene acaulis L.
Dianthus caryophyllus L.
Libocedrus decurrens Torr. =
Calocedrus decurrens [Torr.] Florin.
Thuja plicata Donn ex D. Don
Apium graveolens L. var. dulce Pers.
Castenea dentata (Marsh.) Borkh.
Prunus serotina Ehrh.
Cicer arietinum L.
Trifolium rep ens L.
Trifolium pratense L.
Zea mays L.
Gossypium hirsutum L.
Populus deltoides Marsh
Lepidium sativum L.
Cucumis sativus L.
Rumex obtusifolius L.
Trigonella foenum-graecum L.
Festuca elatior L. = Festuca praetensis Huds.
Abies balsamea (L.) Mill.
Pseudotsuga menziesii (Mirb.) Franco.
Pseudotsuga macrocarpa (Vasey) Mayr
Abies balsamea (L.)fraseri (Pursh) Poir.
Abies alba Mill.
Abies concolor Lindl.
Pelargonium x hortorum Bailey
Koelreuteria paniculata Laxm.
Grape
Grape, wild
Vitis labruscana Bailey
Vitis spp.
B-2
-------
Grapefruit, Ruby Red
Grass, colonial bent
Grass, orchard
Grass, red
Grass, rye
Gum, sweet
Hemlock, eastern
Citrus parodist L.
Agrostis tennis Sibthorp.
Dactylis glomerata L.
Festuca rubra Gaud.
Lolium perenne L.
Liquidambar styraciflua L.
Hemlock, western
Ivy
Kenaf
Juniper, shore
Lemon, Volkamer
Lettuce
Lichen
Lichen, parmelia
Lichen, umbilical
Lilac
Locust, black
Locust, honey
Lupine
Mangel
Maple, red
Maple, sugar
Milkweed
Milkweed
Mint
Oak, California black
Oak, Canyon live
Oak, red
Oak, white
Oats
Tsuga canadensis (L.) Carr.
Tsuga heterophylla (Raf.) Sarg.
Hedera helix L.
Hibiscus cannabinus L.
Juniperus conferta Parl.
Citrus volkameriana Ten. & Pasq
Lactuca sativa L.
Lobaria spp.
Flavoparmelia caperata
Umbilicaria mammulata
Syringa vulgaris L.
Robinia pseudoacacia L.
Gleditsia triacanthos L.
Lupinus bicolor Lindl.
Beta vulgaris L.
Acer rubrum L.
Acer saccharum Marsh
Asclepias syriaca L.
Asclepias sp.
Mentha piperita L.
Quercus kelloggii Newb.
Quercus chrysolepis Liebm.
Quercus rubra L.
Quercus alba L.
Avena sativa L.
-------
Onion
Orange
Pea
Peach
Pepper
Pear
Petunia
Pine, eastern white
Pine, Coulter
Pine, Jeffrey
Pine, loblolly
Pine, pitch
Pine, ponderosa
Pine, Scots
Pine, shortleaf
Pine, Sierra lodgepole
Pine, slash
Pine, sugar
Pine, Table Mountain
Pine, Virginia
Plane, London
Plantain, (plantago) common
Plum
Poinsettia
Poplar, hybrid
Poplar, yellow or tulip
Potato
Radish
Radish
Rape, spring
Rhododendron, azalea
Rice, domestic
Sassafras
Sequoia, giant
Allium cepa L.
Citrus sinensis (L.) Osbeck
Pisum sativum L.
Prunus persica (L.) Batsch cv. Halford
Capsicum annuum L.
Pyrus pyrifolia Rhd. cv. 20th Century
Petunia hybrida Vilm.
Pinus strobus L.
Pinus coulteri D. Don
Pinus jeffreyi Grev. & Balf.
Pinus taeda L.
Pinus rigida Mill.
Pinus ponderosa Laws.
Pinus sylvestris L.
Pinus echinata Mill.
Pinus contorta var. murrayana (Grev. & Balf.)
Engelm.
Pinus elliotti Englem. ex Vasey
Pinus lambertiana Dougl.
Pinus pungens Lamb.
Pinus virginiana Mill.
Platanus x acerifolia (Ait.) Willd.
Plantago major L.
Prunus domestica L.
Euphorbia pulcherrima Willd.
Populus maximowiczii x P. trichocarpa
Liriodendron tulipifera L.
Solanum tuberosum L.
Raphanus sativus L.
Raphanus sativus L. cv. Cherry Bell
Brassica napus L. var. napus
Rhododendron obtusum (Lindl.) Planch.
Oryza sativa L.
Sassafras albidum [Nutt.] Nees
Sequoiadendron giganteum Buchholz
B-4
-------
Sorghum, hybrid
Sorghum bicolor (L.) Moench x Sorghum x
Soybean
Spinach
Spruce, Norway
Spruce, red
Spruce, sitka
Strawberry, cultivated
Strawberry, wild
Sunflower
Skunk bush
Sycamore
Timothy
Tobacco
Tomato
Virgin's Bower
Watermelon
Wheat
drummondii (Steudel) Millsp. & Chase
Glycine max (L.) Merr.
Spinacea oleracea L.
Picea abies (L.) Karst.
Picea rubens Sarg.
Picea sitchensis (Bong.) Carr.
Fragaria x ananassa Duch.
Fragaria virginiana Duch.
Helianthus annuus L.
Rhus trilobata Nutt.
Platanus occidentalis L.
Phleum prat ens e L.
Nicotiana tabacum L.
Lycopericon esculentum Mill.
Clematis virginiana L.
Citrullus lanatus (Thunb.) Mastsum & Nakai
Triticum aestivum L.
B-5
------- |